From d30ed259688bd06c3ec4038f8739f9fb407327df Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Sat, 28 Mar 2015 21:04:21 +0530
Subject: [PATCH 01/26] Added implementation and doc string

---
 doc/examples/plot_seam_carving.py   |  19 +++
 skimage/transform/__init__.py       |   4 +-
 skimage/transform/_seam_carving.pyx | 210 ++++++++++++++++++++++++++++
 skimage/transform/seam_carving.py   |  73 ++++++++++
 skimage/transform/setup.py          |   3 +
 5 files changed, 308 insertions(+), 1 deletion(-)
 create mode 100644 doc/examples/plot_seam_carving.py
 create mode 100644 skimage/transform/_seam_carving.pyx
 create mode 100644 skimage/transform/seam_carving.py

diff --git a/doc/examples/plot_seam_carving.py b/doc/examples/plot_seam_carving.py
new file mode 100644
index 00000000..1c2dcfc9
--- /dev/null
+++ b/doc/examples/plot_seam_carving.py
@@ -0,0 +1,19 @@
+from skimage import io, data
+from skimage import transform
+from skimage import color, filters
+from matplotlib import pyplot as plt
+
+def custom_sobel(img):
+    if img.ndim == 3:
+        img = color.rgb2gray(img)
+
+    return filters.sobel(img)
+
+img = data.coins()
+out  = transform.seam_carve(img, 'vertical', 80, energy_func = custom_sobel)
+out  = transform.seam_carve(out, 'horizontal', 70, energy_func = custom_sobel)
+
+io.imshow(out)
+plt.figure()
+io.imshow(img)
+io.show()
diff --git a/skimage/transform/__init__.py b/skimage/transform/__init__.py
index 3049ebb0..05484887 100644
--- a/skimage/transform/__init__.py
+++ b/skimage/transform/__init__.py
@@ -12,6 +12,7 @@ from ._geometric import (warp, warp_coords, estimate_transform,
 from ._warps import swirl, resize, rotate, rescale, downscale_local_mean
 from .pyramids import (pyramid_reduce, pyramid_expand,
                        pyramid_gaussian, pyramid_laplacian)
+from seam_carving import seam_carve
 
 
 __all__ = ['hough_circle',
@@ -43,4 +44,5 @@ __all__ = ['hough_circle',
            'pyramid_reduce',
            'pyramid_expand',
            'pyramid_gaussian',
-           'pyramid_laplacian']
+           'pyramid_laplacian',
+            'seam_carve']
diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
new file mode 100644
index 00000000..3d07beb5
--- /dev/null
+++ b/skimage/transform/_seam_carving.pyx
@@ -0,0 +1,210 @@
+# cython: cdivision=True
+# cython: boundscheck=False
+# cython: nonecheck=False
+# cython: wraparound=False
+import numpy as np
+cimport numpy as cnp
+
+
+cdef cnp.double_t ABSOLUTE_MAX = np.finfo(np.double).max
+
+
+cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
+                 cnp.double_t[::1] current_cost, cnp.double_t[::1] prev_cost,
+                 Py_ssize_t cols):
+    """Find a single vertical seam in an image that will be removed.
+    
+    Parameters
+    ----------
+    energy_img : (M, N) ndarray
+        The energy image where a higher value signifies a pixel of more
+        importance.
+    track_img : (M, N) ndarray
+        The image used to store the optimal decision made at each point while
+        finding a minimum cost path.
+    current_cost : (N, ) ndarray
+        An array to store the current cost of the optimal path for each column
+        in row currently being processed.
+    prev_cost : (N, ) ndarray
+        An array to store the current cost of the optimal path for each column
+        in row prior to the one being processed.
+    cols : int
+        The number of cols to process for seam carving. Columns with indices
+        more than `cols` are ignored.
+
+        
+    Returns
+    -------
+    seam : (M, ) ndarray
+        An array containing the index of the row of the pixel to be removed
+        for each column in the image.
+        
+    Notes
+    -----
+    `track_img`, `current_cost` and `prev_cost` are passed as arguments to
+    avoid memory allocation at each iteration of `_seam_carve_v`.
+    """
+   
+    cdef Py_ssize_t rows, row, col
+    rows = energy_img.shape[0]
+    cdef cnp.double_t tmp, min_cost
+    cdef Py_ssize_t offset, idx, offset_clip
+
+    cdef Py_ssize_t[::1] seam = np.zeros(rows, dtype=np.int)
+
+    for idx in range(cols):
+        prev_cost[idx] = energy_img[0, idx]
+
+    for row in range(1, rows):
+        for col in range(0, cols):
+            
+            min_cost = ABSOLUTE_MAX
+            for offset in range(-1, 2):
+                idx = col + offset
+                
+                if idx > cols - 1 or idx < 0:
+                    continue
+                
+                if prev_cost[idx] < min_cost:
+                    min_cost = prev_cost[idx]
+                    track_img[row, col] = offset
+
+            current_cost[col] = min_cost + energy_img[row, col]
+
+        prev_cost[:] = current_cost
+
+    seam[rows-1] = np.argmin(current_cost)
+
+    for row in range(rows-2, -1, -1):
+        col = seam[row + 1]
+        offset = track_img[row, col]
+        #print offset
+        seam[row] = seam[row + 1] + offset
+
+    return seam
+
+
+cdef remove_seam_h_2d(cnp.double_t[:, ::1] img, Py_ssize_t[::1] seam,
+                      Py_ssize_t cols):
+    cdef Py_ssize_t rows, row, col, idx
+    rows = img.shape[0]
+    """ Removes one horizontal seam from the image.
+
+    The method modifies `img` so that all pixels to the right of the vertical
+    seam are pushed one place left.
+
+    image : (M, N) ndarray
+        Input image whose vertical seam is to be removed.
+    seam : (M, ) ndarray
+        An array use to store the index of the column in the seam for each row.
+    cols : int
+        Number of columns in the input image to process. Column indices more
+        than `cols` are ingored.
+
+    Notes
+    -----
+    `seam` is passed as an argument so that we don't have to reallocate it for
+    each iteration in `_seam_carve_v`.
+    """"
+    
+    for row in range(rows):
+        for idx in range(seam[row], cols - 1):
+            img[row, idx] = img[row, idx + 1]
+
+
+cdef remove_seam_h_3d(cnp.double_t[:, :, ::1] img, Py_ssize_t[::1] seam,
+                      Py_ssize_t cols):
+    """ Removes one horizontal seam from the image.
+
+    The method modifies `img` so that all pixels to the right of the vertical
+    seam are pushed one place left.
+
+    image : (M, N, 3) ndarray
+        Input image whose vertical seam is to be removed.
+    seam : (M, ) ndarray
+        An array use to store the index of the column in the seam for each row.
+    cols : int
+        Number of columns in the input image to process. Column indices more
+        than `cols` are ingored.
+
+    Notes
+    -----
+    `seam` is passed as an argument so that we don't have to reallocate it for
+    each iteration in `_seam_carve_v`.
+    """"
+    cdef Py_ssize_t rows, row, col, idx
+    rows = img.shape[0]
+
+    for row in range(rows):
+        for idx in range(seam[row], cols - 1):
+            img[row, idx, :] = img[row, idx + 1, :]
+
+
+def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
+    """ Carve vertical seams off an image.
+
+    Carves out vertical seams off an image while using the given energy
+    function to decide the importance of each pixel.[1]
+
+    Parameters
+    ----------
+    image : (M, N) or (M, N, 3) ndarray
+        Input image whose vertical seams are to be removed.
+    iters : int
+        Number of vertical seams are to be removed.
+    energy_func : callable
+        The function used to decide the importance of each pixel. The higher
+        the value corresponding to a pixel, the more the algorithm will try
+        to keep it in the image. For every iteration `energy_func` is called
+        as `energy_func(image, *extra_args, **extra_kwargs)`, where `image`
+        is the cropped image during each iteration and is expected to return a
+        (M, N) ndarray depicting each pixel's importance.
+    extra_args : iterable
+        The extra arguments supplied to `energy_func`.
+    extra_kwargs : dict
+        The extra keyword arguments supplied to `energy_func`.
+    border : int
+        The number of pixels in the right and left end of the image to be
+        excluded from being considered for a seam. This is important as certain
+        filters just ignore image boundaries and set them to `0`.
+
+    Returns
+    -------
+    image : (M, N - iters) or (M, N - iters, 3) ndarray
+        The cropped image with the vertical seams removed.
+    
+    References
+    ----------
+    .. [1] Shai Avidan and Ariel Shamir
+           "Seam Carving for Content-Aware Image Resizing"
+           http://www.cs.jhu.edu/~misha/ReadingSeminar/Papers/Avidan07.pdf
+    """
+    cdef Py_ssize_t[::1] seam
+    cdef Py_ssize_t ndim = img.ndim
+    cdef Py_ssize_t cols = img.shape[1]
+
+    track_img = np.zeros(img.shape[0:2], dtype=np.int8)
+
+    current_cost = np.zeros_like(track_img[0], dtype = img.dtype)
+    prev_cost = np.zeros_like(track_img[0], dtype = img.dtype)
+
+    for i in range(iters):
+
+        sliced_img = img[:, 0:cols]
+        energy_img = energy_func(sliced_img, *extra_args, **extra_kwargs)
+        
+        # So that borders are ignored.
+        energy_img[:, 0:border] = ABSOLUTE_MAX
+        energy_img[:, cols-border:cols] = ABSOLUTE_MAX
+        
+        seam = find_seam_v(energy_img, track_img, current_cost, prev_cost,
+                           cols)
+
+        if ndim == 2:
+            remove_seam_h_2d(img, seam, cols)
+        elif ndim == 3:
+            remove_seam_h_3d(img, seam, cols)
+
+        cols -= 1
+
+    return img[:, 0:cols]
diff --git a/skimage/transform/seam_carving.py b/skimage/transform/seam_carving.py
new file mode 100644
index 00000000..c716c0d1
--- /dev/null
+++ b/skimage/transform/seam_carving.py
@@ -0,0 +1,73 @@
+from _seam_carving import _seam_carve_h
+from ..import filters
+from .. import util
+from .._shared import utils
+import numpy as np
+
+
+def seam_carve(img, mode, num, energy_func, extra_args = [],
+               extra_kwargs = {}, border=1, force_copy = True):
+    """ Carve vertical or horizontal seams off an image.
+
+    Carves out vertical/horizontal seams off an image while using the given
+    energy function to decide the importance of each pixel.
+
+    Parameters
+    ----------
+    image : (M, N) or (M, N, 3) ndarray
+        Input image whose vertical seams are to be removed.
+    mode : str {'horizontal', 'vertical'}
+        Indicates whether seams are to be removed vertically or horizontally.
+        Removing seams horizontally will decrease the height whereas removing
+        vertically will decrease the width.
+    num : int
+        Number of seams are to be removed.
+    energy_func : callable
+        The function used to decide the importance of each pixel. The higher
+        the value corresponding to a pixel, the more the algorithm will try
+        to keep it in the image. For every iteration `energy_func` is called
+        as `energy_func(image, *extra_args, **extra_kwargs)`, where `image`
+        is the cropped image during each iteration and is expected to return a
+        (M, N) ndarray depicting each pixel's importance.
+    extra_args : iterable, optional
+        The extra arguments supplied to `energy_func`.
+    extra_kwargs : dict, optional
+        The extra keyword arguments supplied to `energy_func`.
+    border : int, optional
+        The number of pixels in the right and left end of the image to be
+        excluded from being considered for a seam. This is important as certain
+        filters just ignore image boundaries and set them to `0`. By default
+        border is set to `1`.
+    force_copy : bool, optional
+        If set, the image is copied before being used by the method which
+        modifies it in place. Set this to `False` if the original image is no
+        loner needed after this opetration.
+
+    Returns
+    -------
+    out : ndarray
+        The cropped image with the seams removed.
+    """
+
+    utils.assert_nD(img, (2,3))
+    img = util.img_as_float(img)
+    
+
+    if mode == 'horizontal':
+        img = np.ascontiguousarray(img)
+        return _seam_carve_h(img, num, energy_func, extra_args ,extra_kwargs,
+                             border)
+    elif mode == 'vertical' :
+        if img.ndim == 3:
+            img = np.transpose(img, (1, 0, 2))
+        else:
+            img = img.T
+
+        img = np.ascontiguousarray(img)
+        out = _seam_carve_h(img, num, energy_func, extra_args , extra_kwargs,
+                            border)
+
+        if img.ndim == 3:
+            return  np.transpose(out, (1, 0, 2))
+        else:
+            return out.T
diff --git a/skimage/transform/setup.py b/skimage/transform/setup.py
index 22f31696..ff2e9bc6 100644
--- a/skimage/transform/setup.py
+++ b/skimage/transform/setup.py
@@ -16,6 +16,7 @@ def configuration(parent_package='', top_path=None):
     cython(['_hough_transform.pyx'], working_path=base_path)
     cython(['_warps_cy.pyx'], working_path=base_path)
     cython(['_radon_transform.pyx'], working_path=base_path)
+    cython(['_seam_carving.pyx'], working_path=base_path)
 
     config.add_extension('_hough_transform', sources=['_hough_transform.c'],
                          include_dirs=[get_numpy_include_dirs()])
@@ -27,6 +28,8 @@ def configuration(parent_package='', top_path=None):
                          sources=['_radon_transform.c'],
                          include_dirs=[get_numpy_include_dirs()])
 
+    config.add_extension('_seam_carving', sources=['_seam_carving.c'],
+                         include_dirs=[get_numpy_include_dirs()])
     return config
 
 if __name__ == '__main__':

From 29ec5ee3ecd7dac999cbbc49d149978c520f9ab9 Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Sun, 29 Mar 2015 15:25:08 +0530
Subject: [PATCH 02/26] Added example and test

---
 doc/examples/plot_seam_carving.py            | 37 +++++++++++++++-----
 skimage/transform/_seam_carving.pyx          | 12 +++----
 skimage/transform/seam_carving.py            | 12 +++++--
 skimage/transform/tests/test_seam_carving.py | 33 +++++++++++++++++
 4 files changed, 76 insertions(+), 18 deletions(-)
 create mode 100644 skimage/transform/tests/test_seam_carving.py

diff --git a/doc/examples/plot_seam_carving.py b/doc/examples/plot_seam_carving.py
index 1c2dcfc9..0583cd5d 100644
--- a/doc/examples/plot_seam_carving.py
+++ b/doc/examples/plot_seam_carving.py
@@ -1,19 +1,38 @@
+"""
+============
+Seam Carving
+============
+
+This example demonstrates how images can be resized using seam carving [1]_.
+Resizing often distorts contents in the image. Seam carving tries to resize
+images while trying to keep important content intact. In this example we are
+using the Sobel filter to signify the importance of each pixel.
+
+.. [1] Shai Avidan and Ariel Shamir
+       "Seam Carving for Content-Aware Image Resizing"
+       http://www.cs.jhu.edu/~misha/ReadingSeminar/Papers/Avidan07.pdf
+
+"""
 from skimage import io, data
 from skimage import transform
 from skimage import color, filters
 from matplotlib import pyplot as plt
 
-def custom_sobel(img):
-    if img.ndim == 3:
-        img = color.rgb2gray(img)
-
-    return filters.sobel(img)
 
 img = data.coins()
-out  = transform.seam_carve(img, 'vertical', 80, energy_func = custom_sobel)
-out  = transform.seam_carve(out, 'horizontal', 70, energy_func = custom_sobel)
+out  = transform.seam_carve(img, 'vertical', 80, energy_func = filters.sobel)
+out  = transform.seam_carve(out, 'horizontal', 70, energy_func = filters.sobel)
+resized = transform.resize(img, out.shape)
+
+plt.title('Original Image')
+io.imshow(img, plugin='matplotlib')
 
-io.imshow(out)
 plt.figure()
-io.imshow(img)
+plt.title('Resized Image Image')
+io.imshow(resized, plugin='matplotlib')
+
+plt.figure()
+plt.title('Resized Image Image')
+io.imshow(out, plugin='matplotlib')
+
 io.show()
diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index 3d07beb5..188c8dcb 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -84,7 +84,7 @@ cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
     return seam
 
 
-cdef remove_seam_h_2d(cnp.double_t[:, ::1] img, Py_ssize_t[::1] seam,
+cdef remove_seam_v_2d(cnp.double_t[:, ::1] img, Py_ssize_t[::1] seam,
                       Py_ssize_t cols):
     cdef Py_ssize_t rows, row, col, idx
     rows = img.shape[0]
@@ -105,14 +105,14 @@ cdef remove_seam_h_2d(cnp.double_t[:, ::1] img, Py_ssize_t[::1] seam,
     -----
     `seam` is passed as an argument so that we don't have to reallocate it for
     each iteration in `_seam_carve_v`.
-    """"
+    """
     
     for row in range(rows):
         for idx in range(seam[row], cols - 1):
             img[row, idx] = img[row, idx + 1]
 
 
-cdef remove_seam_h_3d(cnp.double_t[:, :, ::1] img, Py_ssize_t[::1] seam,
+cdef remove_seam_v_3d(cnp.double_t[:, :, ::1] img, Py_ssize_t[::1] seam,
                       Py_ssize_t cols):
     """ Removes one horizontal seam from the image.
 
@@ -131,7 +131,7 @@ cdef remove_seam_h_3d(cnp.double_t[:, :, ::1] img, Py_ssize_t[::1] seam,
     -----
     `seam` is passed as an argument so that we don't have to reallocate it for
     each iteration in `_seam_carve_v`.
-    """"
+    """
     cdef Py_ssize_t rows, row, col, idx
     rows = img.shape[0]
 
@@ -201,9 +201,9 @@ def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
                            cols)
 
         if ndim == 2:
-            remove_seam_h_2d(img, seam, cols)
+            remove_seam_v_2d(img, seam, cols)
         elif ndim == 3:
-            remove_seam_h_3d(img, seam, cols)
+            remove_seam_v_3d(img, seam, cols)
 
         cols -= 1
 
diff --git a/skimage/transform/seam_carving.py b/skimage/transform/seam_carving.py
index c716c0d1..e9c700b0 100644
--- a/skimage/transform/seam_carving.py
+++ b/skimage/transform/seam_carving.py
@@ -1,4 +1,4 @@
-from _seam_carving import _seam_carve_h
+from _seam_carving import _seam_carve_v
 from ..import filters
 from .. import util
 from .._shared import utils
@@ -47,6 +47,12 @@ def seam_carve(img, mode, num, energy_func, extra_args = [],
     -------
     out : ndarray
         The cropped image with the seams removed.
+
+    References
+    ----------
+    .. [1] Shai Avidan and Ariel Shamir
+           "Seam Carving for Content-Aware Image Resizing"
+           http://www.cs.jhu.edu/~misha/ReadingSeminar/Papers/Avidan07.pdf
     """
 
     utils.assert_nD(img, (2,3))
@@ -55,7 +61,7 @@ def seam_carve(img, mode, num, energy_func, extra_args = [],
 
     if mode == 'horizontal':
         img = np.ascontiguousarray(img)
-        return _seam_carve_h(img, num, energy_func, extra_args ,extra_kwargs,
+        return _seam_carve_v(img, num, energy_func, extra_args ,extra_kwargs,
                              border)
     elif mode == 'vertical' :
         if img.ndim == 3:
@@ -64,7 +70,7 @@ def seam_carve(img, mode, num, energy_func, extra_args = [],
             img = img.T
 
         img = np.ascontiguousarray(img)
-        out = _seam_carve_h(img, num, energy_func, extra_args , extra_kwargs,
+        out = _seam_carve_v(img, num, energy_func, extra_args , extra_kwargs,
                             border)
 
         if img.ndim == 3:
diff --git a/skimage/transform/tests/test_seam_carving.py b/skimage/transform/tests/test_seam_carving.py
new file mode 100644
index 00000000..0c476b2d
--- /dev/null
+++ b/skimage/transform/tests/test_seam_carving.py
@@ -0,0 +1,33 @@
+from skimage import transform
+import numpy as np
+from numpy import testing
+
+def energy(img):
+    if(img.ndim == 3):
+        return np.ascontiguousarray(img[:, :, 0])
+    return (1 - img)
+
+def test_seam_carving():
+    img = np.array([[0, 0, 1, 0, 0],
+                    [0, 0, 1, 0, 0],
+                    [0, 0, 1, 0, 0],
+                    [0, 1, 0, 0, 0],
+                    [1, 0 , 0, 0, 0]], dtype = np.float )
+
+    out = transform.seam_carve(img, 'horizontal', 1, energy, border=0)
+    testing.assert_allclose(out, 0)
+
+    img = img.T
+    out = transform.seam_carve(img, 'vertical', 1, energy, border=0)
+    testing.assert_allclose(out, 0)
+
+    img = img.T
+
+    img3 = np.dstack([img, img, img])
+
+    out = transform.seam_carve(img3, 'horizontal', 1, energy, border=0)
+    testing.assert_allclose(out, 0)
+
+
+    out = transform.seam_carve(img3, 'vertical', 1, energy, border=0)
+    testing.assert_allclose(out, 0)

From 1f570a145f9a61d25f025db1ec42dc66d109ac5c Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Sun, 29 Mar 2015 15:34:06 +0530
Subject: [PATCH 03/26] Formatting changes

---
 skimage/transform/seam_carving.py            | 16 +++++++---------
 skimage/transform/tests/test_seam_carving.py |  5 +++--
 2 files changed, 10 insertions(+), 11 deletions(-)

diff --git a/skimage/transform/seam_carving.py b/skimage/transform/seam_carving.py
index e9c700b0..0777e377 100644
--- a/skimage/transform/seam_carving.py
+++ b/skimage/transform/seam_carving.py
@@ -1,12 +1,11 @@
 from _seam_carving import _seam_carve_v
-from ..import filters
 from .. import util
 from .._shared import utils
 import numpy as np
 
 
-def seam_carve(img, mode, num, energy_func, extra_args = [],
-               extra_kwargs = {}, border=1, force_copy = True):
+def seam_carve(img, mode, num, energy_func, extra_args=[],
+               extra_kwargs={}, border=1, force_copy=True):
     """ Carve vertical or horizontal seams off an image.
 
     Carves out vertical/horizontal seams off an image while using the given
@@ -55,25 +54,24 @@ def seam_carve(img, mode, num, energy_func, extra_args = [],
            http://www.cs.jhu.edu/~misha/ReadingSeminar/Papers/Avidan07.pdf
     """
 
-    utils.assert_nD(img, (2,3))
+    utils.assert_nD(img, (2, 3))
     img = util.img_as_float(img)
-    
 
     if mode == 'horizontal':
         img = np.ascontiguousarray(img)
-        return _seam_carve_v(img, num, energy_func, extra_args ,extra_kwargs,
+        return _seam_carve_v(img, num, energy_func, extra_args, extra_kwargs,
                              border)
-    elif mode == 'vertical' :
+    elif mode == 'vertical':
         if img.ndim == 3:
             img = np.transpose(img, (1, 0, 2))
         else:
             img = img.T
 
         img = np.ascontiguousarray(img)
-        out = _seam_carve_v(img, num, energy_func, extra_args , extra_kwargs,
+        out = _seam_carve_v(img, num, energy_func, extra_args, extra_kwargs,
                             border)
 
         if img.ndim == 3:
-            return  np.transpose(out, (1, 0, 2))
+            return np.transpose(out, (1, 0, 2))
         else:
             return out.T
diff --git a/skimage/transform/tests/test_seam_carving.py b/skimage/transform/tests/test_seam_carving.py
index 0c476b2d..d17eee25 100644
--- a/skimage/transform/tests/test_seam_carving.py
+++ b/skimage/transform/tests/test_seam_carving.py
@@ -2,17 +2,19 @@ from skimage import transform
 import numpy as np
 from numpy import testing
 
+
 def energy(img):
     if(img.ndim == 3):
         return np.ascontiguousarray(img[:, :, 0])
     return (1 - img)
 
+
 def test_seam_carving():
     img = np.array([[0, 0, 1, 0, 0],
                     [0, 0, 1, 0, 0],
                     [0, 0, 1, 0, 0],
                     [0, 1, 0, 0, 0],
-                    [1, 0 , 0, 0, 0]], dtype = np.float )
+                    [1, 0, 0, 0, 0]], dtype=np.float)
 
     out = transform.seam_carve(img, 'horizontal', 1, energy, border=0)
     testing.assert_allclose(out, 0)
@@ -28,6 +30,5 @@ def test_seam_carving():
     out = transform.seam_carve(img3, 'horizontal', 1, energy, border=0)
     testing.assert_allclose(out, 0)
 
-
     out = transform.seam_carve(img3, 'vertical', 1, energy, border=0)
     testing.assert_allclose(out, 0)

From 2e237887ba91be44e95d4b437324621e3f03df68 Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Sun, 29 Mar 2015 15:41:36 +0530
Subject: [PATCH 04/26] corrected title in example

---
 doc/examples/plot_seam_carving.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/doc/examples/plot_seam_carving.py b/doc/examples/plot_seam_carving.py
index 0583cd5d..c989f0c4 100644
--- a/doc/examples/plot_seam_carving.py
+++ b/doc/examples/plot_seam_carving.py
@@ -28,11 +28,11 @@ plt.title('Original Image')
 io.imshow(img, plugin='matplotlib')
 
 plt.figure()
-plt.title('Resized Image Image')
+plt.title('Resized Image')
 io.imshow(resized, plugin='matplotlib')
 
 plt.figure()
-plt.title('Resized Image Image')
+plt.title('Resized using Seam-Carving')
 io.imshow(out, plugin='matplotlib')
 
 io.show()

From 222221c028f1333fe2bf9ccdd82d871451e4eb16 Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Sun, 29 Mar 2015 15:52:48 +0530
Subject: [PATCH 05/26] Added bento.info entry

---
 bento.info | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/bento.info b/bento.info
index cba403d9..9659df2a 100644
--- a/bento.info
+++ b/bento.info
@@ -161,6 +161,9 @@ Library:
     Extension: skimage.external.tifffile._tifffile
         Sources:
             skimage/external/tifffile/_tifffile.c
+    Extension: skimage.transform._seam_carving.pyx
+        Sources:
+            skimage/transform/_seam_carving.pyx
 
 Executable: skivi
     Module: skimage.scripts.skivi

From 1a8ce172a117791c2f14aef83913d9de102836bc Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Sun, 29 Mar 2015 15:55:30 +0530
Subject: [PATCH 06/26] Corrected bento entry

---
 bento.info | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/bento.info b/bento.info
index 9659df2a..7d21bfa8 100644
--- a/bento.info
+++ b/bento.info
@@ -161,7 +161,7 @@ Library:
     Extension: skimage.external.tifffile._tifffile
         Sources:
             skimage/external/tifffile/_tifffile.c
-    Extension: skimage.transform._seam_carving.pyx
+    Extension: skimage.transform._seam_carving
         Sources:
             skimage/transform/_seam_carving.pyx
 

From 12fb4cbb860c992f035e235e28fec829b9e55726 Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Mon, 30 Mar 2015 13:02:25 +0530
Subject: [PATCH 07/26] Removed debug print and corrected formatting

---
 skimage/transform/__init__.py                | 4 ++--
 skimage/transform/_seam_carving.pyx          | 1 -
 skimage/transform/seam_carving.py            | 2 +-
 skimage/transform/tests/test_seam_carving.py | 4 ++++
 4 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/skimage/transform/__init__.py b/skimage/transform/__init__.py
index 05484887..9b40b3a9 100644
--- a/skimage/transform/__init__.py
+++ b/skimage/transform/__init__.py
@@ -12,7 +12,7 @@ from ._geometric import (warp, warp_coords, estimate_transform,
 from ._warps import swirl, resize, rotate, rescale, downscale_local_mean
 from .pyramids import (pyramid_reduce, pyramid_expand,
                        pyramid_gaussian, pyramid_laplacian)
-from seam_carving import seam_carve
+from .seam_carving import seam_carve
 
 
 __all__ = ['hough_circle',
@@ -45,4 +45,4 @@ __all__ = ['hough_circle',
            'pyramid_expand',
            'pyramid_gaussian',
            'pyramid_laplacian',
-            'seam_carve']
+           'seam_carve']
diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index 188c8dcb..86f3634a 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -78,7 +78,6 @@ cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
     for row in range(rows-2, -1, -1):
         col = seam[row + 1]
         offset = track_img[row, col]
-        #print offset
         seam[row] = seam[row + 1] + offset
 
     return seam
diff --git a/skimage/transform/seam_carving.py b/skimage/transform/seam_carving.py
index 0777e377..beba4bb4 100644
--- a/skimage/transform/seam_carving.py
+++ b/skimage/transform/seam_carving.py
@@ -1,4 +1,4 @@
-from _seam_carving import _seam_carve_v
+from ._seam_carving import _seam_carve_v
 from .. import util
 from .._shared import utils
 import numpy as np
diff --git a/skimage/transform/tests/test_seam_carving.py b/skimage/transform/tests/test_seam_carving.py
index d17eee25..40758648 100644
--- a/skimage/transform/tests/test_seam_carving.py
+++ b/skimage/transform/tests/test_seam_carving.py
@@ -32,3 +32,7 @@ def test_seam_carving():
 
     out = transform.seam_carve(img3, 'vertical', 1, energy, border=0)
     testing.assert_allclose(out, 0)
+
+
+if __name__ == '__main__':
+    np.testing.run_module_suite()

From c332fd53637fbfb198c9c3e161a258c4e46aca9d Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Sat, 6 Jun 2015 21:46:50 +0530
Subject: [PATCH 08/26] docstring change

---
 skimage/transform/_seam_carving.pyx | 36 ++++++++++++++---------------
 1 file changed, 18 insertions(+), 18 deletions(-)

diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index 86f3634a..b91887a5 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -13,7 +13,7 @@ cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
                  cnp.double_t[::1] current_cost, cnp.double_t[::1] prev_cost,
                  Py_ssize_t cols):
     """Find a single vertical seam in an image that will be removed.
-    
+
     Parameters
     ----------
     energy_img : (M, N) ndarray
@@ -32,19 +32,19 @@ cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
         The number of cols to process for seam carving. Columns with indices
         more than `cols` are ignored.
 
-        
+
     Returns
     -------
     seam : (M, ) ndarray
         An array containing the index of the row of the pixel to be removed
         for each column in the image.
-        
+
     Notes
     -----
     `track_img`, `current_cost` and `prev_cost` are passed as arguments to
     avoid memory allocation at each iteration of `_seam_carve_v`.
     """
-   
+
     cdef Py_ssize_t rows, row, col
     rows = energy_img.shape[0]
     cdef cnp.double_t tmp, min_cost
@@ -57,14 +57,14 @@ cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
 
     for row in range(1, rows):
         for col in range(0, cols):
-            
+
             min_cost = ABSOLUTE_MAX
             for offset in range(-1, 2):
                 idx = col + offset
-                
+
                 if idx > cols - 1 or idx < 0:
                     continue
-                
+
                 if prev_cost[idx] < min_cost:
                     min_cost = prev_cost[idx]
                     track_img[row, col] = offset
@@ -85,27 +85,27 @@ cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
 
 cdef remove_seam_v_2d(cnp.double_t[:, ::1] img, Py_ssize_t[::1] seam,
                       Py_ssize_t cols):
-    cdef Py_ssize_t rows, row, col, idx
-    rows = img.shape[0]
-    """ Removes one horizontal seam from the image.
+    """ Removes one vertical seam from the image.
 
     The method modifies `img` so that all pixels to the right of the vertical
     seam are pushed one place left.
 
     image : (M, N) ndarray
-        Input image whose vertical seam is to be removed.
+      Input image whose vertical seam is to be removed.
     seam : (M, ) ndarray
-        An array use to store the index of the column in the seam for each row.
+      An array use to store the index of the column in the seam for each row.
     cols : int
-        Number of columns in the input image to process. Column indices more
-        than `cols` are ingored.
+      Number of columns in the input image to process. Column indices more
+      than `cols` are ingored.
 
     Notes
     -----
     `seam` is passed as an argument so that we don't have to reallocate it for
     each iteration in `_seam_carve_v`.
     """
-    
+    cdef Py_ssize_t rows, row, col, idx
+    rows = img.shape[0]
+
     for row in range(rows):
         for idx in range(seam[row], cols - 1):
             img[row, idx] = img[row, idx + 1]
@@ -171,7 +171,7 @@ def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
     -------
     image : (M, N - iters) or (M, N - iters, 3) ndarray
         The cropped image with the vertical seams removed.
-    
+
     References
     ----------
     .. [1] Shai Avidan and Ariel Shamir
@@ -191,11 +191,11 @@ def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
 
         sliced_img = img[:, 0:cols]
         energy_img = energy_func(sliced_img, *extra_args, **extra_kwargs)
-        
+
         # So that borders are ignored.
         energy_img[:, 0:border] = ABSOLUTE_MAX
         energy_img[:, cols-border:cols] = ABSOLUTE_MAX
-        
+
         seam = find_seam_v(energy_img, track_img, current_cost, prev_cost,
                            cols)
 

From cf98558196d5b0d57c413b82a0873462234689ce Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Sat, 6 Jun 2015 21:51:36 +0530
Subject: [PATCH 09/26] pep8 changes

---
 doc/examples/plot_seam_carving.py   |  6 +++---
 skimage/transform/_seam_carving.pyx | 10 +++++-----
 2 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/doc/examples/plot_seam_carving.py b/doc/examples/plot_seam_carving.py
index c989f0c4..6c31fbb8 100644
--- a/doc/examples/plot_seam_carving.py
+++ b/doc/examples/plot_seam_carving.py
@@ -15,13 +15,13 @@ using the Sobel filter to signify the importance of each pixel.
 """
 from skimage import io, data
 from skimage import transform
-from skimage import color, filters
+from skimage import filters
 from matplotlib import pyplot as plt
 
 
 img = data.coins()
-out  = transform.seam_carve(img, 'vertical', 80, energy_func = filters.sobel)
-out  = transform.seam_carve(out, 'horizontal', 70, energy_func = filters.sobel)
+out = transform.seam_carve(img, 'vertical', 80, energy_func=filters.sobel)
+out = transform.seam_carve(out, 'horizontal', 70, energy_func=filters.sobel)
 resized = transform.resize(img, out.shape)
 
 plt.title('Original Image')
diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index b91887a5..3b83e08f 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -6,7 +6,7 @@ import numpy as np
 cimport numpy as cnp
 
 
-cdef cnp.double_t ABSOLUTE_MAX = np.finfo(np.double).max
+cdef cnp.double_t DBL_MAX = np.finfo(np.double).max
 
 
 cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
@@ -22,10 +22,10 @@ cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
     track_img : (M, N) ndarray
         The image used to store the optimal decision made at each point while
         finding a minimum cost path.
-    current_cost : (N, ) ndarray
+    current_cost : (N,) ndarray
         An array to store the current cost of the optimal path for each column
         in row currently being processed.
-    prev_cost : (N, ) ndarray
+    prev_cost : (N,) ndarray
         An array to store the current cost of the optimal path for each column
         in row prior to the one being processed.
     cols : int
@@ -35,7 +35,7 @@ cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
 
     Returns
     -------
-    seam : (M, ) ndarray
+    seam : (M, ) ndarray of int
         An array containing the index of the row of the pixel to be removed
         for each column in the image.
 
@@ -58,7 +58,7 @@ cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
     for row in range(1, rows):
         for col in range(0, cols):
 
-            min_cost = ABSOLUTE_MAX
+            min_cost = DBL_MAX
             for offset in range(-1, 2):
                 idx = col + offset
 

From 2cc4066097bbf42ee965b876d4f1ae5a39adb04c Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Sat, 6 Jun 2015 22:42:18 +0530
Subject: [PATCH 10/26] 2d images converted to 3d before removing seams

---
 skimage/transform/_seam_carving.pyx | 50 ++++++-----------------------
 skimage/transform/seam_carving.py   | 27 ++++++----------
 2 files changed, 20 insertions(+), 57 deletions(-)

diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index 3b83e08f..3b64413c 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -9,7 +9,7 @@ cimport numpy as cnp
 cdef cnp.double_t DBL_MAX = np.finfo(np.double).max
 
 
-cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
+cdef _find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
                  cnp.double_t[::1] current_cost, cnp.double_t[::1] prev_cost,
                  Py_ssize_t cols):
     """Find a single vertical seam in an image that will be removed.
@@ -18,10 +18,11 @@ cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
     ----------
     energy_img : (M, N) ndarray
         The energy image where a higher value signifies a pixel of more
-        importance.
+        importance. Pixels with a lower value will be cropped first.
     track_img : (M, N) ndarray
         The image used to store the optimal decision made at each point while
-        finding a minimum cost path.
+        finding a minimum cost path. For each pixel it stores the offset that
+        produced that least cost.
     current_cost : (N,) ndarray
         An array to store the current cost of the optimal path for each column
         in row currently being processed.
@@ -83,35 +84,7 @@ cdef find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
     return seam
 
 
-cdef remove_seam_v_2d(cnp.double_t[:, ::1] img, Py_ssize_t[::1] seam,
-                      Py_ssize_t cols):
-    """ Removes one vertical seam from the image.
-
-    The method modifies `img` so that all pixels to the right of the vertical
-    seam are pushed one place left.
-
-    image : (M, N) ndarray
-      Input image whose vertical seam is to be removed.
-    seam : (M, ) ndarray
-      An array use to store the index of the column in the seam for each row.
-    cols : int
-      Number of columns in the input image to process. Column indices more
-      than `cols` are ingored.
-
-    Notes
-    -----
-    `seam` is passed as an argument so that we don't have to reallocate it for
-    each iteration in `_seam_carve_v`.
-    """
-    cdef Py_ssize_t rows, row, col, idx
-    rows = img.shape[0]
-
-    for row in range(rows):
-        for idx in range(seam[row], cols - 1):
-            img[row, idx] = img[row, idx + 1]
-
-
-cdef remove_seam_v_3d(cnp.double_t[:, :, ::1] img, Py_ssize_t[::1] seam,
+cdef remove_seam_v(cnp.double_t[:, :, ::1] img, Py_ssize_t[::1] seam,
                       Py_ssize_t cols):
     """ Removes one horizontal seam from the image.
 
@@ -189,20 +162,17 @@ def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
 
     for i in range(iters):
 
-        sliced_img = img[:, 0:cols]
+        sliced_img = np.squeeze(img[:, 0:cols])
         energy_img = energy_func(sliced_img, *extra_args, **extra_kwargs)
 
         # So that borders are ignored.
-        energy_img[:, 0:border] = ABSOLUTE_MAX
-        energy_img[:, cols-border:cols] = ABSOLUTE_MAX
+        energy_img[:, 0:border] = DBL_MAX
+        energy_img[:, cols-border:cols] = DBL_MAX
 
-        seam = find_seam_v(energy_img, track_img, current_cost, prev_cost,
+        seam = _find_seam_v(energy_img, track_img, current_cost, prev_cost,
                            cols)
 
-        if ndim == 2:
-            remove_seam_v_2d(img, seam, cols)
-        elif ndim == 3:
-            remove_seam_v_3d(img, seam, cols)
+        remove_seam_v(img, seam, cols)
 
         cols -= 1
 
diff --git a/skimage/transform/seam_carving.py b/skimage/transform/seam_carving.py
index beba4bb4..28d92b74 100644
--- a/skimage/transform/seam_carving.py
+++ b/skimage/transform/seam_carving.py
@@ -55,23 +55,16 @@ def seam_carve(img, mode, num, energy_func, extra_args=[],
     """
 
     utils.assert_nD(img, (2, 3))
-    img = util.img_as_float(img)
+    image = util.img_as_float(img)
 
-    if mode == 'horizontal':
-        img = np.ascontiguousarray(img)
-        return _seam_carve_v(img, num, energy_func, extra_args, extra_kwargs,
-                             border)
-    elif mode == 'vertical':
-        if img.ndim == 3:
-            img = np.transpose(img, (1, 0, 2))
-        else:
-            img = img.T
+    if image.ndim == 2:
+        image = image[..., np.newaxis]
 
-        img = np.ascontiguousarray(img)
-        out = _seam_carve_v(img, num, energy_func, extra_args, extra_kwargs,
-                            border)
+    if mode == 'vertical':
+        image = np.transpose(image, (1, 0, 2))
 
-        if img.ndim == 3:
-            return np.transpose(out, (1, 0, 2))
-        else:
-            return out.T
+    image = np.ascontiguousarray(image)
+    out = _seam_carve_v(image, num, energy_func, extra_args, extra_kwargs, border)
+    if mode == 'vertical':
+        out = np.transpose(out, (1, 0, 2))
+    return np.squeeze(out)

From 66caaa5b9fccd3daabff9fbcc42e50cfa0df41fa Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Sun, 7 Jun 2015 00:40:05 +0530
Subject: [PATCH 11/26] Corrected testing code

---
 skimage/transform/tests/test_seam_carving.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/skimage/transform/tests/test_seam_carving.py b/skimage/transform/tests/test_seam_carving.py
index 40758648..57df75e0 100644
--- a/skimage/transform/tests/test_seam_carving.py
+++ b/skimage/transform/tests/test_seam_carving.py
@@ -5,7 +5,7 @@ from numpy import testing
 
 def energy(img):
     if(img.ndim == 3):
-        return np.ascontiguousarray(img[:, :, 0])
+        img = np.ascontiguousarray(img[:, :, 0])
     return (1 - img)
 
 

From 37d8a8447b49f29f1782e4bc8e97f9712e26c36f Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Mon, 8 Jun 2015 23:11:57 +0530
Subject: [PATCH 12/26] Modifed code to use seam map

---
 doc/examples/plot_seam_carving.py   |  14 ++--
 skimage/transform/_seam_carving.pyx | 107 ++++++++++++++++++++++------
 skimage/transform/seam_carving.py   |   4 +-
 3 files changed, 94 insertions(+), 31 deletions(-)

diff --git a/doc/examples/plot_seam_carving.py b/doc/examples/plot_seam_carving.py
index 6c31fbb8..a994fef2 100644
--- a/doc/examples/plot_seam_carving.py
+++ b/doc/examples/plot_seam_carving.py
@@ -18,21 +18,23 @@ from skimage import transform
 from skimage import filters
 from matplotlib import pyplot as plt
 
+def nothing(img):
+    return img
 
-img = data.coins()
+img = data.coins()#io.imread('/home/vighnesh/images/seam_bw.png')
 out = transform.seam_carve(img, 'vertical', 80, energy_func=filters.sobel)
-out = transform.seam_carve(out, 'horizontal', 70, energy_func=filters.sobel)
+#out = transform.seam_carve(out, 'horizontal', 70, energy_func=filters.sobel)
 resized = transform.resize(img, out.shape)
 
 plt.title('Original Image')
-io.imshow(img, plugin='matplotlib')
+#io.imshow(img, plugin='matplotlib')
 
 plt.figure()
 plt.title('Resized Image')
-io.imshow(resized, plugin='matplotlib')
+#io.imshow(resized, plugin='matplotlib')
 
 plt.figure()
 plt.title('Resized using Seam-Carving')
-io.imshow(out, plugin='matplotlib')
+#io.imshow(out, plugin='matplotlib')
 
-io.show()
+#io.show()
diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index 3b64413c..3be4c54f 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -1,7 +1,3 @@
-# cython: cdivision=True
-# cython: boundscheck=False
-# cython: nonecheck=False
-# cython: wraparound=False
 import numpy as np
 cimport numpy as cnp
 
@@ -112,6 +108,65 @@ cdef remove_seam_v(cnp.double_t[:, :, ::1] img, Py_ssize_t[::1] seam,
             img[row, idx, :] = img[row, idx + 1, :]
 
 
+cdef _preprocess_image(cnp.double_t[:, ::1] energy_img,
+                       cnp.double_t[:, ::1] cumulative_img,
+                       cnp.int8_t[:, ::1] track_img,
+                       Py_ssize_t cols):
+
+    cdef Py_ssize_t r, c, offset, c_idx
+    cdef Py_ssize_t rows = energy_img.shape[0]
+    cdef cnp.double_t min_cost = DBL_MAX
+
+    for c in range(cols):
+        cumulative_img[0, c] = energy_img[0, c]
+
+
+    for r in range(1, rows):
+        for c in range(cols):
+            min_cost = DBL_MAX
+            for offset in range(-1, 2):
+
+                c_idx = c + offset
+                if (c_idx > cols - 1) or (c_idx < 0) :
+                    continue
+
+                if cumulative_img[r-1, c_idx] < min_cost:
+                    min_cost = cumulative_img[r-1, c_idx]
+                    track_img[r, c] = offset
+
+            #print "min_cost = ", min_cost
+            cumulative_img[r,c] = min_cost + energy_img[r, c]
+
+    #print "-------Cumulative Image --------"
+    #print np.array(cumulative_img)
+    #print "-------Energy Image --------"
+    #print np.array(energy_img)
+
+cdef cnp.uint8_t mark_seam(cnp.int8_t[:, ::1] track_img, Py_ssize_t start_index,
+                          cnp.uint8_t[:, ::1] seam_map):
+
+    cdef Py_ssize_t rows = track_img.shape[0]
+    cdef Py_ssize_t[::1] current_seam_indices = np.zeros(rows, dtype=np.int)
+    cdef Py_ssize_t row, col
+    cdef cnp.int8_t offset
+    cdef Py_ssize_t seams
+
+    current_seam_indices[rows - 1] = start_index
+    for row in range(rows - 2, -1, -1):
+        col = current_seam_indices[row+1]
+        offset = track_img[row, col]
+        col = col + offset
+        current_seam_indices[row] = col
+
+        if seam_map[row, col]:
+            return 0
+
+
+    for row in range(rows):
+        col = current_seam_indices[row]
+        seam_map[row, col] = 1
+
+    return 1
 def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
     """ Carve vertical seams off an image.
 
@@ -120,7 +175,7 @@ def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
 
     Parameters
     ----------
-    image : (M, N) or (M, N, 3) ndarray
+    img : (M, N) or (M, N, 3) ndarray
         Input image whose vertical seams are to be removed.
     iters : int
         Number of vertical seams are to be removed.
@@ -151,29 +206,35 @@ def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
            "Seam Carving for Content-Aware Image Resizing"
            http://www.cs.jhu.edu/~misha/ReadingSeminar/Papers/Avidan07.pdf
     """
-    cdef Py_ssize_t[::1] seam
-    cdef Py_ssize_t ndim = img.ndim
+    last_row_obj = np.zeros(img.shape[1], dtype=np.float)
+    seam_map_obj = np.zeros(img.shape[0:2], dtype=np.uint8)
+
+    cdef cnp.double_t[::1] last_row = last_row_obj
+    cdef Py_ssize_t[::1] sorted_indices
+    cdef cnp.uint8_t[:, ::1] seam_map = seam_map_obj
     cdef Py_ssize_t cols = img.shape[1]
 
-    track_img = np.zeros(img.shape[0:2], dtype=np.int8)
+    cdef cnp.double_t[:, :, ::1] image = img
+    cdef cnp.int8_t[:, ::1] track_img = np.zeros(img.shape[0:2], dtype=np.int8)
+    cdef cnp.double_t[:, ::1] cumulative_img = np.zeros(img.shape[0:2], dtype=np.float)
+    cdef cnp.double_t[:, ::1] energy_img
 
-    current_cost = np.zeros_like(track_img[0], dtype = img.dtype)
-    prev_cost = np.zeros_like(track_img[0], dtype = img.dtype)
+    energy_img_obj = energy_func(np.squeeze(img))
+    energy_img = energy_img_obj
 
-    for i in range(iters):
+    energy_img_obj[:, 0:border] = DBL_MAX
+    energy_img_obj[:, cols-border:cols] = DBL_MAX
 
-        sliced_img = np.squeeze(img[:, 0:cols])
-        energy_img = energy_func(sliced_img, *extra_args, **extra_kwargs)
+    _preprocess_image(energy_img, cumulative_img, track_img, cols)
+    last_row[...] = cumulative_img[-1, :]
+    sorted_indices = np.argsort(last_row_obj)
+    #print "Sorted Indices = ", np.array(sorted_indices)
+    #print "First sorted Index = ", sorted_indices[0]
+    #print "Last Row = ", np.array(energy_img[-1, :])
+    #print np.array()
 
-        # So that borders are ignored.
-        energy_img[:, 0:border] = DBL_MAX
-        energy_img[:, cols-border:cols] = DBL_MAX
-
-        seam = _find_seam_v(energy_img, track_img, current_cost, prev_cost,
-                           cols)
-
-        remove_seam_v(img, seam, cols)
-
-        cols -= 1
 
+    from skimage import io
+    io.imshow(seam_map_obj*255)
+    io.show()
     return img[:, 0:cols]
diff --git a/skimage/transform/seam_carving.py b/skimage/transform/seam_carving.py
index 28d92b74..af329036 100644
--- a/skimage/transform/seam_carving.py
+++ b/skimage/transform/seam_carving.py
@@ -60,11 +60,11 @@ def seam_carve(img, mode, num, energy_func, extra_args=[],
     if image.ndim == 2:
         image = image[..., np.newaxis]
 
-    if mode == 'vertical':
+    if mode == 'horizontal':
         image = np.transpose(image, (1, 0, 2))
 
     image = np.ascontiguousarray(image)
     out = _seam_carve_v(image, num, energy_func, extra_args, extra_kwargs, border)
-    if mode == 'vertical':
+    if mode == 'horizontal':
         out = np.transpose(out, (1, 0, 2))
     return np.squeeze(out)

From 49037f69b0041ddc3bd512c9b6697e5c9adb5340 Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Wed, 10 Jun 2015 00:11:03 +0530
Subject: [PATCH 13/26] added removal function

---
 doc/examples/plot_seam_carving.py   |  18 +--
 skimage/transform/_seam_carving.pyx | 175 +++++++++-------------------
 2 files changed, 65 insertions(+), 128 deletions(-)

diff --git a/doc/examples/plot_seam_carving.py b/doc/examples/plot_seam_carving.py
index a994fef2..bd935a31 100644
--- a/doc/examples/plot_seam_carving.py
+++ b/doc/examples/plot_seam_carving.py
@@ -15,26 +15,28 @@ using the Sobel filter to signify the importance of each pixel.
 """
 from skimage import io, data
 from skimage import transform
-from skimage import filters
+from skimage import filters, color
 from matplotlib import pyplot as plt
 
 def nothing(img):
     return img
 
-img = data.coins()#io.imread('/home/vighnesh/images/seam_bw.png')
-out = transform.seam_carve(img, 'vertical', 80, energy_func=filters.sobel)
+#img = io.imread('/home/vighnesh/images/castle.jpg')
+#img = color.rgb2gray(img)
+img = data.camera()
+out = transform.seam_carve(img, 'vertical', 50, energy_func=filters.sobel)
 #out = transform.seam_carve(out, 'horizontal', 70, energy_func=filters.sobel)
 resized = transform.resize(img, out.shape)
 
 plt.title('Original Image')
-#io.imshow(img, plugin='matplotlib')
+io.imshow(img, plugin='matplotlib')
 
-plt.figure()
-plt.title('Resized Image')
+#plt.figure()
+#plt.title('Resized Image')
 #io.imshow(resized, plugin='matplotlib')
 
 plt.figure()
 plt.title('Resized using Seam-Carving')
-#io.imshow(out, plugin='matplotlib')
+io.imshow(out, plugin='matplotlib')
 
-#io.show()
+io.show()
diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index 3be4c54f..d63374d4 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -4,111 +4,7 @@ cimport numpy as cnp
 
 cdef cnp.double_t DBL_MAX = np.finfo(np.double).max
 
-
-cdef _find_seam_v(cnp.double_t[:, ::1] energy_img, cnp.int8_t[:, ::1] track_img,
-                 cnp.double_t[::1] current_cost, cnp.double_t[::1] prev_cost,
-                 Py_ssize_t cols):
-    """Find a single vertical seam in an image that will be removed.
-
-    Parameters
-    ----------
-    energy_img : (M, N) ndarray
-        The energy image where a higher value signifies a pixel of more
-        importance. Pixels with a lower value will be cropped first.
-    track_img : (M, N) ndarray
-        The image used to store the optimal decision made at each point while
-        finding a minimum cost path. For each pixel it stores the offset that
-        produced that least cost.
-    current_cost : (N,) ndarray
-        An array to store the current cost of the optimal path for each column
-        in row currently being processed.
-    prev_cost : (N,) ndarray
-        An array to store the current cost of the optimal path for each column
-        in row prior to the one being processed.
-    cols : int
-        The number of cols to process for seam carving. Columns with indices
-        more than `cols` are ignored.
-
-
-    Returns
-    -------
-    seam : (M, ) ndarray of int
-        An array containing the index of the row of the pixel to be removed
-        for each column in the image.
-
-    Notes
-    -----
-    `track_img`, `current_cost` and `prev_cost` are passed as arguments to
-    avoid memory allocation at each iteration of `_seam_carve_v`.
-    """
-
-    cdef Py_ssize_t rows, row, col
-    rows = energy_img.shape[0]
-    cdef cnp.double_t tmp, min_cost
-    cdef Py_ssize_t offset, idx, offset_clip
-
-    cdef Py_ssize_t[::1] seam = np.zeros(rows, dtype=np.int)
-
-    for idx in range(cols):
-        prev_cost[idx] = energy_img[0, idx]
-
-    for row in range(1, rows):
-        for col in range(0, cols):
-
-            min_cost = DBL_MAX
-            for offset in range(-1, 2):
-                idx = col + offset
-
-                if idx > cols - 1 or idx < 0:
-                    continue
-
-                if prev_cost[idx] < min_cost:
-                    min_cost = prev_cost[idx]
-                    track_img[row, col] = offset
-
-            current_cost[col] = min_cost + energy_img[row, col]
-
-        prev_cost[:] = current_cost
-
-    seam[rows-1] = np.argmin(current_cost)
-
-    for row in range(rows-2, -1, -1):
-        col = seam[row + 1]
-        offset = track_img[row, col]
-        seam[row] = seam[row + 1] + offset
-
-    return seam
-
-
-cdef remove_seam_v(cnp.double_t[:, :, ::1] img, Py_ssize_t[::1] seam,
-                      Py_ssize_t cols):
-    """ Removes one horizontal seam from the image.
-
-    The method modifies `img` so that all pixels to the right of the vertical
-    seam are pushed one place left.
-
-    image : (M, N, 3) ndarray
-        Input image whose vertical seam is to be removed.
-    seam : (M, ) ndarray
-        An array use to store the index of the column in the seam for each row.
-    cols : int
-        Number of columns in the input image to process. Column indices more
-        than `cols` are ingored.
-
-    Notes
-    -----
-    `seam` is passed as an argument so that we don't have to reallocate it for
-    each iteration in `_seam_carve_v`.
-    """
-    cdef Py_ssize_t rows, row, col, idx
-    rows = img.shape[0]
-
-    for row in range(rows):
-        for idx in range(seam[row], cols - 1):
-            img[row, idx, :] = img[row, idx + 1, :]
-
-
-cdef _preprocess_image(cnp.double_t[:, ::1] energy_img,
+cdef _preprocess_image(cnp.double_t[:, :, ::1] energy_img,
                        cnp.double_t[:, ::1] cumulative_img,
                        cnp.int8_t[:, ::1] track_img,
                        Py_ssize_t cols):
@@ -118,7 +14,7 @@ cdef _preprocess_image(cnp.double_t[:, ::1] energy_img,
     cdef cnp.double_t min_cost = DBL_MAX
 
     for c in range(cols):
-        cumulative_img[0, c] = energy_img[0, c]
+        cumulative_img[0, c] = energy_img[0, c, 0]
 
 
     for r in range(1, rows):
@@ -135,7 +31,7 @@ cdef _preprocess_image(cnp.double_t[:, ::1] energy_img,
                     track_img[r, c] = offset
 
             #print "min_cost = ", min_cost
-            cumulative_img[r,c] = min_cost + energy_img[r, c]
+            cumulative_img[r,c] = min_cost + energy_img[r, c, 0]
 
     #print "-------Cumulative Image --------"
     #print np.array(cumulative_img)
@@ -159,14 +55,29 @@ cdef cnp.uint8_t mark_seam(cnp.int8_t[:, ::1] track_img, Py_ssize_t start_index,
         current_seam_indices[row] = col
 
         if seam_map[row, col]:
+            #print "---------- Seam conflict at ", row, col
             return 0
 
-
     for row in range(rows):
         col = current_seam_indices[row]
         seam_map[row, col] = 1
 
     return 1
+
+cdef remove_seam(cnp.double_t[:, :, ::1] img,
+                cnp.uint8_t[:, ::1] seam_map, Py_ssize_t cols):
+
+    cdef Py_ssize_t rows = img.shape[0]
+    cdef Py_ssize_t channels = img.shape[2]
+    cdef Py_ssize_t r, c, ch, shift
+
+    for r in range(rows):
+        shift = 0
+        for c in range(cols):
+            shift += seam_map[r, c]
+            for ch in range(channels):
+                img[r, c, ch] = img[r, c + shift, ch]
+
 def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
     """ Carve vertical seams off an image.
 
@@ -213,28 +124,52 @@ def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
     cdef Py_ssize_t[::1] sorted_indices
     cdef cnp.uint8_t[:, ::1] seam_map = seam_map_obj
     cdef Py_ssize_t cols = img.shape[1]
+    cdef Py_ssize_t rows = img.shape[0]
+    cdef Py_ssize_t seams_left = iters
+    cdef Py_ssize_t seams_removed
+    cdef Py_ssize_t seam_idx
 
     cdef cnp.double_t[:, :, ::1] image = img
     cdef cnp.int8_t[:, ::1] track_img = np.zeros(img.shape[0:2], dtype=np.int8)
     cdef cnp.double_t[:, ::1] cumulative_img = np.zeros(img.shape[0:2], dtype=np.float)
-    cdef cnp.double_t[:, ::1] energy_img
+    cdef cnp.double_t[:, :, ::1] energy_img
 
-    energy_img_obj = energy_func(np.squeeze(img))
+    energy_img_obj = energy_func(np.squeeze(img))[:, :, np.newaxis]**2
+    energy_img_obj = np.ascontiguousarray(energy_img_obj)
     energy_img = energy_img_obj
 
-    energy_img_obj[:, 0:border] = DBL_MAX
-    energy_img_obj[:, cols-border:cols] = DBL_MAX
+    energy_img_obj[:, 0:border, 0] = DBL_MAX
+    energy_img_obj[:, cols-border:cols, 0] = DBL_MAX
+    energy_img_obj[rows-border:rows,:,0] = energy_img_obj[rows-2*border:rows-border,:,0]
+
 
     _preprocess_image(energy_img, cumulative_img, track_img, cols)
     last_row[...] = cumulative_img[-1, :]
     sorted_indices = np.argsort(last_row_obj)
-    #print "Sorted Indices = ", np.array(sorted_indices)
-    #print "First sorted Index = ", sorted_indices[0]
-    #print "Last Row = ", np.array(energy_img[-1, :])
-    #print np.array()
+    seam_idx = 0
 
 
-    from skimage import io
-    io.imshow(seam_map_obj*255)
-    io.show()
-    return img[:, 0:cols]
+    while seams_left > 0:
+        #print "sorted indices", np.array(sorted_indices)[:10]
+        #print "sorted array ", np.sort(last_row_obj)[:10]
+        #print "Seam starting at : ", sorted_indices[seam_idx]
+        if mark_seam(track_img, sorted_indices[seam_idx], seam_map):
+            seams_left -= 1
+            cols -= 1
+            #print "Seam marked ", seam_idx
+            seam_idx += 1
+            continue
+        else:
+            print "Seams removed = ", seam_idx
+            seam_idx = 0
+            remove_seam(image, seam_map, cols)
+            remove_seam(energy_img, seam_map, cols)
+            seam_map[...] = 0
+            _preprocess_image(energy_img, cumulative_img, track_img, cols)
+            last_row[:cols] = cumulative_img[-1, :cols]
+            sorted_indices = np.argsort(last_row_obj)
+
+    #from skimage import io
+    #io.imshow(seam_map_obj*255)
+    #io.show()
+    return img#[:, 0:cols]

From eac5663251f599791c3db9475f62b5f516adf459 Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Wed, 10 Jun 2015 21:39:09 +0530
Subject: [PATCH 14/26] Changed signatures to use energy map

---
 doc/examples/plot_seam_carving.py   |  7 ++--
 skimage/transform/_seam_carving.pyx | 56 +++++++++++------------------
 skimage/transform/seam_carving.py   | 43 ++++++++++------------
 3 files changed, 43 insertions(+), 63 deletions(-)

diff --git a/doc/examples/plot_seam_carving.py b/doc/examples/plot_seam_carving.py
index bd935a31..67da1395 100644
--- a/doc/examples/plot_seam_carving.py
+++ b/doc/examples/plot_seam_carving.py
@@ -21,10 +21,11 @@ from matplotlib import pyplot as plt
 def nothing(img):
     return img
 
-#img = io.imread('/home/vighnesh/images/castle.jpg')
+img = io.imread('/home/vighnesh/images/rocket.jpg')
 #img = color.rgb2gray(img)
-img = data.camera()
-out = transform.seam_carve(img, 'vertical', 50, energy_func=filters.sobel)
+eimg = filters.sobel(color.rgb2gray(img))
+#img = data.camera()
+out = transform.seam_carve(img, eimg, 'vertical', 200)
 #out = transform.seam_carve(out, 'horizontal', 70, energy_func=filters.sobel)
 resized = transform.resize(img, out.shape)
 
diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index d63374d4..7dfcc8c4 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -78,11 +78,11 @@ cdef remove_seam(cnp.double_t[:, :, ::1] img,
             for ch in range(channels):
                 img[r, c, ch] = img[r, c + shift, ch]
 
-def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
+def _seam_carve_v(img, energy_map, iters, border):
     """ Carve vertical seams off an image.
 
     Carves out vertical seams off an image while using the given energy
-    function to decide the importance of each pixel.[1]
+    map to decide the importance of each pixel.[1]
 
     Parameters
     ----------
@@ -90,25 +90,21 @@ def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
         Input image whose vertical seams are to be removed.
     iters : int
         Number of vertical seams are to be removed.
-    energy_func : callable
-        The function used to decide the importance of each pixel. The higher
+    energy_map : (M, N) ndarray
+        The array to decide the importance of each pixel. The higher
         the value corresponding to a pixel, the more the algorithm will try
-        to keep it in the image. For every iteration `energy_func` is called
-        as `energy_func(image, *extra_args, **extra_kwargs)`, where `image`
-        is the cropped image during each iteration and is expected to return a
-        (M, N) ndarray depicting each pixel's importance.
-    extra_args : iterable
-        The extra arguments supplied to `energy_func`.
-    extra_kwargs : dict
-        The extra keyword arguments supplied to `energy_func`.
-    border : int
-        The number of pixels in the right and left end of the image to be
-        excluded from being considered for a seam. This is important as certain
-        filters just ignore image boundaries and set them to `0`.
+        to keep it in the image.
+    num : int
+        Number of seams are to be removed.
+    border : int, optional
+        The number of pixels in the right, left and bottom end of the image
+        to be excluded from being considered for a seam. This is important as
+        certain filters just ignore image boundaries and set them to `0`.
+        By default border is set to `1`.
 
     Returns
     -------
-    image : (M, N - iters) or (M, N - iters, 3) ndarray
+    image : (M, N - iters, 3) ndarray of float
         The cropped image with the vertical seams removed.
 
     References
@@ -118,11 +114,10 @@ def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
            http://www.cs.jhu.edu/~misha/ReadingSeminar/Papers/Avidan07.pdf
     """
     last_row_obj = np.zeros(img.shape[1], dtype=np.float)
-    seam_map_obj = np.zeros(img.shape[0:2], dtype=np.uint8)
 
     cdef cnp.double_t[::1] last_row = last_row_obj
     cdef Py_ssize_t[::1] sorted_indices
-    cdef cnp.uint8_t[:, ::1] seam_map = seam_map_obj
+    cdef cnp.uint8_t[:, ::1] seam_map = np.zeros(img.shape[0:2], dtype=np.uint8)
     cdef Py_ssize_t cols = img.shape[1]
     cdef Py_ssize_t rows = img.shape[0]
     cdef Py_ssize_t seams_left = iters
@@ -134,33 +129,25 @@ def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
     cdef cnp.double_t[:, ::1] cumulative_img = np.zeros(img.shape[0:2], dtype=np.float)
     cdef cnp.double_t[:, :, ::1] energy_img
 
-    energy_img_obj = energy_func(np.squeeze(img))[:, :, np.newaxis]**2
-    energy_img_obj = np.ascontiguousarray(energy_img_obj)
-    energy_img = energy_img_obj
-
-    energy_img_obj[:, 0:border, 0] = DBL_MAX
-    energy_img_obj[:, cols-border:cols, 0] = DBL_MAX
-    energy_img_obj[rows-border:rows,:,0] = energy_img_obj[rows-2*border:rows-border,:,0]
+    energy_map[:, 0:border] = DBL_MAX
+    energy_map[:, cols-border:cols] = DBL_MAX
+    energy_map[rows-border:rows, :] = energy_map[rows-2*border:rows-border, :]
 
+    energy_map = np.ascontiguousarray(energy_map[:, :, np.newaxis])
+    energy_img = energy_map
 
     _preprocess_image(energy_img, cumulative_img, track_img, cols)
     last_row[...] = cumulative_img[-1, :]
     sorted_indices = np.argsort(last_row_obj)
     seam_idx = 0
 
-
     while seams_left > 0:
-        #print "sorted indices", np.array(sorted_indices)[:10]
-        #print "sorted array ", np.sort(last_row_obj)[:10]
-        #print "Seam starting at : ", sorted_indices[seam_idx]
         if mark_seam(track_img, sorted_indices[seam_idx], seam_map):
             seams_left -= 1
             cols -= 1
-            #print "Seam marked ", seam_idx
             seam_idx += 1
             continue
         else:
-            print "Seams removed = ", seam_idx
             seam_idx = 0
             remove_seam(image, seam_map, cols)
             remove_seam(energy_img, seam_map, cols)
@@ -169,7 +156,4 @@ def _seam_carve_v(img, iters, energy_func, extra_args , extra_kwargs, border):
             last_row[:cols] = cumulative_img[-1, :cols]
             sorted_indices = np.argsort(last_row_obj)
 
-    #from skimage import io
-    #io.imshow(seam_map_obj*255)
-    #io.show()
-    return img#[:, 0:cols]
+    return img[:, 0:cols]
diff --git a/skimage/transform/seam_carving.py b/skimage/transform/seam_carving.py
index af329036..3caba0de 100644
--- a/skimage/transform/seam_carving.py
+++ b/skimage/transform/seam_carving.py
@@ -4,43 +4,36 @@ from .._shared import utils
 import numpy as np
 
 
-def seam_carve(img, mode, num, energy_func, extra_args=[],
-               extra_kwargs={}, border=1, force_copy=True):
+def seam_carve(img, energy_map, mode, num, border=1, force_copy=True):
     """ Carve vertical or horizontal seams off an image.
 
     Carves out vertical/horizontal seams off an image while using the given
-    energy function to decide the importance of each pixel.
+    energy map to decide the importance of each pixel.
 
     Parameters
     ----------
     image : (M, N) or (M, N, 3) ndarray
-        Input image whose vertical seams are to be removed.
+        Input image whose seams are to be removed.
+    energy_map : (M, N) ndarray
+        The array to decide the importance of each pixel. The higher
+        the value corresponding to a pixel, the more the algorithm will try
+        to keep it in the image.
     mode : str {'horizontal', 'vertical'}
         Indicates whether seams are to be removed vertically or horizontally.
         Removing seams horizontally will decrease the height whereas removing
         vertically will decrease the width.
     num : int
         Number of seams are to be removed.
-    energy_func : callable
-        The function used to decide the importance of each pixel. The higher
-        the value corresponding to a pixel, the more the algorithm will try
-        to keep it in the image. For every iteration `energy_func` is called
-        as `energy_func(image, *extra_args, **extra_kwargs)`, where `image`
-        is the cropped image during each iteration and is expected to return a
-        (M, N) ndarray depicting each pixel's importance.
-    extra_args : iterable, optional
-        The extra arguments supplied to `energy_func`.
-    extra_kwargs : dict, optional
-        The extra keyword arguments supplied to `energy_func`.
     border : int, optional
-        The number of pixels in the right and left end of the image to be
-        excluded from being considered for a seam. This is important as certain
-        filters just ignore image boundaries and set them to `0`. By default
-        border is set to `1`.
+        The number of pixels in the right, left and bottom end of the image
+        to be excluded from being considered for a seam. This is important as
+        certain filters just ignore image boundaries and set them to `0`.
+        By default border is set to `1`.
     force_copy : bool, optional
-        If set, the image is copied before being used by the method which
-        modifies it in place. Set this to `False` if the original image is no
-        loner needed after this opetration.
+        If set, the `image` and `energy_map` are copied before being used by
+        the method which modifies it in place. Set this to `False` if the
+        original image and the energy map are no longer needed after
+        this opetration.
 
     Returns
     -------
@@ -55,7 +48,8 @@ def seam_carve(img, mode, num, energy_func, extra_args=[],
     """
 
     utils.assert_nD(img, (2, 3))
-    image = util.img_as_float(img)
+    image = util.img_as_float(img, force_copy)
+    energy_map = util.img_as_float(energy_map, force_copy)
 
     if image.ndim == 2:
         image = image[..., np.newaxis]
@@ -64,7 +58,8 @@ def seam_carve(img, mode, num, energy_func, extra_args=[],
         image = np.transpose(image, (1, 0, 2))
 
     image = np.ascontiguousarray(image)
-    out = _seam_carve_v(image, num, energy_func, extra_args, extra_kwargs, border)
+    out = _seam_carve_v(image, energy_map, num, border)
+
     if mode == 'horizontal':
         out = np.transpose(out, (1, 0, 2))
     return np.squeeze(out)

From ae35dcceb4372ecefd9439a073dbb8ea09f90d50 Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Wed, 10 Jun 2015 22:28:09 +0530
Subject: [PATCH 15/26] docstring updates

---
 skimage/transform/_seam_carving.pyx | 69 +++++++++++++++++++++++------
 skimage/transform/seam_carving.py   |  1 +
 2 files changed, 57 insertions(+), 13 deletions(-)

diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index 7dfcc8c4..c2ad8bfa 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -8,6 +8,25 @@ cdef _preprocess_image(cnp.double_t[:, :, ::1] energy_img,
                        cnp.double_t[:, ::1] cumulative_img,
                        cnp.int8_t[:, ::1] track_img,
                        Py_ssize_t cols):
+    """ For each row, compute the lowest seam value for all its columns.
+
+    This function updates `cumulative_img` such that `cumulative_img[r, c]`
+    is the total energy of the lowest energy seam ending at `(r, c)`.
+
+    Parameters
+    ----------
+    energy_img : (M, N, 1) ndarray
+       The array of cost of removal of each pixel. Seam carving tries to avoid
+       pixels with high costs.
+    cumulative_img : (M, N) ndarray
+       The array to be updated with the total cost of lowest energy seams.
+    track_img : (M, N) ndarray
+        For each pixel, `track_img` stores the relative column offset in
+        the previous row which has the lowest value in `cumulative_img`. This
+        helps in in re-tracing the minimum cost seam.
+    cols : int
+        Number of columns to process.
+    """
 
     cdef Py_ssize_t r, c, offset, c_idx
     cdef Py_ssize_t rows = energy_img.shape[0]
@@ -16,7 +35,6 @@ cdef _preprocess_image(cnp.double_t[:, :, ::1] energy_img,
     for c in range(cols):
         cumulative_img[0, c] = energy_img[0, c, 0]
 
-
     for r in range(1, rows):
         for c in range(cols):
             min_cost = DBL_MAX
@@ -30,17 +48,33 @@ cdef _preprocess_image(cnp.double_t[:, :, ::1] energy_img,
                     min_cost = cumulative_img[r-1, c_idx]
                     track_img[r, c] = offset
 
-            #print "min_cost = ", min_cost
             cumulative_img[r,c] = min_cost + energy_img[r, c, 0]
 
-    #print "-------Cumulative Image --------"
-    #print np.array(cumulative_img)
-    #print "-------Energy Image --------"
-    #print np.array(energy_img)
+cdef cnp.uint8_t _mark_seam(cnp.int8_t[:, ::1] track_img,
+                            Py_ssize_t start_index,
+                            cnp.uint8_t[:, ::1] seam_map):
 
-cdef cnp.uint8_t mark_seam(cnp.int8_t[:, ::1] track_img, Py_ssize_t start_index,
-                          cnp.uint8_t[:, ::1] seam_map):
+    """ Re-trace the optimal seam from a given column in the last row.
 
+    This function tries to re-track an optimal seam from `start_index` and
+    tries to mark it in `seam_map`. If this seam intersects with any existing
+    seam in `seam_map` the function returns `0` without marking anything. Else
+    it marks the seam in `seam_map` and returns `1`.
+
+    track_img : (M, N) ndarray
+       The array of relative column indices as updated by `_preprocess_image`.
+    start_index : int
+       The column number of the bottom most row from where to start re-tracing
+       the seam.
+    seam_map : (M, N) ndarray
+        The array used to mark seams. If a pixel is marked as as seam it is set
+        to `1`, else `0`.
+
+    Returns
+    -------
+    out : int
+        `1` if seam was marked, `0` is seam inersects and was not marked.
+    """
     cdef Py_ssize_t rows = track_img.shape[0]
     cdef Py_ssize_t[::1] current_seam_indices = np.zeros(rows, dtype=np.int)
     cdef Py_ssize_t row, col
@@ -55,7 +89,6 @@ cdef cnp.uint8_t mark_seam(cnp.int8_t[:, ::1] track_img, Py_ssize_t start_index,
         current_seam_indices[row] = col
 
         if seam_map[row, col]:
-            #print "---------- Seam conflict at ", row, col
             return 0
 
     for row in range(rows):
@@ -64,9 +97,19 @@ cdef cnp.uint8_t mark_seam(cnp.int8_t[:, ::1] track_img, Py_ssize_t start_index,
 
     return 1
 
-cdef remove_seam(cnp.double_t[:, :, ::1] img,
+cdef _remove_seam(cnp.double_t[:, :, ::1] img,
                 cnp.uint8_t[:, ::1] seam_map, Py_ssize_t cols):
+    """ Removes marked seams from an image.
 
+    Parameters
+    ----------
+    img : (M, N, P) ndarray
+        Input image whose vertical seams are to be removed.
+    seam_map : (M, N) ndarray
+        Array with seams to be removed marked by non-zero entries.
+    cols : int
+        The number of colums to process.
+    """
     cdef Py_ssize_t rows = img.shape[0]
     cdef Py_ssize_t channels = img.shape[2]
     cdef Py_ssize_t r, c, ch, shift
@@ -142,15 +185,15 @@ def _seam_carve_v(img, energy_map, iters, border):
     seam_idx = 0
 
     while seams_left > 0:
-        if mark_seam(track_img, sorted_indices[seam_idx], seam_map):
+        if _mark_seam(track_img, sorted_indices[seam_idx], seam_map):
             seams_left -= 1
             cols -= 1
             seam_idx += 1
             continue
         else:
             seam_idx = 0
-            remove_seam(image, seam_map, cols)
-            remove_seam(energy_img, seam_map, cols)
+            _remove_seam(image, seam_map, cols)
+            _remove_seam(energy_img, seam_map, cols)
             seam_map[...] = 0
             _preprocess_image(energy_img, cumulative_img, track_img, cols)
             last_row[:cols] = cumulative_img[-1, :cols]
diff --git a/skimage/transform/seam_carving.py b/skimage/transform/seam_carving.py
index 3caba0de..e037b6e1 100644
--- a/skimage/transform/seam_carving.py
+++ b/skimage/transform/seam_carving.py
@@ -62,4 +62,5 @@ def seam_carve(img, energy_map, mode, num, border=1, force_copy=True):
 
     if mode == 'horizontal':
         out = np.transpose(out, (1, 0, 2))
+
     return np.squeeze(out)

From 1e49f498465afbff28a15655c67afcf9e261f648 Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Thu, 11 Jun 2015 00:14:39 +0530
Subject: [PATCH 16/26] Added rocket image

---
 doc/examples/plot_seam_carving.py |  82 ++++++++++++++++++++++++------
 skimage/data/__init__.py          |  19 +++++++
 skimage/data/rocket.jpg           | Bin 0 -> 112525 bytes
 3 files changed, 85 insertions(+), 16 deletions(-)
 create mode 100644 skimage/data/rocket.jpg

diff --git a/doc/examples/plot_seam_carving.py b/doc/examples/plot_seam_carving.py
index 67da1395..9ed7181c 100644
--- a/doc/examples/plot_seam_carving.py
+++ b/doc/examples/plot_seam_carving.py
@@ -13,31 +13,81 @@ using the Sobel filter to signify the importance of each pixel.
        http://www.cs.jhu.edu/~misha/ReadingSeminar/Papers/Avidan07.pdf
 
 """
-from skimage import io, data
-from skimage import transform
+from skimage import data, draw
+from skimage import transform, util
+import numpy as np
 from skimage import filters, color
 from matplotlib import pyplot as plt
 
-def nothing(img):
-    return img
 
-img = io.imread('/home/vighnesh/images/rocket.jpg')
-#img = color.rgb2gray(img)
+hl_color = np.array([0, 1, 0])
+
+img = data.rocket()
+img = util.img_as_float(img)
 eimg = filters.sobel(color.rgb2gray(img))
-#img = data.camera()
-out = transform.seam_carve(img, eimg, 'vertical', 200)
-#out = transform.seam_carve(out, 'horizontal', 70, energy_func=filters.sobel)
-resized = transform.resize(img, out.shape)
 
 plt.title('Original Image')
-io.imshow(img, plugin='matplotlib')
+plt.imshow(img)
 
-#plt.figure()
-#plt.title('Resized Image')
-#io.imshow(resized, plugin='matplotlib')
+"""
+.. image:: PLOT2RST.current_figure
+"""
 
+resized = transform.resize(img, (img.shape[0], img.shape[1] - 200))
+plt.figure()
+plt.title('Resized Image')
+plt.imshow(resized)
+
+
+"""
+.. image:: PLOT2RST.current_figure
+"""
+
+out = transform.seam_carve(img, eimg, 'vertical', 200)
 plt.figure()
 plt.title('Resized using Seam-Carving')
-io.imshow(out, plugin='matplotlib')
+plt.imshow(out)
 
-io.show()
+"""
+.. image:: PLOT2RST.current_figure
+
+As you can see, resizing as distorted the rocket and the objects around,
+whereas seam carving has reszied by removing the empty spaces in between.
+
+Object Removal
+--------------
+
+Seam Carving can also be used to remove atrifacts from images. To do that, we
+have to ensure that pixels to be removes get less importance. In the following
+code I approximately mark the rocket with a mask, and then decrease the
+importance of those pixels
+
+"""
+
+masked_img = img.copy()
+
+poly = [(404, 281), (404, 360), (359, 364), (338, 337), (145, 337), (120, 322), (145, 304), (340, 306), (362, 284)]
+pr = np.array([p[0] for p in poly])
+pc = np.array([p[1] for p in poly])
+rr, cc = draw.polygon(pr, pc)
+
+masked_img[rr,cc,:] = masked_img[rr,cc,:]*0.5 + hl_color*.5
+plt.figure()
+plt.title('Object Marked')
+
+plt.imshow(masked_img)
+"""
+.. image:: PLOT2RST.current_figure
+"""
+
+eimg[rr, cc] -= 1000
+
+plt.figure()
+plt.title('Object Removed')
+out = transform.seam_carve(img, eimg, 'vertical', 90)
+resized = transform.resize(img, out.shape)
+plt.imshow(out)
+plt.show()
+"""
+.. image:: PLOT2RST.current_figure
+"""
diff --git a/skimage/data/__init__.py b/skimage/data/__init__.py
index 2be829ac..98fb4e5f 100644
--- a/skimage/data/__init__.py
+++ b/skimage/data/__init__.py
@@ -26,6 +26,7 @@ __all__ = ['load',
            'chelsea',
            'coffee',
            'hubble_deep_field',
+           'rocket',
            'astronaut']
 
 
@@ -241,3 +242,21 @@ def hubble_deep_field():
 
     """
     return load("hubble_deep_field.jpg")
+
+def rocket():
+    """Launch photo of DSCOVR on Falcon 9 by SpaceX.
+
+    This is the launch photo of Falcon 9 carrying DSCOVR lifted off from
+    SpaceX's Launch Complex 40 at Cape Canaveral Air Force Station, Fla..
+
+    Notes
+    -----
+    This image was downloaded from
+    `SpaceX Photos
+    <https://www.flickr.com/photos/spacexphotos/16511594820/in/photostream/>`__.
+
+    The image was captured by SpaceX and `released in the public domain
+    <http://arstechnica.com/tech-policy/2015/03/elon-musk-puts-spacex-photos-into-the-public-domain/>`_.
+
+    """
+    return load("rocket.jpg")
diff --git a/skimage/data/rocket.jpg b/skimage/data/rocket.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..32bdd33960c60d012fc71b5fd80a7ccc24d3b906
GIT binary patch
literal 112525
zcmb@tby!<L*Do4EfCNoK3lu022vUj{DOMo3JH?AzvEmeHkpw9&#ofJFakm1+t+*AZ
zP_$5}-1K|TbKmpdz2|w(B-#1R8e4m3W|qv#--W*$0H~6jq8tDR2M3^leF6S%0VNfs
zrC(@hsmm#<$YKis03eT)3`!OV1pu6!J>0b9Q4IP9h71G~0DJ&0;D3_2rMs(?hK9<&
zjI)cg^?#`U`}M~>0NWD)FvIz;um8Q|{|6?tbanH<*87G1@L5^ATLJ($JXrkF+rt%0
z$H~OvdloPM2iN?=ZaP}hSb9GeliL0V&;G*}|H0e;up9O!uyro~)wlg03<Uqf7XQJt
z|FDO(&%bd2(k`yPZuYiz9t^y~!omzvRxTFS4BEc#9@b9o42sT{E^e+aZss1=R@l1#
z;qPB70PH^%|2;7_Y{RFI{)hU1@$&C?_)m8in*Z~BaVSRrOZ)FYmY!~2|0{<Bd;$Rc
zf93z%M<x{jpk4$3fPVd##vcs;<Z=K2Wbgk=W8nn=pxgjJXXf8O09s2Y*C$-OJX}w>
z_*7m!Vc_u*5)ibpv9J*W{9Oh}|C??9E8*he0D-u`e+h((3&h0(;o;$dAP_u!GJHHD
z0(=OBgph;)42HpAc=&`wgkT~vFbw>!6C7M@eQYc28;l3R1OFeVzuf>B7^ealh>HUQ
z0AV<|Fr2^r06OfH0O4Y}`+pxEE)ax=k3#?kKnStrmQVl=mQ_4lEN@_Zd;)v|APz17
zga^lmkrN;odHL>1p}<U<TILq7?mux03XZ2hN^2+N<=6H=_$}S*dZ(r-nPoyYPgw+X
z3hK{ft<Z@{Tjx})x*nblGeXv0n7)4OAhb9D;6EPz%f!Dy@o@18z}Qk%7yt*$6fTGW
z9~U1F7l@CGtq#LQfZ)7Rc$((qjIV<3@uA}LYJ2dRv@E7h-GVpoKS{{1J45nIYxfdZ
zP74%JFuNz#Z|Q`b|6KwQ|ED<&AOZOGX7NP6Y|(-d{QagEb>1i}O>F|PqlsJw%n{*D
z6jR@SLDj-FM@w>z=1aB1e0t;fhBv1_oj!0xML7fX%lDXl4LtZ*vG(`fe0};G-SbkB
z+1dKKy7TP|W_Ahwjo7!5GgP|9^Z1%^R!wT@hd5o|s`Q4_(Xi?c-B)D>7S16{k1UR2
zXFS{coQ}kth4;^{Z)?e$6U%m6CAskNLlP3wY_c68*>ouo%Z&>W85z^{=TmW?_)hK&
z6)6&|20H03whL)iky7V$JGRal63VX*b>RkD@88rYjZvnhU3HvpT?H)EhxO~;yJNi$
z$xV)vxp^o#sUWmZNjY7(N6^$N8$=xob+p|)XJN@s&&VGe9wkI?#$-fwaT%slzB==0
z&0c%uRb`{sCtB~h#SlzO5Vhmy?_9x`m}XMtS}PL5$=bbIcv)*Cmm!}}q^$H=;d5??
zV<eL|Z1F7moerD%GJggrGHW7Y{nFq)yKcGpp0s(5R-<2gLEUa0;xve8{&W=xKc=Ij
zg4P$!0O6Ov?HHmpYF(s^fUt(&=0OlcDImOI12q<hcs&H|umUUA(V7A&JgU$F!I&(r
zRM#mtbk$P@rI|A$U>r1wvpan}rFS1%Te2BMME$k^_-06fWkZo69b<*dsKuXGsPrSJ
zEfYe-VISQGp%x&Lzsyx+RsIqr4!z7P(BX5*2PJ@g%ngVrJiSpOB>@_jzZVz^dYIHJ
z-hoky0uWB<5a(%k5l@G$Mvs+Bp8c7}X^(yDOAs^{gjl>nF?ljraX?>B#@8f8GEDJD
zKNZeHrr)4w<Iqwky(mf*C<+c{K&f`^qO(TscV0fiW7|!%?dSu!t0D6iZ#5G4J6vEU
zNnq5;`n6PbsU&htsJ9Wr#;2v?t{#?;OO6M$(Bc~=Vw2Z_Nbyc2zr_jHAy@pG_w&A>
zVS(cAGY6a2W8O-3Hwi)Rx|OzvX@3FfDKvpiH!W<jC5qE$@>vznnJIW6li9gJxLd}r
zlbaTeNNgNp-hMZ_YoN3BEGd_GH2p`)*)*4{p@9Dgb;me?BS@>z!RlqqJC-XZTn*zb
z+Wm7jr>3mHbu?pD?pS1Kopun$Z0uY(vzivng})QvHqmdg{!`x2$8gT!9e?1i%?w41
z6^V+9bM9foK~qE7X}e#`PoD~rCecJ+D1-;GxSV3=Kl3Kv<oeeeUe8q9P31E2?x>D^
zxPnNUXmnL!HCg4-7+Ep#Sj4l*)EP;CPxQMuqOo6<M~B-pE}vT}ZM`#QYr2{!#OcY<
z6>?QL$5K!DXi4@~>DasMpF>BYDsI%{K}J{dh+(1F*;-#6q1NtD<Hxrfa_NPoHW|67
z`%M8<(9pN_z?XhTbH|Yd>6J;DdOcv%lcpIGh@M9`CQE!B42`#a{=-}*+|5lY?jzmg
zLDC}RQA-51pVQdAsKK(j)$}I)_3LEbn`m~%DkjxjgZ}OqNq?oup3Z`@*_WBBO9Y*N
z0U!r18mF@H&pbU5BRrp<(8nZNID_M#`mTO<lX+zRNO;jftIbTr<R#F=Pm&S5;KTLt
zaYt_Qv+)G5bMPeN1trL#(mr^94ib8(%ICDx|M_;DrXFuoVD|J*^lNjs*_qa%)6ha~
zd)tN7vk#oBRcLkc$bR$gm#b;Zk$|a@wJ+_Ya!0I^#VuSLBx0TVNgrn{8&xRU-W-1#
z`U`jqk(b8wq}#CdrGX+?B)4s)eO9ROWVT*0e-%(9t>NK#y{zH6i;Rxr<)Z5=!jwOm
zc0(29lkuDsV76$K!*lyLYgBrkbx@Gt;NN{Rl@+;PYfCkvo9Sp4830@a)5@+=Gw9<<
zyJ;l_m@G<%v{E<*LRu+g-aug@2%EKR4Y_^?OgvdRco$tILHsm{naGk5*-PPZ5=dkM
zp1*18mO_m+CV})(F#b<rp<vXHlwJUMj%dki6=a9mbouJ|tkkqVW<k?Om0XS15N<m2
z6+_@*%yDqLC@I&&JH|ewr1;uj1Hq~s3i9EwCz(W8s`oj96{dqZLh>I74dBiCD#^tT
zl6c4v6&l>~NU6y7%I4?*l;m}?B&xv(UTCNkDd9W1l&}Fpa0^N<ww#A4UoEE>v~5>g
z1mV6X&W7MEmlC)_p^dDbkUn<OS2)z-IK7~<zk~=H^$`lm2Fv9TL%wyCoCZ-pP+-`?
z5E|w#dalq)EMX|oUmA|l;V8k%z_p<yI4l@D&pL3L(F}3KY;1}%TvSSi2qVv08XGjC
zk5P+u8X>iaKd40rw{wVNTt<}cd7lL@<sMkU_u8(caO9Lwq^^~j@m7MLp1c(Qy00JE
zKjWHO`}0vlXuD|qEe3~UM%UigAYv2Fh{J2MOk*WI*2tC}MD@9ElP7lMHn3Hn)}`f{
z9Ea}hLi%M$hI$(*18j(_D(Ma50}Ka0TqY}*Rb>(HPeqw{TjCGPDn0FjcLnOi-7^FV
z6u&(qR^2SKv_jTR`d41Qd3{c;E-R{@ieHreQ|$b`-f{gry{;DLcAjatqr&W)0@v3^
zd=+~YYGtElk@S-n5Bbq4{XL@E#PFJ8k~o5pO<LL)VFNSV%;#YGcVF6xR*o!u;cbEJ
z=4;fr7G$a&Oa3-J@}nX%b)Jo@j+ij3^)*%Kfw<F;4@(MP4Q3am>t5z6a4w^mQ<|gY
z6=BxwPQK-IKZb)Jp47LFhYGcd%8@p5<G7|8v)2|dtgvf>Fk3LK^|^x}2%%_OE_R28
zEV8Vz!R(ASao_fGm46mYcAX55?#3L`=q*6w?Joa}f|X!c$)aXC95bau?LHo?Omn%Y
z#k5&>EuegjwEPMYlN9I&P*2U;c>D!?-EQFN<!MPLt5lC7B8yG0QEnDs?~tqITl9}!
z>U0_>3Jd+BC{D@le5d)P&FJ1Q>EH4qUoCTA#tH_T#mK1!n;iOhP14m+?5lj1-wjj{
z`zBnKoHsD=h1yjqZzEl8*ejaSfFH8koa(in!F_s_!0^Kc=a|PR@7Cwxr|dk%>d&06
zah~){TY@@DvxCuOu<1-Z$Kx+fv17iIv-XrzZo7~7CCp`srDOB;`f!sLK5?BUUy-lb
zQ#XW94lI$-y;rL3>_6b)@V^w7*A(SxeaGKjlPJ-lOQX_zj(@fpnsAB+%gk(DN@61A
z%V`l%gCKCL27<%rrU#D*_~p3=41Zs~{vDjck*>nL7yd0U95tTP0U=s4t>kC6|7C_o
z9$ZSw++!qq^9|~W8py7MJzC^2q*9Co<+;oqg!Bkxh-u{Ju}^s2GbMyNhT%|GE^Y+_
zp$)$m{7~{T!7i7%Q;hMqdGTNh#rh+3L?nm(-C_^W8u^gB95n41gcO&krkh4*i1B=!
zgCN)2;(B><unW10a8E|Aj6$j^Sg@{#U~oYp6;(d~GK9WH+P_Jnlx)2pG6!bT_B;|Z
z+Cc9DRcjKZ8$rFYnHr2PRv8vNC^rBKdrS5$Ag}^M5Ew)c#%dZ{Xh}#g#x{n@Nn4R6
z>{}lSr^(*68_nH$4;0H7T97W|n8MGe(qiif39({UNxYFv&I6z^KArpI?Ugl@&jyQ6
zc$*{pz_-Pqg<nZwm8|JO=0BKsc2^zm8xTZ?O1{pzD6#v~oEPH#i`!0Wv7nNKAWcHx
zC48}xii}VF(fZ0gU8yE&x9nd%jH4uS&ovU)yHe_3qc}MS2|w_;hh9Eh44PIU16|-^
z_?MVkP0k|3ozAqJSuI?Pg_YVaDt;DRZH;<yXJwtfn!F@<)3j2)>rjW&JX5-}C3vur
zHEOz@S91XW+_Z;j`(-TjW<=)kjJe-qSTpFre!oQagv67_6E5akI?|Fh@}5gSW#>WH
zz=G5!ch}UB%;U*?eOZ<s5xFzJ3!34{@3rr#KeCt};AV%sbYm?^l`4!ZP)@Y}3)myU
z4~h65r2nE-blNolU(6Z!RkP@Q_Dror*E@gdgzqu$z&62Uogjv{HXx<3EaMH-`u^IV
zBa8Fuq=HTo6@)LY2Tc#E_aYd%2vNMt4aQ*TK3Ku5I0*z-`hU#sP<bpr!YlX9_3%+B
z+W89@qna2qib@tF8EkquDdSI?bHA|nk?Xf{r2c{B+XW+O61x_Tv$4N`kK|!{A||u0
zcUwCIACEcu8=qWUnWUzDxA!|Dh<l9D6E37?aT`nz$8|DO%a_-+jGHy5D^a0yshyPl
zsprr9>)AvjhZdLpQ}2~Jf!T&pdQLJ8-|G7SS0}H#7X;2?R}TsOFA}ClGro_4mC3YY
z?&tF^j$2ZhUcky&p0NXj`vr2-7C6mgBrEuTWF{o?LWGkQhg1B&;5=+q$TB&LkMZMG
z9g{_GGn(Xv<VLGx3@-I^5<h+s_I6E-_#D@hPiO@k%zD^3QoBdC?BSUDKJ$G+TQL~T
zZ0~Ot*@I%q38e)w+jk?WP$qFuyvc~7TaDYi$4dLqcGa6ansBfjMZ3l&N}~x#A=?a1
zDLB)I)|dPRFyR$z)R$n@Z+{+1HYJK1ZnTsL3s$7AhVrBIa@vyQKyhiN)4`l<*~;@$
z9$y+ZJyB-@o|6bx`BmLewD?o}GCBp3w_sPe{&Z6@7=kna036Vf3}_Df_mjDKL~84f
zPwU*UD{7DM0AbKapo2WP@n-9-dkI>B;iH&I7Z?zv3z~K^N@|(h_g;9AlTC-nKoslX
zt`Mw1her@mxb`yMi@Usu%$h_HsE1Nui155AJs=@tq~oQU_}X)1>Z^gns=y%y6w4-p
z^s|o{Z{8|zK3%5FB0nR2H3&g<7l#mH*-$tPbKW;R3=A&l761ATEsvC(BGDFDUtfjb
z<Hbb@bq8B)GI2c)fW|~zX=phCN5^&z4up5Z=GvVZe>>#Zux+@t20GprhMx&N6IEyM
zeeZJ5c&RX`lpe?YN?eKaO-d1_#NmQ8PD~fSjfcVWpcD8t2WKS*{zG2mJ<+*9PONC`
z=BR?%wjyP8;NiWcZrsQ&QVrp5yfq_rzPkO<;O3cckyRe2u7C3X0{Bf3r~12L7m@zN
zv9Aa#RC%P+w-)wZeW#=|D;07`m<DKZ8&}5}aD7g}jk(J-{ZP!_Ss}|h>Dy+(s>QNq
zDLvkA{C+ixrQV^DyV73llk^AWBJD>P9FZR(dm0toVGmfh<$IU>7{^$eNtenJ_OnCu
z={ufNV@ggA7RYi0*w4vEJa@zLY*jZu&v{6F|1QJxBj5fH<x_v{N}FQyW562-;j^r>
z)mI8HY_V1z$CVihLJT$`jvlx{DzdkZfOy2qP=nb=W{ev3&VK=D7+4v$x(<YEwygk~
zFlJHxAVW%tDShGA+Ank|j!O{3q`!bHQW$~phWpnMv18#6Q56Q?NTC<LjWtdI9!%pD
zEq9fVAEY*1%Qj`i``MTNteB^LxqZm_ls#IRBO|)$X62J|jDct5@%q)7?2^v1Qk;|5
z3}g4x5PbYF4dMr#O+oA(@+q@-CC_9xWDcE$4qisjw)Z`a-*NCJ`%Ear#o{$al<n`d
z58@U+-nx<_xV07%>;wbx0>h-@T_@G-V1hv}-Bb&RcUzO$;)Rgb@dVFds~Mj}*M|E)
z*p`ld7Uud{`RRA389H`fXk%hYtTV~Oaxucskk;x>Eo4VYw0@qsE%=B{I3~AO5_TiA
zUh%thN6>3^Ie%->tEi^$e6%rmNrGMQFW?yG;VVaLG2GN1N|#&A_;!|}(Btff<7H7V
z*K9tmP<t$p2KSigb_$7OD2Eox`y}f)Z&xRMu<Luhh(L$=fEX+{9fTFc4^I&^Y9MB?
z-z0*42wF2E7@B=Wg@=A~l{9s^q3DM#!U{#O>C8b8k%(y}Say12YXrm*5@4M7V3M`H
z=%9yEHV)Hl&uwr>MoYaYT4dj2hyKLZVH@eegzUwe=fCysd8oM+Z`%w0IMqs#`K(My
zHB$o_Rx$&mfP5eZE6m(<qobcf<4Zii`ZHMBuD~z{0l}yiArzi3NI5Xk&r;t7Xn8bu
zKv@OrV1g13_e>sTAez@Jh@>Hr2y(sdxH+?oAbd&bnZq4mas=ohPb=fFNB|Qd7aHzz
zX>N=l&Il5&gAh>afV=ql$cPzXuFjYe5LgQ}eQoZkE*Woy1WPeAm4l(_F80K14f)of
zkm7p}wb+!>s*fk_%UKhg%@ckxEWs6YR#(#c-GcJXNHpfe^%-*QJs$I^qbcLkrdV-D
z#3lH*3v9w42@6-Nk&HEVi^y4-PJB5W)W)zIi&+e{S&>Ax4bq5&uzXs#e;!gUL1FiX
zO{}}7okpI=wrHmHNM(MZKjHfyE@JJR2Z}Fp$en6uiDTCe?_{CG{Zl=HiOMExc>P0&
z?uaE~8N&^ny3FaaaF(t4@Nm(KE|cFQbe>JyEBG(oyOyX_IpvTSnufQl5+?OsP~K#y
z6-$Q46sYPxEnmt?b=E(4W0;oK7}g`JHll=XIJ%hL3djmm+>&ZN3V305PNoi{d`@Cp
zmtikTt{;@nC)m4e^dYRGYA(5DH@86TOgN(JQ}N1#>_!6b&j^h3I+4HQ^5!tnJiQ)k
zn3n?yf4xF2jrs+YAb$ED`)7yDGyJcXbT2f^rHDW4!i5un^9d<DdcErzOqSnQ;-^Dj
z=~b|Un+2bg{{DH=GB;$P`t6B@_<%6-asbl<nvoy|OP}6h<;GPvW=bT0_Em00Wrg33
z+4PWs-1V7r5K;fa-YSsdo`eoqr96>Iy;#>aypF=!)a!db9?W-Gbz&QGbFX`JLj+Sm
z=(>~f;R%Nzd}=2h!3sR)gY6t~!{5yMmXnsr#al8^)}2e%hd&wj2n4U$$~NgX976Y0
z1eS4YK3?v<NyaohO7eA2oOP3Wm`0N!WRmGJ_S=Lp=fdNqb|L!=m%DqMO2QdXg2Ng(
zlcC=pTsm87-%9NkjPzEy*_};WmMVM*T}b^f_L$GTZund~1}Y@H%<-t^G`OmX?9%{@
zbYP1JPii#<@U7n_6`*EhD9Q6gF~n8D2$TCTd9c3`=QB+_qs1T#{uW1xzJr(UdctZk
zUjp~qcAG4Z9zpe&vEOft9xg34thw4Gnok=NRXG_Q9`kWBkW(8aMhnF%l6i!X^7`!u
zm{*4xFsL=X8*o(TRkTy;k2g3#)MT?dU3~o<cBlCnKB_B%QyA5fX3xdc&W{8Vpwtyi
z>{=r$6w|^?acx^Y0r97FFD=OZ)m$&wW>+<-nVH^Hf|X{1!0D!seR5GON!ZMdWF(`#
z2@rm1hSJS6IYrni=jNgq!rp;67LPz+1By?JShXP+H6C|6iGUhm{d!{3_y<J>Km`$h
zDIf*s>36IvP!j9PBbTnpRb)~xY0j0apnh`+R$w{4)j*JrS7SXjs@<4Kd=T=W8!gTq
z1y-zo2SP$Au~VhSC%1D*ai$TLTK5{ccoxhMs4X`6ZL`@8EyI>~(5nJLF?`)TJjB4Q
z#W^VW|3IhP3ytP_2${tvhy0B;Wj7zAMPHj~q3S@0g=bWTdBnjBjFV_4(o#kN)Wd8l
zg~t6C>cw!gG-kqaDoIqr1G4>IMrg1(gfeL*<1jQ;isumH1hayZR>-Fm1DTO4eaO|@
zEglq9#rrQH3I>+qqH|^P5XA~OJWLs66q7D`K6}e-qo!p}`r5|Lda)1oGidI@@l9Lo
z?=ShhOY}NcekuYxRWxvwmM?)|J5Kmp*s9&rkLZ>{h#iYXw;(l>iONqvwl<VL^&2LK
zhMXg~r1NvmSBf)|@*mGC@h2)-tegq~a{c6tgQ%~vv*|ykk*z;?Vc(w+j2xz*_)-&#
zW);t$d%R`Y#1CJ-U^?_$Ti9k3dT<@{3eF`PBv#<cpmOL`tL>>BdRH@LX=c^^r_GJ9
zm%q%n_S@`MJF(-pl8eSW3E>~!ceP7+(9LkO=#@!@*iqt^4wd*>yB0-X{K)8zgg5fi
zllsSQx!eAinN?lmL2*FYv*m87+pfK*U!_~r8%unUBHWw{G1je-tMyZ!Rir+JcWPN=
zrfB{1TBR&a&Au7V-Kw&2rIqZ{#?wDrF+tsyp+<W)1<K^DJETvg(-%!H^i8XFu4LbL
zKhW1)ozBRW%L7j6fgFJH9Y7%yg!5ehEVtr=!DbW325JQVPpT^r9$ygvb;dgNPUFgN
z7VEr;^sJ@J&4ADU0(>@NZw8D#rR2lr-tyx1SKrwDNNsgGBx^m4ehRl$svb9FC$+$_
z59Vv|SpL{`BTF3R{BREB?ErvGQ%w?Iig@r?jGWj~k6GW1AqPfV=9^?|%N+e{nI8%9
zJ<xXANcnZ(_ZRT8@76SYvQS(w(c!7|ThbHJy=@l1wuQEY;&YylV3(>1o9MbLM$sj~
zpDv!3ZpL&3IJ?#MJs*1zM^ufwbp<3s5%^Ox*S{$N%3w}j`mS{W0wIn&lBmQBVF8~c
z^iifP*L5J-we7lR(ZHPUvAfYi$8zu|H=)3jTKgYC4A&`o$%&aHF*X-8$w$StwLp%7
zTG7?IFx3kX*?mI=+L!%Pkt>p29_|BkPAcSd)XqN?fC`A!4Sz6GRZDdemC-wQTe;?t
zT9E!-)ZN^#^ex6{Tnkh=Khe;aK{%`-zy}(Uw-toPtvmqKwvKb@0|J?odpsKI7@B}?
z?@O+_Bd{x8@WxD1Y##5!Opy5WWh(<pFB5DqbFdi%uP+(+E{RqZb>0Xf!hgoChh5`Q
z*gfU&Di<}F36`D-f`}N4^B^phKpgg0lISdMG*}U9e_O3!^#*t&2yt-Zdk?Ho{}l*V
z>@0;Q7GpPzA!w-O01(c<fx61lkQ6DIB2r%+0V}eqF#;a#huFfTnOq~#Sx?{b3rWq;
zn8_3xfk?D@!mMi~CD~+{H!(m>Cyn7|4Y{&B8h;+P;UrN8E$Nzeo;)!1n+C?rRmKIy
zA*oww{6-~fP%gtmO(7|6Vm5{$H7V6mgp|C@O2)9z^)&}_Qbp)}-*q>d+{K%69cL9x
z_@MgF$z^yR#Ap@hhr$##nSB}Xem9r7mNOUe7jQ4?7SNe%%Q`1PTku=fsC?i4t>6s3
z_PJjB)18ltCLG^^07kXAIo_9dpWy>fTe71j?ACJutzs&ENv_8v-p*fhYc674bNI@1
z<|E1jo~;uxlD_k67m`W!qjsw1cYV-Gb=5BD-DQy&X0JJ0ypbx_f7_*Z*F-22#(No?
z&vB)@jSEJu)N4P<Z>?5t@eMirk=W$9n*0|4<9_a4@nF=3xj8*QlP58q#faRga44FQ
z`0)Ppt#~xwpG)xc1s!kgycDIT=U(B#aM;ey-utoF;V{P7X6dfEL-e{rL3Hs*FTrcZ
zt=m9fdkY+Vxwtfb2lH}~v_zn-k?GeQO@@UMG&1um>M*$%YE0AI<cyB5QK;Lu8OlIX
zwn0R>5~r^%<v7_}wLW%rV(r*SuhYKh$)u{70G{hObglDx5&pg|t+{4=(eTuW?@F3i
z5!br9=BxW%<xg%R(?=g09<ME);;RoE_7cy1ZV~vjiM#vpnAFHLC1lmhdF$2BmSfm~
z-Ob3(Rf@lME{TIjeLGFZqKy89=<p940IaoMkLo(Kkz;=1SxE0ii^!tWm6g5w_1pcQ
zB_0#Ek76MK9m8k!Gt@DE0ipdiKOCq!Cno84fn>Y5>G_ZfWv*|{hpP&IOyqYOM_kl6
zxD>59mIGDAz^ZWz7M8uz;Lwhu59ESu1dl^9>gE<c7{+S%HUtX@Ke#Q-6ME%4IGphP
z>PE=+BsQP#jI$!FN-b%q{70NxZbQ%;HwK$E9H9gkO^)(UU8upfUa*K;d-tSj6_Hbc
z->vXq?VLsi)|d~Ih(b=`RcBMdMe_Ggue;w*U;YKik!UYIR`EB{*qT?Csgj6#7Op(K
zi76TpRGw<2Rh_BR%Vd**=+88?n)kE78#)|wQ)?70)eqT7W@ecWV(X~&rfk*jfp+yA
z&roP?dt<>~G&0<;<_DVjz=+5%k4%l(+!dX{D3WKXMxv4E@@`Ha+F?5gLM|RjO+i4<
zzJaCcj<D*1eK%T^D+ofZfv{9tfF`K^c12+A^sOL>;0@Ms&!h%I8UUdyj%FyioHih|
zs`W$>fQ40rIIYmCgH5#AKP{h^CT|fUN|a}RdROx>VeNDOdM_-aBmfHO@*=X90X=bf
z2e+f}=tdKXX*8i^YjX35LoV~|&<med^j}#q*so~Bx8jgJe+p~QgW7I>W|IMH5m*Es
ziFOS#raMZfJcWiT_O+WuqM$kiTx^CgNcbu_wE0xMDOpxRuunah`-&6q$i<u3#(arw
ztD-Tyg=Z}Vnb>aMF^cC_IomhwDcz9RvViktr+kb#FY%|4I-loA2y;%v^4p*9M3ecS
zI|*MIUc)_^R~IjE*7T;_@x}}H9u5KABD*hQ7d1F5;4|+DEyN-OxP|$~-_Pd|?U9k*
zI`_>uMiYGhkW~8R<+8y^NOkSqP45|gWaIK~N{IwcpUa9wjR^hoq7La9uZ!kcJW7~B
zXyey;8v4;^YKJ=hzqT6^0e8ENd8dI_ZLXIkO>Qy{!VSE@`fGbS_mY^tGHpv$N8#et
zaHi|sg26aXMUjjvL70Og@@~4y)9VRfj<(~3?$vnnTk8V>PkVdZxH#eqq=QZwg;y^|
z!VaqbgkVFEP!p_*x4Hl9X^W+&krdcL_|`j#bHqqdx~XsE{&?wRF6FV5d;V)0t{ly^
zSNTYXW^M+TlT{Lk-TFHaBa>gpJ~Ro@3F4IfD8FoP$I;BzNIZK@*-R03%whG)cE^d`
z(p>`fTlW`{yYLB(<j=gBqC1X1r1@iNg6TSg%ho@}*9u7Swl+=Q48*9CdK&UO2nUAp
z3rufVWlkzEf^iiYj;1`gPUcFxzjE<0$X1!z4EoEZb4DG++P*e@?HQ5A>OU1eDmXL$
z=RM{Q=F{F7UldU~k~d8(T;<`L`ik~0)U3Qk%EyzUHH{Juap{H=w4?5G!)LOE$~r!p
zR@Uf|lL&b9ef{~_|JilpLt<;fsBf9y#_huaDWdTtyE1CsV_Wcwu+o{Vsz4WHAOSCH
z@^XAb33<JzeHc(NP;NJh`+<#d$A-&As0ne&6kPYTKSvvWP3ZVQxxp#&Yep1BZGqbp
z@&e~X1L;S8F?MU4^+QC$pIk)?va0jgZ@je9ZUlF;uH-A~UeCz)+s=)gaO8hf;S?b)
zGFZJ9sbC&d{a}uad&a6g>{ViN<vwFJGWkhb#-k+=`Ky7*CeokDZd<dgHE)VuIdWBB
zKuyUsHk9_EW|>?aEv`o^H?D?k9ueaTKcoGPFKSIAnG@@HClU$ZnT*69(6e+YQRKrW
zIS#_eD)6plzcf^bW`+l2kLI(Fs9l*|HTm;A)bj`5ugigPICmI~ws`L}V#<o*Jf+Iv
z9(nO>(v1L459EOncx91@`WgK_d_oEiY-EO7P74sOAie+%Dd{72KxddE7#E;nB|x~-
znIt+|qWa%90e7NDgB5}x4|x!bYh3$KjO`a7^zzmWC0hi9db}IIkxz#c2iP&5rb`hF
z!sbMK<uZ{(Y0`q5QtkOOR^(LR2GB7N##V)MG9$G+DbMI@5I++=>0K@gCtx!7UNJEP
z)lKex1RfwHyhC4cO#_TTfWweM!a+Eo5PsN;PB554HKMKagNn?eA|T|=pM6dUlFfO2
z9lXFIf_0uUkLZ*;HTJ~m&ZAV@6UV)JW?E;}jdDI6c<z!Z6hwRoFT{D!);+XjBgrGY
z@JaKP>km)1>HKrqLccAh-z`e-4cz+fCBI@qM+mBtdwi1K?Y{YtI@)0}M0d4T=zjbl
zdQhiN=OpIDb&G2f$pbap<znt9Uy1E6il1c@tZ;To#;IkUXA}g2zHhQ88bDYbZqjwq
z<f)W0dm_CY$R51-WikG6b~W@{_7f>Vpm&qh_trgqo%?r^7Lx5fGuMLb;2mOq>mo>x
zQ2UJA_I_&16}>v?xltH$EByIgyJ!>zsm$0_y^HHxp5QF5D-ByAjb}F9ckeJM4rR+j
zB|%NQAZPpCkqJJr=lQ}ZOSjS|pJ=Zzdd~;+8yfMi*MwO6p*9w%{Risxo+30Me#bsp
zuVdn>_|tBEbjvmccT%1KoH|8h&0SNT4|2}Bvug%0n#DOfH8uZ)S7^UX)BH%y0v7DL
zBBV|w6;gX|KX$?|0xCOkeGATFwaxu>f#Y7v_r=eb#)mbutqbj;xdhE~Ml)VD^@dLY
zDK&}wh6-eqwvw5=DMP6#G&_FN=AsPo3igA*r4*iIY|sX}`47%s8)F<6ruuK!L<fGa
zPsK_M7_(bqs4}AxJBFJE9L$Wa@jflNUI1RKRKw4wweK(f6!GL_MQRooHAn^VdS)@5
zdAxH`D{UCuJb3=pf%upKKn+SI*~_JzTl@WmQl!Yxr0Tg9GQn5{l5Mn{(xS{H!0zyQ
zWb080RsIJ6&5NfKl(^2qPW{=ZMu$1Y907OWl6Nc_YUwufJrfMzC8Lckk_o282LW_H
z*E|lTJ<mOx0Q*@w`w1mC*;ILqYKjgHDtuqYTz^f!@%YJIAG&oC!=1krTl(w1@V8-m
zp>W)GoH*liMMjdk2v>YEy*YR4`7L_ei0JQm;U^6M78&UXK>SM*F3Ihps{`kPb8`Bq
zwKM&@FFq${zuN+g@>d0)MLyYFVJGNGfzmovJCR9l0H$?c$2!S-d4|~-*#QkDQe2=a
z=bBUrZl*Twc%Sjj1<toyXh@@D$RO8mlZBsGC{-DYGnoB|PR@AvW5pAMVb?j-b}~$l
zifL|M4zSGwTpEE4KN2L^Qg-i_=N}lqpz@B)qm(}kMd!U-#hyF9c(E8n&0&AqjjrGf
zg6L5;f|wV*_4K;YZ49z`HWT3PtRU&>t%HdHAZfDl^tGAdWA)V_-Z>w`AaXogC6KAL
z#RN<;Yue+OAqZD?i_8$jP5u@vf*_rk3Zh)a>O5$xBn|@*jvFTW>2s`)a-EnnF3ubX
z9SX3+?m?uarP7Gg6HGbW>Ba;<T)Ga#0ezbcLK*>~-)<z)k417(pW1+Muo;RxWDfhy
zfRc@42`M=#1TR{QBZ6p<1|CcxMhyj@wgkAHjD#^VIy$b3l#7a_pGhYSHxz&%$aj?P
zH?3sbjzrkUk}Pk+fWwJ(IN&=5*-4W85PM~-m5>3TD>RdoO+kbKD(N9P1qgXuf0xIS
z5x2LAj^jg1TX?S?1g`7!1P4|nQ0yC_1hBM(@2ju~EpkFRy;K4AuCJsO9=l~)+T<=_
z6p8-kZEK9fxJ%Oa{9edcguj%0<=$;4n%Q*#pnUpm(i>w}-(XEC(#@aX3;0^X=Yi{%
z@x_+Da_d3s4v%8X_f+$wi#01B$)~qp?w-G1HO2;kNrRU@%DMh@zsrqN$?TRdmx1Hp
z<r>bLu+MW<@AB+$)_jNYej6RMPen>Kv%eS616F;7XU=-n^J*LuM|!>CQ&VoVZacT1
zc#FKU%)b98Om(y`aD!G(vyueNXod4~Xw6fraw^<MKoK7&tHf2dT0MKR{t>r-f%b<w
zTcIcOx2I2Re`f|{{aK^>HWEf@^kl&i2he95PqRZK$>)3=7@5kZJwl708~;Wx-hR&)
z*yrKBcx<s!ni{gyggd`xW+LJBhg(IRT<N1yKwJ0&E=INM1j`qst)-jmim~_5nM-N7
zFO7=7wD;86+PC&aw=W$soYe&g8Qm2WiKV8SPawy0Q$v~=Bxm|z0fgJ<hxZxyiS&a?
z&z{_l21oPv@m(2i*OTOq{OAqHrK2(Q!P)#X%OT_yW)qOGdlwxa-l3^|1_;;``Yf;_
zp`&lC(U}DCP~fI~oJS6+uGIVdj4pM`SS!w35|adM+XTX=I;o&Z(Wgn^_G<URhA+IN
ztkmu<%e+G2sqy9$+Hx;`eWrVRUv;4qp_?>HDV)@^toRJIs@M3Sk1W)3g@yW>eqEr(
zHH$axOUd4zb*i&&Q_j52MQx-b(d%PU<xWns{h3k{<7hj)fo@z}=T~;Dg0;SwKA7(4
z1Le#EAV~wCt>TkHGPn9_>Vhpm)3)B=yEji>j*t6u{jB|Q@xnV(|C`-cC)J=z(Sh|}
z45LC98EuQP2CweFG*_AR<)Tr|_+`BM(}V5{b8zpaAt6l{qT!gJ^i#b<8{#`_ME4JO
zb)VNZZ04t-v*|`3Dws3Q+gQgsP!Ux|u+?#?v%v3v0Zl!tq#ugxW)-PZ{{qT)Vp23O
z>SFEqEy#>K!#l(9xXAKdm*q+18s%o}?{h8Vl16Y&Bl_A`D>sKp%S?+NjUU)FG0ODE
zkJ`PK5<zT<IA(rS2SgQym}fb4rCZ^ibKTa|-aFvyq^!4dSt9|^4e!5GFWsFV-|`F7
z7R3a}-QfSZY2|ONjEOqAPN0VIck(cKsM33Zm=8R$o{PuLazuHya?@L3Qw#`>jJ7Ec
z*h|(atPO`Lhe4_tL-Hvg&2dvi^R7YnWDbMM8Jha$fUh$u?IqL4;+Q;-xl&<5m?M!~
znKd{u7w|WmDY*TE(Mcgc(&(f+X$|hNp=z;K3azybT`<gJnV1e7A5v_Z4^c)V4*LBy
zLZxN~;E@Q1jI(=8l?ZSstz5VQ;yUl4MxqrBf^a;Q69*z_LevAc;;}<VnxP)XVYMjw
z*C1rG*jopzj|F>H!~M_8($`_8wvQCoKqg2zK8C05hm_#JH3X!2gRx;8jj-qhDc(({
zJ;)PC3js;z(^stuR`fc^UJ(B9J~JTW+9wbG{=E1{x0H}vDH2*mA=R7&+&b;f#)e3?
zJB`CRh>QEz)l60=@1b*{+gSUn5N+N%8w6$-PB{x)FyOHnp8fS}Qb0W(2up(C+pIJ^
z&Ea)31MXas-f7=TN2OOX!6N(l<D?Bi@(qB7QN@U8$q4Gl-@I2Q3uT61iQFNxTJ89$
z5A*ol@s~Nigsz_XnC~~>!_qmBiA7Z`VjSh^9~F^Q*{EmF*{d|)d;WSE!D92Gqbx$i
zJyGt()9)F)d;2X<JELqg`0pGB0XA;L`Zz4C=W3lsN-Y{<+9hpu%Vm+DcB93*+aLXw
z#t~~JEA?l+k}FtLt8mkG{JuHsOs0X#|J0c_rt%l?0#?$_VwL==E<j8#<<}~$!xQL|
zn$wd^SV7cNBU`(->BP>hwX>0Zt1En5{MSx9C#nug-zxj+zSmf4r+5wyCpw}dW1}SY
zy1w4u*Db!~=r<pYn@X#(9$UC+u#*`|PRu{nj6F7yD-Gf!V{|zYjosai7w4Jz_yZGJ
z<rU551@p*|aily^@UbPZc)VSft!56(>l*pA!!8iyOc(FR*WWPSc((lNv~D=YIDdb;
zJnssHaycUTrNPW-?wKysubJDz3|`_vJfd0mAQVcIsrQErC%KYxyAR+MACM2bBQd&x
z12=GI-%n)qKW^)9-ELzXIxhUeDsJR6`@UGda{OR&K~{Q1Gz#DlV8ID*I5!e_)HSuP
zmT1{LAHnz~(Y+98HL+jStiZoR-A{ggt@!A0Y((&(xgTcf`9=A>e~a~LBtKtHNBOeR
z6DAGhp2whLf@r>+$uQA-$qi8vx$BXG$>mlKw_V)%+U!VQWv{Z7RaUA-^|3_RuYNx|
z>5FJox&0!?+U4qZorZ1V=EFhsC0C+l{3O>hoAP$L=@@b)Q>PXRm9kihf@w$B3^fz4
zwl`%aoomXMtra>232FCwI7J#9W}c*HUE%-qjx>(*t1FC8$Dsi2OVyG#hAA%lv72ap
zJTLOIu8QhT>Yk^wu~pAMvpAMfC+YiLK)PR~yGQu~^lhh5hhaysChZ~ZGf=c_<&0hK
zsN`pcI+-JKsKN@38xJYQnV(GX*}^<r&ofX)ual>HH=QSJ4#YITt+W8BBdrjWmKjo-
z;l2l9UOeeeAE;rR!m*j~0yW{VPef0$)LQqg#W7Gl@<Iib!v`w-SCCh&ELdaeHIhm)
zI_}%Gc|Ufkb=dX-q5O*VA6nsdl9v5~l#t~qun9y=;^lsyjH#BgdcOw3FXW3l_Sl#Q
zfen#^#;GQO6%mJQ8iderhR8hPzQf@D5vXHm^gPGr0j3qs$sI+YoJY8ps%ejs?cnLf
zYCpAX(mWU(u<3^?Y!KKRHfX3c)<28Y)+l_cu$N^Zv@!_nb=hMzc^o{<5O-ID0EdJU
zA#l1|@0oyQ#58`OoTY82Fy#5<ZuDqq`u%Kd95l<egbor3J$r4z(!0L|<p?HLg5G~}
z*&PBLZo&VJ-Tf&_<A6BCFgq1G`}&Z9Hu1B2E^IQQk!?@Or)(M)1h0q%WImQ5n#Jg2
zy80eC`@X&ZhN_Wy^a+!k5dUC}Q+2k(Ra)Tvph=NQM`O$0x8~C2QI}t=XLa5Kf7x&U
zE}mIFdd)cCojsE46k@0xtNE>|^3bb>QgBBxP`qeuDp2^vyoA)Q<2~i*q)yfp58!-G
zzTmFn5oo61mt`nLqMZf%W>&p{#3uRKdok%~LypSz{_kl_``M{PVs(=*<U{_z%<IE^
za*sl6vW>PSZn91Uh|(U`tJ)ljJ9lOL1*915YcjLg6x!_US=mIi4AN3k`*t?Q&Z|3m
zx<8sPwx4?tR7(SGsU0soz4c;3rP4U%ow!s=)|qmX{duJnO)M@a@?LNJITV@UN!b*R
zv8ghwba&A5umS9g98F*&ClgOpJ0i;7#{t=;k44o9lA0k?4^`K_pWW<9RrY-yvX47e
zU=0w);qi_<no}zi;IR~l1jm1V;Z|F^H^MK8Q3U^Kj;KZ9NT$3yDzv?{lyZ}s*=oxK
zv%o8Z<aMgV2iAf2-Y}#(=SWpeBp*Lu&Ab(ZjaOjU@3XtvOW?5(Us3n8(DSsghwn&h
z73rkrjJz?7<rplkPFu2dAtZa=H$|tFW3NE2M5^KGrG?`PXEiZ9$KT7oJBN||h<!w1
zFSJ19*SpeeQKek0_(Z0>la23KsQ1_U)~?qz^l~tio(5eUfD1pJo_avHnG|3BoIeif
z&T@;V*GLaNJK#4;_qO})8}gyY(UIqKgTZTEQkQ2Vmle+1LtQL}-Ypzm@~^|{zUv8`
z_AJ<F_dau;G&%v*4>WPEG=AY^$5=dWxmvD8texYwUwSibePEow^IoNx^D7-b{3w`S
z{gMB$vd_HqD+$T&!tKDJzuLl%)<<Gt&eEeG9xlsFN0C+nT|uo1eVbyjOS<2`K-rP1
zx>hzNf+h80a>qjTwSt%3_B7d^cho<()Crer6W6K+d75kG>&aM<tCNsX$!VLx?W2xM
zkXHBe-qvEbshWD1!|`9E`gW~2jZNUY+GYydckBXPB}x1rc{Bl>2UO%TO><*rN|FZC
zK>}Iym>OmwgPN{eGr7b^dOaJF_mAlfY30g_>~Fk#QCD#Y_Rq~#8$l2mpdiJKrZp!}
zFB4|vfIY+QX-7YhVmL&ECI?f<JZmP6h~|iQ!upw++^ska2Krhb#yd?r22ZiosaN%Z
z5RB_lNDgw-T~`E?invS=__fp_609)gGN%B`R}C8I141Kex3D2joIBvxrksBPMI>OE
z^?_%z6ezj4v(rDb|KdkhFY_GG;`BdT!3_u|8MDYdR)qjl`XU@Sn$H1B*$Ri$Hlg9*
zb_8xc07ey~Mj7s}kx$&~W<huzIrCjIp?i4~yY)z-#U&Plz<NfG|Bjl#ctGf=5{SV*
zR__M;=YWqua!Z^aSRZx30tb2_n9&i9OC^8}V?rLc-tW*bwu=R=%17ogve@h|f<gqi
zZ9N`p);U|SwM&$PHidXbz8fuq(G5JoQv2a0b37lON<KAcyz#{Rv`qUoK~r*l5ml1<
zHO9tlWfbtRKZv&>j=^XAvsBk_o~Mm&&8d5!h0Bg#a8Xgd>1sYea4%jOci12Af!1T-
zdod=0>K~0FxT(Kr+*>~=j&ldN*|M>`?k9_Y3K|S}lpP8v;}a!&ZN=v}TNp=$)&HC~
zoqq9pTK0#Di;iB2{`{7DMHn0$s1mp_18Rt92*#Q0KLB03<B5?QXzU*iPIHjyw>$Bc
zqaWouY-YUYndr~+p`3@=pS(=9x3G^RyCCy(2G3GKoB5#|w_NyQvCaKA#*V>H+`Qh`
zo2hsbr5D_8YM0OOm$!@Xd))S=ZErYk{+hxQnqj0HjAz;Srp^_n;h!jPHPXhsw3eoK
zfTDB4z*`G}dk@(Es1Qh0`2^VIFRga+UY5x{hw)k1m=Qg6*_krn_;$KEF_%B(GyNCP
zF<A@0S461hY#oW0j!yde*!4`vm+wWL=5&4C7Taw2kl6{tIa4o5Wu9tDUHO<hg!Xvc
zU%QUT=Xoo4dHO2hI34Gt38ZT}Jj)8nZ-TLUVa(4KJh!$0D`#uE-S{HIXw<pyRY%dW
z)Y3kWQ&syoeM~i}Lp<*)`b!{Vw)kKQY57U~clJS8%UM#-!bFR`l8{AO{?``+(zT=N
zirV_AjEHaZt|vxMEs03yXZvPLOylMPh35goEW{!t-;qWZ93LH02%|iO=<VW!8X$0_
zTf|{2Vb&APq~wz6LyuwI<P5Tb-KgV~bVk09MIS3<y0mV@VNELHRZi5GkFGJ%dDE(0
zZ45LAc|UwPz9e^?l#YC8P-qXQzTFsTmph2McFr=twRyncxpAue?CbeCSFEC`<x8b!
z2Y&%=F}fJHPZ~Y_O=f%Im5duH6oEs3xGXsgJj_cN8(Cb082DQ~p%CtzP4S11&gV;(
zAEc>R`w6fNrGMj3_~UK2I_Fi=bZ(d3Kr*OPI=XOB^3$Lrq{B?Pn<1EEKIj-kD@t{`
z!uE2{JA>#`#bE*&-<Dc5`<P<hlRQo#9^%3}ipt8&G?PbMrk!t7Bz!t2iaDnWR7YQ~
z%?CPD&v0ftOI$r{Pa_d=H)hxMc~v}Zi&8*L4Hqo91SXb6-#4PbyOUph?EmXnvG*@v
zbBiiS>^VM}P|o>ZfG#@^k5Bmadiqi@EI^kB#((Y4Aba)bwNm><5JKc$2iK1Jhs>vO
zk-(%tX$RJ~9P^}hA{>#v#aG)I$z=HgL?XUrt|ebU_Z%o>E3aAoDIpFAN*Z)U!I0L3
zUZ9Op+fONccxtV|@bmNl<4s0ubDCl2iO3l;d(}rU%51$4dam)zJ}^yAtt~Z&8rH}C
z?;g1{f~Z9DsCGNgy}J|cw`R7w1wDk%x+`})8{)6?rqq}D3mAdHql<jC*KVXBh;o}0
z?$^@$snc+d^=7xr?HX1gObE&HUw{W^ykJtKD%a-@1d|ThdAn(`jXBg2{T?o}3c(|h
z{Jx$pi5B(#{#|tv!4TqEnAu_tL?TkzxqST@P&hyvRK&oyphu>fS|@i2AeE&1PQt9H
z>Kk%XDNW=08*2#If#@^xN1wt2HkozkcE==3TyAh0!!L6O=U9*^pULk@Nftu54xaDz
z9is?>#0Y!SM+(po`<N@*b2^I}#G`wZKHOy3*qPMw@7o6cW+=R}Wc%wKY&75N2y%`V
zmmo+jkbOJt`wvO<d)X$8a3NTA=)O>F36|&kM4g2}U@k&R3AL*c5QBAJ-@#rcSdu|T
zCDH4|HUAm~BlGO425+N*vTYolmHBLv8pA@?T~pwgbEq-xY~9IaPk|(x>L{P`Tf<&8
z1xcQp?<WdSXSC`{rbId8>taT7;k^fs%0ab2=)J-pxjD)qmBVsQN;<W)x|ih<>6ZOa
zM|rlNFBu1q*q_DVeSAa}mXrKj_%~t0h`^&zHYbbW2n!)X7p07wrnCn^<QLVfa;hm6
z&F?0p3bj5gM`=HGH#rP8@w0ZVvA-&7IZlJSH7d(IeSGguNYm@l*i}BKA9<xgPjxuY
z6MdS<$;aB%3C&8LKj+_>9oFZ0+s!l1COd!oa^MpAu;LTc(m<Kjn@!LikXQ%>Q%edq
z^wjOemw?Y-<aDs~i&*@+N}4yHOtRZ8E}k`{JEi~0Hp%Hqu9rs#AoGlj*v8Stc^M<e
zL;pY`_pHwRd7sD6-6e|;42pVA)2vJGm_<JNVcV@|7bo-?hqL-)PGraI9Szpb(NEIG
z#kPk=_=Bo8jRXQn(gavF%_fD7P;Xa7%L}&~3J3UKhR(_2C3jK%3|%@+F!_XyVG>Mv
zz4pyUr}~?K`C>fhNu1x8+UwLr5isrC{EU;F^rg_pF023$+Ae#4AToLduR>zDdHLOJ
zUN@ZbZjp*#Z;LZ+M4}p58t>1&)Z+XMHT&}@OWY`q43qbYqi|%*T4;2v&@J|Dbo=bD
z`mdsoA3uGde`I#@R>9Jdwp%pweyR+7rp|zs)ZqFgw7?gP*X~_xaL34Fv0pRig%Wb}
zx0Hk!!!)Z-x4;N+DJ828H4%bR*}IK&Det9Io&?u&s)*E4N`tPa5k=qN3@P-_&Cplo
zPk3q};!O>{WZC7{WB>~j-L*bNLuFfC#gfy^j0ApdmaGpQ8sw)PW+o*mt-{<(@1DM%
zdl1H1Ak)cC?!+LZD1nxmTsZD%olQNg5@5%_W~xkViV0Sjrod|Gz3B-i>3ImguMQ;Z
z95Y&D3R#`WS31UbuL^m-2Z((%y!+(YLfuEZoSC8?PN0vcEs*|YjBlCT+lGsdfJ0CW
zvZNG~Qe{!jI>OIl$IPqzp2#3=gFE3bAm8ni-Zou;xAojXjB`a~k>U-ZN1-5SD3m`H
z@zXdUZE6}Pam>R><i6>sat`^&oloZp_`ex4gc2H+YBiGdj`)&%!^*Mij?aC1(ES@F
zhJSoYb+Iw1?fW=K@|XP0*b>zV5APTo+R=Gx9NrW|^Tyc+V~=ATJPlpv=%~rc5%XRN
zTXMHz*`fIQT9tL$LFduW;#6t7BeU}L0)pnR-;0nP%WTd`en4d#Pr3^h$V`Y*wEYa)
z&ST73)$`v$TDvR81aHJL-lr})FUU||tk9LCQKl+u#-BP9-u0U)v<$keA2_Y-4_y!o
zCYkU$#|w)O8RE({AH;9lPy9B-*w@a_vNxet;T~yJplyUF6VYlY)JhArD&ix~XJ9u<
z@Yi~Pf1ns{FePJ}gXTq=@B#FvHP>C(xLXy%msU^r*74$z92Ls(Hd!n6<Degp8Rr#G
z_keiQj?60WMRS#Px9xQazODkjBJ@AH+(-tMk2JX%m#rRPa_fi>;U2Ro7+;plz+z%_
z0+A!XsXu#7IB-8zNpFT@_{rz&m!!fh!O?3Ox6gm%IcUuf8~RfG`F{ZIKoY<1=tzMw
z7A?1zr@kB&DYC|&qw3BQTS`F=?cFKK5nx7(1RK6Sm=h2#k#2uWzcqmd##e|x^HhQp
zLr~B7!2lxD0VM+@-RS2LI-y{Mc)9N8xxkH32tVTjSJ`1lHMx6}x&#CQ9Fu-ryP!*W
zNVhaBm%0Sbu!f?E-2zc`i*n|@-S8<PWp?jJJkkWWNwG4xlj?yul82^(xo7ZL5*9<X
za?kF~AWdOJhrheN6a*otE?F<#oEt(A$as2_*ttT2gez9;g)_-}p%Dm4-Ihz~?pP3l
zO{RJHph802o0T*4uP6|Lwyy~HcO(cyQR-3sP$4$S6?0MeZ-EH}nXSKfbO=H*R7|q_
zmINm1IqxLz^uUBQ6-3QZAVO}byCqKj@F5|=Ro<C$^gx8%+f1`hL@B%>0Gmv=m+ORt
z-9yp(;6e(G-C2B65)ept+`UT($OaIAp}ey1h1CcoIVZ=_0$`PyYILmCR);lHnpUV%
zt<Id|t(a<BS~fN}c$_W6=;A(<jGR?RVYr|)z$Mh-vu4bm^N!7Kr*j!Zk^<19O}LB;
zn)3iYOkBiV0nF$*YIIT++(j`zGbHIW2*^qcP06qXOtQ&wXeYAEf49{J9hh&v12su-
z8Gt#9nk2qOfiH2mHUln#5mnWqnJ~y2NPrB0*gE$>wdUq!YiwlXw3MzHtI1x1HMV~j
z2Y(uD$+8P`@$=zd(vDb;*V1VU^rmwOAPn~rF)&2jPT-D0bo(=hsA)e)JNJAE)N<va
z=DkP?qg9!D)Aa-ywyx1u*#|;I{b1#B53T9xhA`NQ;I@myS}hlFldD?eCN?e>xd1>i
z8q{ex7qU-~Ff=4u`+??QsRwCB(<b6f6Vj>qP;-U9if$nLU!@jDV#ue9mS_Mcx_UIW
z8{h~BqT=k%m(YiEAa}o*89sLjkxI(D;<3~nMqVB-{-lUvX^c3hUkcW=wcMZ?+0l4B
z?*{5r9%|>MSdtt$%th_i>D@Nas#E(*X~QrLx+quhm+-Y?twP*h4x==X;6c@{0NC;%
z!Rju!v>F(Ok$ngwaXOFFP;73gcyI}tETobA^sc0#C-r7S{gg;pT8*uPW7>9u(DL^u
zh;L<ydt-SDT~L)X&$=uwG6}L74r!iU+@R=@)>mmGTl#!ad;l)EqG@`t(24=PE*KD(
zX8XM`A!vQ6&jS_dX^qM>oI6#h^?p^r({J}iKM`{sjH67E+-u%PgKhDGui+d201s8)
zjJ3Y=2jEpx3sI=_8a~e7n3T3isQtd8ME9GP7aY^qTfJM63;q#Wg&GtYMyti1$$ZQV
z&WzgW5F9)XfU6a1RI5_0CZAQ%-ev@r^*p+m<hn7Irt2WvH1IiGHa#wfo0*#*ck?6f
zbuWg_ak9|r%@pVZO@h;KNc@xPxAVlGWDi+q7^HJn@4J@5KNm7T2-Dg-D@2r(tc0|@
zq>Q>ZD%Z_}d^%BKjHgM0ogtJCU%QjO=W@9@=A(^Ax@Uj(w}8<_ojz`mM>j;4Qk5dq
zdYn8TZNCT?yUBgR-yE_(4D9%EBT~V)R!)cVXnmME%P_1WwT{lWx!nA98^~Ct%}Rr!
z{XN|1lWvf=;OujEaWtvmyV2ZV7=<^}qfj=PmQ^P&<$k4?(=n0(0EdFZs8gud)M3r*
zQy2Wr*!3;kjavq6RZ_Cu`(41H&rXw2{lEO9g<6#V0Jm_oiMeC;JBLN|*s0QM&kL^j
z-M<1qRkgIoGPk}o<8rELDM%2~OVhb6Qt#x82xt+iT83-OFJv?&t9$cbxl#l#3wa>h
z=<k6G3esiE>dFK!ATMz9_Ca!m?KDKoEd4McVOm8k@Zko_+IR7LIiw^kC?az7Zh;9o
zS|#jF>Z<}13HhxFNtc1NTN@_~-90v&4-u;RK8$j&QP^xXO7z%YloJ4fG1TeJb*8~u
zt7$#ia}XvbC!p3+SxN9WCFc#dNtW2TSO*&GZ%|9le|MB-;5dz5Csw5gP%P}gGnf{c
z+eVO1q5%*M)vQ(xm>O)W+J|N&L=bJSI}aUpPLNGT;%!qtd*fJXq$(PeyK^28RTx_-
zq%QT`3vad^_CmrA?ZUkG@Wgyjhi>kIzuGV%Ae2=%W}EjWP$6MZbKE@r*b#kJ6m#_Q
z=adRvN=9RM7kc_2gR)kc2`g%@@4qND(J8%>5QElT>E8;#ClPcNvp%lqESy;)?3al6
zdzBzdL=#cZguhl$plpGAloXhiSKO5ef$ad7lVW+0!VYVUG=hR04Z_NhWtty+JrSU~
z66vDmikqPEM9M;-<<H}iK!`e&oQJDFqPakXk`&LME=!aMvPM~YcdqUj5qLsU9;MHE
za?2<Mgai&qYMF0AUEb_1lSn+$<W|?0c)}<;B$B2?stfg`5<)J5sH&MLFTr6UA#Oy@
z)!w||LJ6TJ@(vukmINS}sVItv!+aNnT2|88*2wC6NuuL=q`oU+OlZt9>x{W3<lFD+
z-Z!TZhYnsMjNE=R7k~{EDF<er@pXY68YG*8%c{AB!{PCHyH}ByU@k|AK32NajxzkY
z9IVLZ^JS)_A#L{iKZIPv)T@ZXQ?FUMON78OJ@)Zc>oJt8zMCDI3`|GD=t+?;0iyeP
z$7Rg~!AS`-=8z!*#Z?qQ_@Y3BGC@(#K76-8ge6RuJhIQ(0uyx%x8<L~AVLy?7VQ2T
z;YH;L1cIy0Q+IqjAvY%TE#vpXHV}jt13pm{AQO3~4*dJzLIDQt`~EP9x)X96Q+D`C
zjZ{@eP#&7X&LH{KoSS0eA6i&qg;Lh>dM|MHA3g$4!CPdz0a2~p%^v%FQ<;PZ9(};x
z**IUL{{T5hY%qDkQ*(vfGana{@2cfq4kg)k?vLdW^i*7Wk}hFAS92m7#F-MIS~@^>
zp8o)(N%}FHj;PxBs#IBU-A`8-vFVib@mDa+bu2wfmXah5Yy{uNN0OaSH0eqf^8A^8
zbN(@<NzSC<gso!HW1w?J;0*}&hSnm*Q+R8e+tSeLH6Wg6k=Xei5m4M<!RB04aCn5n
zgoZknwZOH*UB6I4P?t1<4xhUnQJI%!pyTB>ZMR@+o7Y)?ho|z2S*K3;%*L=`u!P~s
z7XUUhHv98n0`d%PmPl7{&+TyowE)`W*#7|iiPN7|1|d*0Vr7djXl~VP&xhRU;p6Y*
zolytkCBnI;MrZ_kFKPV52Id~fFi4i%8#d8^Xot*6ABm=U12AI#@dsLN9&}Qvjp^#{
zdNpSq`Z}Nd$gw^e=DQ47(GYFmbonUb0TsiEnFIQ(F^9A2NWOJdBXESMh*_uI0ySN-
zAUVc{i-ICI#nS{zR!g$$VcZl%rKoS>>4GBZ6W*1`+hD3O@|M!9nGW8vQZ%V>61mah
zZNn_S(MOx&4`=p{AfIBJT0C4qG2{sITNpMtVQ92_PbByNW8?DZw!R#oxc)DfX0pp<
z<u;v9lWtr%*>BC8o9_A5-SGYKp+<?Rhz!KE9F1dg>MYW4vy6^eLECj&tw7pteHR}2
zdCj4@%Wc5gW=^}p*_6RNJv`1Gm?ils+rCI}__X{*D8EOWew5+x#p-#m^*tH-8OV7?
ztS6ZEPqfh*wN`)Mu;LqHBQW_m4o1N8H_)qqhWN1(7X)-`eZCkX$eFr;s!)4DzdayA
z6$b-J@pI;9!_+%}WQTbB+@pnTJ8LohowEy#1EWdg;bPZ;Eh6$PyHl!ZmWO*L+{cE*
zA8Dq?p;~UASbnRv+&>{c$pv1YL!4@8T=Cpj0p>hBK&Y{sMy+c6YLzPBhf`z2r+veR
zh#=b2rJ-DXI&}DDT9im4qnjNr0d4PNI^GvW2>4@dX85Hed-@<sLRx~NWtVga5vo!(
zn(oV<_!1`a7If}WO#8R0!-VA#1K7L0FhB?fRuDhqln{vv7H;3@+>qyBiIt5tyW`!S
zaDW@9Wt%v&Mf>62Qq)-u&=%8qW$ai)?3xx%Dr7C>D7?vGAs|>nAvIN8yhR1JLZ8F*
z$!eu;kxiRz7Di%ZCEfzxa3550#7Cl42X>L~1$Cn*5W+Ta`K;C6N7aQ;^Ub%WYxH9q
zfyQi6>cQ>~T3kq+w11fU;X)|nTm2l>BYkAZl!6`q0J;kl$%2xhlKpTX5OqmL@=vdJ
zP<WvY13wSJQUq#YdcslotVF;zO(<H_E+<fPUTX;hk{XVDzjO$+LK+5MFRM5a;E<t`
z^+Di-A{@E>X#y@Rl0lmO^npJ1sS26$-{?pnAi~+WF9N$#q)uYF_H&YK$(c6EGuC&+
ze`ELjB91Q-D7Uh1>dgv9r+wQkD~hNyQZ<g=)?)fE2_$6<O-5AhDVT+H(hp`wT{~!r
z=0MbP&!~M^-J@BUssgh&_z|dCl_{DP4JU2#*<~BecYf&UysmdBNICaIFr3JMutK~l
zN!{p$vLJ4TLA-@mlKLP^L^TEOMCZ>a5psvmdauJh&?W&%vvdjB2^%F#KvhdFNRX|m
zQRs$1Kvt4*BJV{^pxVf~8irf)NGwn~ArSA(Q@eiX1+G&v5Y$ymB>S?35>X`u2QH!e
zpxz2KE3Uki#$INbw<!sSQ4Puri_PxOqG*Qy0NVYi$>QMpGYIB>OGAgBzIf^NT$@&8
z(xFgo@$;hbgMz-dOkBa=htWTkSH)W!RCQ3M7LG-{Ju4Ph1Tt)slJfMd2u;mLJ}*Z+
z^MMN%rg<W8dtgFts7N`FSATp6KqO50{rjLoK>=r$e^yW-1jgYk*Xrj25C}3(<@P{&
zp*K+7{1PH0T?hmN-}>PpHZ&Ac70v`8DsSET;wMxlbutx|q3O#fY?rCN{;FMf^IJ#u
zzqJ_rL9RI3A%y+oOC$K5l-efX-ueqI3>%u<HBvzCknQ#a{Kb@FmuNF$T$i!8z~ian
z_Y?eMLDO8CG_28LdB?ImJ)^dO%*9dnAXKZV(0s}I4fQrvJ)?#AF1QSWek{XVhr?r+
zm+y?+Q$V2l?GY7P*2m#sE%#<Fg?51=qiDNRO~m-R9DV>RE6oJkR~B}koXtPA8kHE@
z^zxG$5JsTvM_c*mRR%rOh%LUv9SG!poV5v?PaVf8F_&u*yEOc;`sx1wJEAn|J;c$j
zA<fn_BWE|E<*QAKli1dQQk}uj@;c=`dhRZ=;u;BCy51<0+cezKxFg;YVGXF)aw<Bh
zJQ8!AnphReqn8^DiCe?#s`wMz6=_u-jZTpniIJCCgTBxQJ(qfv-RTafPEoA*-rLWI
ziBx)BE95*~It4R$)Z{y709qpBZrDPR!wS=e#?`{~1VAx+`-G9?y4NDU9-bg|A5V+|
zu9hZZnXj=kjR=dCN;fIFrb+Sk!b%9SUBPDll#K<(U`jQfBJYVrBcV$cphe@lyiE}$
zw(&0v8>p|J_?d}pPqaFW+?0~JQ=P?qhRsWL+Gx9!yZ3!j<S|Fmc$0$$V)w)d5;FiG
zoz9oC^f+PI;M!heD`t6EnezUDbw(AQo}I+F#-SlCtuiyDFUw3g4a_rZZGdK?R(OlU
zlKRoHhpUP>imsx^REIi4gF~%h-UI_XOJa9f^=Z)TSSq+O!s1BgY~dhUZ0QrC*!3x@
zJxWR1ZGE?8LF;WPi(At^pfGTAYT+&8#o>$3sY0zWOBNkUVlzO3E(mxfxq;KY<pou1
zD^jW2VTh?8hUPZ~qz6IX$KMJ<g*`U@y)v_J%d)~c#Ms*0>t*wEBA$cG>RX08(x~@?
zPOq%pb`G5!-j3unAdRmYjB8=^^g45Fdnc&3h#>P$XX}*8bjQiY={Z*}nHfgO`Jd4q
zpNZO#KD8Py?A~Pl5D9VqQF!WLbqXz~Qmb>cK>S9{9%tz%qN_bWi)+t4Aw$xX0X;Y0
z;hH%@^;Ldq?{&RTc-xB3t&YUF(ZYdFH6|UD^&_|)`2lmk5o+$*cv?&qaRj<_0@=JL
zyhr9aTgx_zkxawVlwG*f8)j+eXO9g$^vulT+K)`Rt8L5v5mJp7;kf%+Uh0I>B)P^(
zGZV-NumZp?&#G&1r?@mU2Y?y`05y(h#P2Jhgf5MhoJ<>{Xm9D|4%#T9BWc{92t`Ol
zLQ*$s<(5Lem~pa1&WkDvJM;ST&H<`RDB3FApI+!KFjKrFqjv5@=goOTI@tm@L`Kn9
zf1{r`EhyDCRc}1OJNd%0V?~4qoZgCw@JM$+jhB^{+f30jF62al0M#{uZB=tL_zFU>
zM?|#*%H$sIe4<t;RvaM4wkV*qq>1>(?b&pR7x0R>uI>u2@SmdVhrTM9>JvKRwo`s>
zv`!8*=fxRZ41L}u-<`kdKUrEFP@L3%g#5yWQOtfF@wl?B2xcOyyjJgwFkuNQ8T)hg
z!cSWy<&xT-o$w-36m#M}{`7^qdfg`9Of=RAbNlT~q836*w{YHp%bKEe!J$UTc8aHe
z-vR}Uq^Rb*`c?!%m`PUpZWOUsdaA&kp<(8%aBXl(<lSoi_ypcja3W;|8zUiYmeF1!
zv<KEGG)P>)UrOt{KW7L~DELBIr7x4Ef>6~H=N%uMKeTzrm&I2vs$O?nG%1GDYd~Cq
zj_`x(QwCB>f-CBhMtQ2~kgbLRWvFN9$1Nc)`i@qWi`^R>BM|3SsfKaXW*#B|_!UDR
zOYYlT4GbSY_V}m;9G9MZz7f5&tg)gB2c>)c5No1cRtQr3l{?|WacL-E1mS6lXPWSe
zCUC40@JTRKB0^3}y*a_H!XV5blB)6^F7#Lvlr<F3E`Hb&DI39c0J1UUNxB3B39*f-
z%Pr7pqB^0HFM56P7&3&7;-IDoxi2N@C~y(9lob`bmp)6%3xN$nN}~6tcX~i$bjGWz
z{E~sE@dni--FXn4!cLkZuc<jN{xOxs!Sp7Pv%ZC1;q&Pv9mmyk7A9@RtF?BX44w&>
z8ZpbnC2V)$=Qj<cdLNlwv0w~sL*elOSCmx?;m>?rRpt{z1r@YD?4U|!$Yhr#*>`VT
z@_`lzX5Gb_u0*}*x&+13auUxZ?(R*{AqcJS&vz{MK!ha~&vJdxAqdeCNPP+1kRb?z
zBI57vSPqs#Y)C4dT;eiH5DatSSS`8`40E@;*$D^)IgiJ>18bo|R~dR{NtqTZdJoJ@
zX*6ZRCg)IsujJ6J%RhQC_>*cJRvWVfnRSMz)v)WTzk@9lYiBW~&VyZgC!r?eAl=er
z+_0X7zncDyeO2}pT5dB2DE5pwGPIB#X$=IxIv9u`83x2IF2iD-S1Iqw0Fis0&f<0D
z)j{H_Kn?=0OwSVb$+;(vS%mR;iRJcp52xyk{6VSc9*0a1x(zi6+2n_05p2@pTxJ*`
z9`S*4h+xL{^(~yZFmx8odO(KrH_%Fnp{(4^7Kt*{+8b_3u;x!U{TTOzX)t0<Ha?TV
zkPJD!IkpeLsxCA`i(bfrbU#85%r8!>h5@fGK<M!_>m)ANUk<O?j&+{qfNP9}rbB@S
z4^VQLIuUstVmeZ2aV?~C0xk$5d0yf{xHk7H+qj(OUHKYCu}nmzW1{~63sRrbE!;)!
zq;qaDwS#HU^^LFUx0xiwj!|>TOyPSc7u7w{%Tw;3J-JGjkf%$?_M3Al=WHp5-ucJN
zr+?hLqF6?DO?w@34bMN*ph~$f$l+^g%l9fL5Dt|?2GmD1wt8Yj*WG=dAKMs(PrKr@
zpZMBA@dU@ZQ)v#s)Cbu!@BaX0G@+jCzNL$%>b0RaQAGa$6-bIkNnOs3#*Lv?Mi3E(
zjV_2R?tuvtPB$kk{OjeDGVPPZUqg%dLL@b%$7}nph?{fzt{vIc264<=vnD1SP^~CT
zVjinN3EQ3D!oCzp@z^0-6U3?3q(`nz=SEA3n0A-g-&N@=x|JLw*zjOBE!F@UkC?WL
ze_*;;Pe=u7a%y&7q37l$stf?SEDNr?)nIh+rssD?-X59k#d@!OG6kG6kTVx+4AR;-
z4b8P#wT@Gfjcdk`Ng4s9yRNOeveU6oXnW*<sLajP$==S}l*#6PPpB1_?r?^|TY7_x
zlln-o+QrAw#Yep9+;a;ZY4wXkM^?Dz<Q%xZk!bA;8o~L4gNEL$lI!SNAl~Cd(yf5W
ziCN4+F^g70d6vA>Qd2Grw;2u1DQ+Uv%P%$o^hzpL!d9nKx20dQA*8@@%a%o)U<rGn
zqbmduOga~OSc-KAWcq!eC8A42ldOixFwrNx;&c|3HA8BpQKjk1QIU)&uDaXul<Q?D
z<{T2Vg~Y$j@n>=imoK}XQEHE*RHQo^c^FBM-sMxwI2Ra$A<+EiUg)ytSF&wSynMO@
z7k8<N+z`<=f+e{|@<P}&*{VgX(<E(Z+6^}3iLTCUkE(!Ke#~8^Qw?lA7;I?B%182;
zJ!}l?)l;a<L7KID)9p@Rd`P#;rIxK}y-mN)4*S`{q5Dt8J$}+2lz;LH#+)3+^*)Pt
zqhlIyRVgOr5gXkPM3RNQ!*`|#lqCptXX%E8m2DGn=fBee5Xey$)imCl$0!hzsJZ2C
zUmfrvVOmAv`;~XX-8~Y~p}Rczm3OZ=A`zlaN=fF^KW<RaAZoIpf<<of?<ItUl5G_&
zyWo&SnMrwDOLE`2JK?4*RhtF2Yl&&ZJwh+IIs2?R6!E>>1yj@N6`VEkbjM$!i>^xp
z6M&*w-rdEVXK*Xwt2ll?hERU0=M+DGNvH77y#D}24xyUycl#r5B&@_W4Ab89qCo`-
zI$5a-_xm9uQiOq;ZtBhiEU4t2Uvvr0l2mtY>Cb*i0wkoXd6mA+`NUv@Wi8Oub6);i
zp`M9M*&zocx6#XdMHk&9gEZgj?#cx9NFdLjcPJ2spexV2(E>NRwcsg6Cvn~clbHFJ
zYqZwR<nZ>kq7_V0dxpULikm<+1oI1_Jij*Ui*TmUXSR2QYK>E+$>;lZRpitlEfIf6
zKs@1ECL8&6UCKP0>Atb<36l{wC-!~zTg@z?NEd2k-#`P1@Fb9Y0P4P9y@@u-LdBD^
za@e9`a3z}UDENbaj$9x@${^%kE+yhl4-{0b=+vk*Q*5}KnXuc>OY$-moH%|Wxn);a
z3kV;eT}y{?N#k9aPSs~*6}qIX*viZ{V{e@k5A6v>t`14!tOItt<g0NyVwM%a8TCl_
zs>&iF@BMM>@1pbF0t$|4@lELo7KD@mMc%wQ`eENx&B7h3oAb!Cg{nw)ck{{wAu3v#
zuI<~PLi4Q?BzKRTMc$d`0wMx#VV%NXd<a2NO!HqyETBRW4okR`dTxOTHtyZ^Zh;9Z
z3WPU)s9q>TR7<T=_^LtG3#vSjnF)=g;;B<aKywIG`R=#Q%SLAvzYd;GA60h(f8xRD
zS3TpFNHJx7vkaaIlLH=nLlWVHo*2M(F8R;ZEO5(>cTf+)K8l!NikD~g@{9WDsi5iK
zyFb5l8^SdyvA20}?ku<WK$zr)L!SQtN2MT6!3Ny9r{4k)ND)UhU#S)ZASfay0_B?b
zEC@nTAmorC2-76^_bdoORLM6&u?S7v9LIOx3sYn!>fOl#5KKKe@9v11*$S<RmDZS{
zxSuwad0W43GqJKUcsKrd{{Tw4hQ1dP(fC36tJN}`*-B1TYA&Ekqem##p_qK4c3Lv)
zZxy4OUR}w<^lNb$bG-)zcUFwT0RU<qXVrPiLw_7e5G8iO<qh<#PL$>j=WMpQY>-dk
zHWx6eFF!koDie)jAbF{|o8{hFW&qk?wC%GYJ_O{R6hw}1r&^y1dvK09VE1B1vhMLE
z;nOAAhO!J<O|GVK&2?ZTT4oNGF=mmlgX+P$q+wH#EIC7AX0x(f8WeH=09S-_OS$fg
z7;2bTQe_-Vh%o>;&dOu(-Za_-E;Z@br%}(K1Wup{9OlI9zv!lguGcP8vciK(M|8Es
ztSQ1Cu?6%8jQ+q}eG;hfYI-jWYg&D$zyX7t?oIOE=S>o0TnBH9E_Wu$B%cqTA><2b
zc4&;W^sbvI+l6**Gku5aWIf)`?{9=@VJNT+JwRMHCpLN1J^``mc9<>d!|^pL9@(Tc
z{{Svtt(^$??wPEc4Wof4EaPqo2%0cfcy94<<<8jSLBgggp{Ca|KsiTuDG?ya4JQ8p
zgpJ1QrAFnT%+70|#E~HD5fVVRFl~FN3_&YZ{2Z&z6u_uS6BR27uRG?R9`WJvL&bVk
zH&Vn^j_bdeP3}Md=?PFccebS(QYE~MTcITGtZty)yP<k=jXl*7piPJqGR=Ox;wMyE
zU@q+zy=c<tt0Ra&givgA_(O;YRDGw;{FAeMZ9E&?N^8xByl3YB0KFBlZZz-}o-N%!
z8+}(w1GL3cX#7Lrf1>7jij{U>YE#E3__se#!vIcMFXouNjYoKVTqWW6eG%ob;=U2|
zn}!lwNzU?h&TeBXfVJj2t@@&!c2qTYcyIVf0zAh3dM)pP@MP507FJ%bQFNsmX_0PL
zVn*3dnja3LYN+S=OQU-Yhs9I{n1_%W00SgbFgF>2aT*-b)2}eID^{aT!PJ^0lYGRu
zU-)1Hv;cXl4kw_=)rH(S!u?d*55;9<TbJz3_#875^C{~0`hy)S;n?r^c92~b4Papd
zV&?-50xV3Q!U?kHN~_1;5tL4@bDSrT2JR=2^CMK+Qx0NGK6L7)D8+70H7drK)LW;6
zhfe;LMa=I0uy?13ZMk7KhARzQ(s9+zOAh`}azyIVG{5(NNS^R$x{<9ywKGEtPb89Z
z<|8!fBpaK0hT25Jo}QAQqNm|hjv{f&OTyXI`#iL<T*bT-V`FxmgVpIDqA}j3I@os{
zTPnI81@>l<>w%Ps9@pY*5$c7>kEK<<jQU^^Id1J{>@q<C$4RqYv=wZAhw_xLF*#Q^
zq|1y1<)s_~^+`O>>lo_nIjqc;YdOPfQ;v4-A0jW+a&2zO*_li<j<24h)>@Aa)51${
z7?w;XopH8s#IF>?*N|NJaQr=ypNFN`{h+H)JDn7T++^L&w$>XBmAbDrcuQ&)fi*S}
zpf~Yb{6l{fZ7kHQE>5iSfe4D)3S3*hJD@^JW^0(DF8$rmBFh%$R-xyT??%@z*8&oB
z@*LA7-N<3Voe`*zm!>EsU~w;W5R`?31VP?eXUS3)heY)VYTetEQrx$qEI72H$cQM1
zzqvc1B0@s7m)5?R5U`*kxi4~WRe=b0<hhD+PEa9QaC=XDxv4IRsuGh%gP&iT{{W0o
z;!erKY7^7u6}%YDjbQOn_>H&Hq}ojaBZ}DMh?f5V&nTt9!+*2tIlI(%$*R{B{{Wk4
zJ2ZSrDh*lnzVt}jg0l^RLXJznFRB4#11SU<^6vD(0uxC<%{OJf6kZUAXs14(O29z~
zNg=s+?}UV$uDd?3Yzjci9l0;OsR9OxDj6re@QB5vrjVh#<J|;B$`TMW{{S4JAq_(>
ztG_r9uNyFzc57-oG?SK=amj~q6*AVP+9zDMwl#TVKB#7;>=KHY%I(8Yf+swUK#N3~
z*iP4-mwFV<XTTt3gSnq~)qMx@tMYW&p4M@VPPd6^uLW!!v1cS-(>aEM?gF?}3(=oE
zqs%&`Og&E4k%I(&`i+jOZH{xOiWNJ+8H3|{>H*vXb@SU0c^YhusoA;5QgO+<x0|yK
z5&h%~@Qmj=pc!ZORh;8CcCZejXXczH6Q<zpoYbgP`7sGrUIG&ZB`L2s9(1npey$!L
zL}>7D{;y1qPX7Q&RXBs4R|NMw&D-!RSes~zlCO6(k7amXejx;db=LeLUT%SMhCr8Y
z`?)}5rbyWi)mxQU?}U;O4t^*QlAwvjpMQJ^Ld!1)+>jv!N4S+O&wA$qEeKX8a&mzP
zMA|6uz=YU^6*zqmA$r0=$us2!@Psl>A_e=Q8QBY_{E%5E_RpZ(h=f>jv%J$k_IS?X
zcj3<?n|^BM1b@ai!2Fh*+KVd2ImC^*o~X>0q88%2r~D!uAct`1<mY@~{Z#mXj9l|*
z`mJ(eqP)^H;1Oj+#3uG4pyZk4oS@X|jY1lZ{$0?zpb&&jCR|C$Il!KwHY8sC+45CL
z657aaB5<KE#b8YbB+oTY_z;AkIo!XkAVLX_e24GxK!nv>$vAubFd+z&HQp)pLbwP`
z%`@ft;wV5d>*oR$e!JKfTAT8u(~Z*7rzYdX!&;8;5n3polI7@2?(U49DcaL+9;F}&
z4-gKM8~()T3crC2CaWr#1-Ff2cIu|<H7Tl`%)jEu+_OSL)Tp_`D%XXy`^$fN@k6Cf
ztvp+4HL_Y9_c_Myd&q8MkU9=lkQ<|006^4x>AHcc=SE^013`e+-zbtM<Pvu>N{nR;
z<|SP>P;4-?ggh5?FZ}NPQPUAe6zX+q(sMy`On?kJ+Q~L>77mN(Z4nAJaQUgys2yS;
zk|bY10QF~4Hq>qwr^8a`TciA)HJ8cqvojLr(s63}+qe3uzn6cyJykH`rY{W4I|+fT
z`lFP92a`dQsVy4Wtm=4b@2W&ew?Cepbl6<0xKkt;ZKj)RA#5xQxWs;<{_U}8xEeTn
z;O)#jX9uq^dbIxlDyr;NO42bCdAm#zC)`Iv?+IpPz}ju(9=LG_{{Sp@75g6UCB2tf
z?jHvTAL0%0Q0%k!s5?Kw!^q$KGMYTgCf>VdovVjP)c6nGU%%>#_<>QF`Z$ULOMz@n
zoCed*;5FI}G+n5cI}1{sLl8lo$5`dec#&|GCoa`ps)am<f@(@`Rlz=?EfVMUvW1I%
zI=nIUm`$MQW<!7=i@_QYbky4D!?o0L%%<)7>IgR+#^CO&bT?LNhe>$N2-aD<i3>Z6
zr}#yhFBP!tsmu*+eGmJtH5zGEb2%THeBvUGXq>*N=tSHuW-hzf1J&P@Xe!8vKcukG
zijBc?5;`qswYD6b!<5P(5B*_Slb0``2dqQ;5tqhZ`dlf@rU}kNM~TzSn+~e?4q$%}
zVR4M0pLg_HNx3;Hv))iO$;OacdbKM`lYH$S&!%?Xi{ZB@qA4<?@yATAr%j~$IY|@b
z1@90{t6ihnPYOQ}d)ziNe~jC4V`aUtrCP6u<qnM`;L7NE*{RmUj&-AO^xj}kqz^8)
zF3YCU=Nh<bwXt;jDt*ZHof?8nI+!!vcFuK%fD#ot)ag*E+EnjOjhzfkJdMQqt=Doi
zTSApbkfY0XI()T~*XoxQ#_OIZ2}_B^_LN;^I&_(qkcl#PB=aQQ#PdE;qNw-2shWm%
z@Em~HbRP9A+jf_u=)~XyIw;0vUULuB=uWv%!tEx4o~*?r&rO$o_SgNK{{Uru&Zw`F
zrX1$NX*XtrsbM~;JGA!?<u_gBwON+Ezc>E?A-LpyVbw$7+4(sqzD<(2aPzPBd1jpr
zH)#`{MIdAOE4j7yfOvGG6lyY)n?{-LxWjE394=`%8@F=6bQ0Mx(xp|7sMAryywNV-
zGk+K&){&pc^HoeeE33Cnq)cn7n9*tNX<Nf@1;591R-$Sjw3v`uAX8|Wy~|?Jz4x?r
zU0;*?5|IbnJ8uq~@&o3#a)|W$BUw<g$~I-xSqj4_5wgg%ih`gwd&>(=QL4%^hFO}b
z@8u3Gk+GDaX*p$(qIw~V2ufr^%C?EU6Ti9|CIkdE70)5+PfvytEy-jUOs^zW&zizF
zC<ZK&t?$WteG0Jj37FYo0XZoBa6r`-A`(%wRLdpy_d-0=-WQR)w`K9p5)jqAli|;N
z2v=MNjx}1Dd9>(~e&k=UKgKh7um1o{e~9|4m=F2p_#dKUjS~Do@{(lQ+iw<j@WpNd
z*x?zD?|;Gv(RbpGoOE()eyXc(YySWU+cqmV#SYBZr=N5P8ll^gZXWH~EDA=*MBaQ$
z-RM|aq$gF2b5G#~_e~<)(~I>&Ba#=CGEvXdy&)-TWepq6cpygUIeIsb#shH)q!4ib
z02GBblpzrKm($#_6eOx-zYhH1Lci>1Ot$H<ITuGZXR#cti^-{A*k%4~qc4fHJkVhq
zxch)0`4KAa6o!hf(!=Ig)Jm(Ey=bgGj#SSZxvw_(ya3VhJVXMv@HfJ#yX^5s4m4rg
ziF0t~MXof`<ZlOQ_$>@KYG?;mZc}?5K-d_OG0ImSIfR~|)GCjvIC6v<a)SLwSSgaX
zU!VzZE55viq3J=4`n3f?0@1CaZK=8Vc<QXQLguV)7|vG~SEVL(zXDyNvg-q+2P)jy
zOf`29^dA02%{~PhoENcnE}NMW3+Covw8gZK-IZ*9X5n5A`Ft!##cR^dE=<0>@{d`y
z))w+tlvPlXMLsDGP=-yHB~JeUL=sS&sxEn=^Im9>>=1-vw&cCu$chlu7d+6p^GF0}
zgf&FZC4)pP5?j3&JfK1W21U4X^d~3~f<hTvmvji)5Y-V!G+)&M5O*%@zNip~Vk+i@
zzZHQS%INPTgM0lw;TGF0wWVr|Di-U%5PzId;zD?3PcQWaX9xcPJ5SH#RSRtl2<;aT
z5@&k5X`v=vKm55xZVqPx_8pz`0a)T7dP`l+zaoECs8zlExP1v4_gUH?5UQ?9O9~9a
z5QtF|dVF%<1RV-2gKCKIrx(XKNsSY81yL2thuW|r^JLk?QQoMkIrD)uJ<GidJ^S)p
z;6f6LXKD|@bAbpdIrIH+Aq7*U6Mk?Z2!k}<KUzS9)xiZ$FS%kd8lg8f#bF^+G0RT5
zhNj+WwOtvBSMwhao*2yHecBXUZ512!S8y_F)My^T`YSBJTQxKG2POS4??sv?w5wjZ
z-Q}affH&q`xK)<ZrAzR%6XN^)fJ|DM0H0;S$y;IIH^_$JAL#w%<s7k9Yj!m#;WhP1
z*g4S9(aqiMG;a5Rx=^OUQMGXv-Z_8?B=Q#XJuawtVUrcagDl!4+ckNSEciv?;XVHV
ze>kauWkH1Ps9dy=M$ZBuat;?QcJr0i)4nLN*q+M0fdGeo6MeTF$C$db1Fg)CsFJ2#
zKtsd@>a1z9qcIf4N=AMVMxc#|+-y}Bwe;VS{{X(dM_tkgZC><0N<_5~Qq#P$K_uja
z?rj&!W%EYWfFeCaYwIrW?QwTTUNVhbIc0nh4FJg!Vs?gTG`R#F02{0LXR@|>IEf^P
z784)Et+pL?QB2FfEi)}xW?>cfjk9^A5aMhP_uJnPuIzOIQlpqV%^}2hn(Gh@Gy>8&
zyVe#bLZwJyYh+_xFl}ffaxO?9@){eB)e{2EH8(+U86~n_ZT1Ci%ZB@h_r;tGyG#>g
z`m#R1RmT<sF>WK;0Q>&{*;^MfOuvj<GqS6;V!QZ8G<mA%$ln(id;$)r2&e;>#w4P3
zg{8Kqis<-iy|bT{LBee&ADRZ(ZsH^JO^^19F?hX3(y4^ca4H9w)*BY{Hd;6ax!#ai
zZdh78O#K$3*_|V{VH_)0sZ5#tql%{Lu1gUQXDtE0Q!mOb-nT1wVzv#AsX>*ix^vZg
zvf4Y17dxBtu)eb@#v%QkPRfvsw7>^`@vt1NuI9s%>uI_J^D<PKi5JnlM<mGhH{Iyy
z8`?iu9LtOBpHfESsS{0)q(E~3f^;(y*M3)T_<dn;vs+Q2?`=UXak%*r)M^g;Zl&mT
zHfu6npjbAhn{z^HRo|SGf5g##{QJ_<j<l*ay*$7c!61XHNY$sP_=qCJE_E1bKPe&8
za7Ki4`bNe^p;{$b+*b80OzMWjC%Seer<VAQhP$aM9x}r8%EYYX%kv$X;A26`h1Ocn
zRfpjr@8Y*5;sOGWHSqMYG~D8Hn1bV^13m77C8k50t4Q6`s_G9AGhRUbIdV3KOb}L&
z*!^N-5$zjkwJG_Kb@t_?WeQ<}xw*|dgIzwJS^cQY;dQFw_?9ZQX9tkhw7I^kvXV0o
z;)y;zRorVS(7;h;Fk3Dr4x&WC@R%JnT`=2ap9thJzj(6AvLY}U`d~sL4&qYuZ&+xX
zC0M*FsrSN0sVayZUIbonAt4(wPAt$rY6K-|jkbzqyK*MzcRC?uK{~hd-2yr!3rlU;
z{C7f0VOu3#<;9lefi#6}QFHr}1YIQqc1oG$=tx8&XH=3L$22Y#PH;d@RoetO?4-qQ
z@nlFl2AYYEf0b@Bc*K8Zo9;eEXmH>8zb~jzSgK{&%B!!qsHBP^awb=Q&T$?F+?*AT
zpIOg8<|04KK5AT1*|>v;Akpxjs;>}czK?`!tmJ|m%YJ_d5U@d)-O2>e$ambHozRge
zYS|)gC?fI`n@T|jU-5v>$*ftG-2J(%E`<pQ87}O%dto41V$ANk1t4Vzo9WMgrXy{W
z)JheZuI}XBlq7UQQO}=ua=?XZ*{FZ_%dpl!fhHZslxEu}OQgGNBlLK)%RfYAaW@Km
zl}8wi^ZpR7;Ra^ZXx<m!9DA}ipRCm6U8=SQx{kKfVc?G$UmVwWnBOwDDbdnT`>tfG
zY+_n;Cy1-68g1$>OlNoPMtiZf0lO8ELA=)`?W<-rOf7&h4NbS3dXz6q0?aeI+usXw
zNZH^Rns|;mNs$LD{{YfeE+t6Q;m>g9-}ME<riz@Cn)gS(h31wJSD7crx(E?;h1M4}
z-IpjMWY7(@6(rB&yC`TtR7G1<d;Kt=Aq<mvqAxVuu*gC}qM>iy`r;@<Azupp$pR3R
z5fd+Gq61_g5+JAWfe8vb^F;gLLK3F@le^oT2tu)QJpCKsLg*hQ8JFbzp*COHkuG|6
zVoZd_;W2Ve;kGGp0h@+a=qKDlv4plVnd|f_H_AD)J*L(sOfoY;xh>0xfBAp26gVDU
z!n0@E2jG6Pt?@DUl-@T7<`#h@TXIy(G~}_kvk+AUF5mY;Hb;sP3UXBC&p1|uYJy|l
zsJ!{Y)iI^hauvHgyZ6A5LLgL4$?#Yc-8WIWZbkT_K$;0<=*k2k5?o)#1RxOZ-TI(H
z5sv-vAvH0{cmB9F2thR6YW}#1*2qogrKYFoxQdG``j2jv31Sp368vSJahb(K1^`xY
zoQ?!}gY(ey(OtmiIIN8^Adi>HLrm0)yE!!MhUrHgGjd=*@Q2&&?v9ijP~oZJX&8at
zaE(pe8P_o~?Hiz%X$pi2udh%{p6gE`U|r+|wL8GoIxRBXyt^&r(iOrZ;1S~f)==Ea
zZLfyH(GVI-;A|Knzvg$3c1nnIyBgSsnJ$LU@fkkmg-OH7r-(7^vhTKehqC%?{{X@`
zaQ^3{Q9jGh)pf=}G+)huSSA>7Tq&7P=N>lHzI=b@zAaR3N^dxfIFfy4{vpr!#@xqr
z>_-+kl^7`2Kf>?BJpTYG-<G3vmfdD)opyJM`hT<WiLvahOjV?09X_Cqh~JXt_PU$9
zH;9kjx)qSz<97<EsauqD#5c6UymX&+*IYZYVjAx+fAv*lnVGAnW~CjnL%{s%@m^k_
zdf!Wud(q9WekPc0J6as?D~8UziO_XYa~p?yi|=y-<WK3Y>K2PjlAvN7MMRKA>@G7W
zQe;f-FE|#T?W9-14#ut{3;11c5B7!E5_?9`u4BC0`2OWtNL3Rh-SuY|)ma8Y`V0f_
zMvN6wEoA8e7z7~LFIl^XIBF}7_UTd=HiA-Q!W%UFF)=a5H+FYixP{M;Pu4>obBonr
z#<r7$!$2M`B+rnag4p5gaKk>MZ7%r%Z#hzUQiofkVa=Ym=N3sQ<vQPI9b=Rkg0P80
zed){QhVP{ip-&k@3rX#pw7Ns24$;QtggVw2U<aJ)y3}?R_3AO;7;tPvPEZU70xu`p
zBwJO~@)Nb^rDf_Jie7S__nc~Gl`(Fhcu5i70{bpUpr-tOK<bp5`Q45;9O1wN<n@yF
z;PYi;+}>fpMEJCm?`n&wz7q8&D|Cr>8^j{evi&x7p2l@%KQ@}YBw@cvN9c5{a(lyS
zj7=7h>jyQ!w7sK9&0^;J1QQ}zxQ>jb^G~s<1-m(<J1H@97Y#Z%5IO2EY5XIhOVgz*
z6E$|in*RXI5%6n!h@Uu1>>cr7)x!!c#JZpjc7&J*!fsDUnY1n<opD*z0|_HgFX!R7
zIzoi0<64_-%Qs4*6CI-9mpY3Cb5o`L9JqPcJ}(<6o+z(UyTAsUGUm{@8Fxlg5chbJ
z!OVcWwwnxxyEI&ekpOdZ&vCw4k04EyYghy2$(Ab^>GPYJZ1otDnQ&cZT)kyL$(;Rw
zIVY^|IOph2#<)L<65*&}B0rcnyZnuj&CWr^3@|OgfDVQ5XUcE<#tdlIs{RG@yI<RZ
zk4Su!$Nj-V!GJi%Jb@5cAZcKB&?#aGcMHOoUQ3V5_u{v52LAo=t*XyNM(wxR2^ttg
z(p<9kcce53gDEN*@PcMYyj$R#LKX#uLb$y<l~N5+WTYEb+d20^1qlk$Wz)_Jge)sc
zib|Wls4Rq?Jd{-xdGdp4LclgT=k4xTNJ&=rlBz(2m2Y-ga`Yq!L(^1SRXMLENDh`v
zL08NU4gUbr9YIYpmt+)|!{@)@7@Tka0H=Y}{Fa{!gNr@`=BxO<v`Y0ww(?fnF>MtS
zs@UtnL2m<~PWg|$?p*kh&BW?P*+=<8t`OOFeu&f(RTPD5(Z6>wfe1<#-Rxb;i2@RA
z&3AqIK%VIYIW4L8z=CBZ2P<>?;7lYCWRRj$Nk+Xp^N67)&OhM-V^jicl(&z5<ycT6
z5G?B<kY~}F{ZNQHsdYfhHQl>*?4czyqP6V%XK!sMX!|^)OI89=$y&Hp9-Mi*^d)4(
zrk%+*TwZMuJ*GYzqXX>)Z$iP+#W}fvCfv!&0P|X$HI1!V^ruihvcA|?Dw|csGxK=j
zleEPK<mwMn>J618tqJ`$`kOwYy5w8Ae|yEzs_b1LJsGffE_8x5v=DEw(|B7>V`(^n
zP}`utpVf1|FQ3Nll|q>>v=`-EG!}|uvxLZx0QgVq>WI{*NuB1>Yol#F#WoW11DrzE
zn^4(v^(~!Zic33VdS-T>V_q6<SY-j}n<brg!dYB^Q%qEvxOXd-m7FUb_%9Tp!}#4l
zNy7&A-*=c2a)&q!!a?2x^zMTM-7IzvrY8X5O04AXM>uIQ($Ys>OXnMaD!Y++b6EOq
z<#{fsY8$P~z76D|43l^-IB3x%#F08zdhdZ7N(zFd*{8)af>45Dw)6O+K$e11EhV`T
zIK7Jqrh=h%<WGtuEiQx=1Qg9ZJ^P^52yO`HYNr$$p$^p~$v(Icf=HQV?cD+pO4MAC
zez*{~SCf3xl>NSJdS#$RHN;Zv+DZA#+Wz!a#<V+!R|oK)qVI+|adgk-3c=+|Dg(6H
z*{0Klz_ga$qV>Meuk6NG1Bk5C`F+>=tK3XLrTpJT%>;_OzgkA}%0M7R+`n4GLK=c9
zWWTNv5@i-bR9y2<*9zetDZWX}cV*5Nso4mVc49vb@E}+q*}pye)&$dcLoMr?Z-EFa
zaF-W+2tqK=)4Bv85a-SWCdN6f$SVpCp$G<~tG|D?B83xKoNk#$tyCPAM>J(k{W9Sf
zEg?I9RA=!)JTd2yY5xE)R`65ltpjj=NsT^4X|fU$60RXgw#bC0uL%=&^k>s>{ha!u
zG-_gS?QRyU1E2`YVq0sp6P3|D;0SYf24y<5=+>I_+*>5f+ylsvBu>T%JjQ}XpHCW;
z<m1k|tx3x=0{gBH6L)cLPa2l%Q{DAp3{xG5s_Fwfi40>PgWUzrE|Dx~bGyzW23tuW
z7HB+DV=DJRmpR&tm;gnbNEQI^$Q7++*uQ94&#{@b!_tyZfQbv9xNo?4qm~k#VT{Dm
zeM2)Owg4pGLpHsp{M7LerRel3F}5Ip988GnBI?5GPckKY%$O%ur<wl%7`pdWpeCh8
ze>5=r!6V=mX11ME4P(ETygrBGtJN8o<dMp=6!$bxe;G!tai@dU&(r(*Vw<saDAqNp
z-bl_GL)z`LNYTJ*NCCh>ycJ|&>QtrK)U<k~-O@Ti)u-X8-9xiJ)+}+>mS|g{kDKh|
z?T^$xKA`(M<EX04cW^%|O`>7W+jxOH3z)ixW^kr*$>=0DE_uoMY^@C-02palB77m?
z4;JsfCx(HAYX`G$uhl~k4{ls{f~)mZ(8U;5ZACQZq$cDDv+Fk8)*sne;3Ah8aQ^_f
zz<kI50P?z?_YP_N4)`DUQiM4!QF&B4$zMbOH;=M3je@CkX4~Hl1`1IJo9pEQONF+#
z{{UQV9=P?&%;iOzGj@O3<(mLM(tAJs)OhYD<6Df^b4T|;Ugiu(F?g6>iv--9K9heL
z`jh&3E!9WCld@HrDF)nbX~|i6HgukynUizowXH=1Uq)RKsY45jquJ4@MM47s8Q#M=
za}Wq|5IhEorCdc9m${&_OW@^Ww<1U&Ypeo#n6h<=8NwKS_W`;a(zu1D-I;QV;XO4k
zQuWD7RMeO+=?c$$PA6D!>=~N0-$;N?M>+0mxsX`*WhCfK55zSp3^=?%ns04EF`1#Q
zBanx*59J5JN>M5JduSy+BPULjla-eKbtrFqqd#xwI+kAFREJ;elueXyJW7l5HLPoW
zY|b?t0KRC<X|xZ-bNLuieN09Qpf7QG+5lr_VkanhJqBCv2sHW~Cl4F7?V8HI)Tpv8
zRjSAM+`OkB8K;V6NAjrO4-Z|*lvl&yh7b&90JsA=oO{BUCP5Q9h-sRQV6nbol!e}{
zZrX5|yrv^7MYE7DHSU6EA#v#hnOlO9L8;KzZhe+ppPQK<J3XtVIWpITbKcFqsyC!D
zu*x{RGmU5!$<^Mg$mMxtX#u5;-X`lc)o4~|rzVhZvl7sECJ1!(aB&?JvDM^ST{?)l
zEbE;;xs4G&fqc<@laAfJ5#6i}K93oF&h-zIis#u?B509Q_((qSXAg4J`!Z<HX*+wF
zF0Ym)8vPbdDPXU~IDsFHzap(-WN_>tJe~glir+e4sx;LF>o;*0Ll0ynSWa-JXxgs-
z09p+Ym7)NGHsz9XNJ$gTc$z}rpALSAV&yc86Fl?yBEy~1NwF=>6XKacwb5-?k+f9l
z`}aTv?wU%V^H07N!V*Xmi#3ByP>|>+H6Ma3IOK$lHATr$Q_`74P@GfgbsD74VWd-@
zesw{|QJG0+A9lb3aLXq6vH(R=UgQ!e32Q{S!eD1G6qP>_(r}#%5LByGFqfB8m9@uQ
z5&Yq9^`jGx0&vV^ZhB}3HK)T1MQ0xkJk@>r-6IjK-Cl<{4^tgC1*e+nJ2(sXMUE09
zf??Wr27GflamU|@Rdhs<egOpsCE@$yNOLNhO}Th;%h3y4p$L=H-R{GJE>N+iDqFkK
z5)j3_*Xw}^D%GeXZ!csDLqvlQ4~sZj*r%c%mP=~qy*Wfts=|%SyLM0}EP`Xdck6*0
zrs_OgJvnn&M8%3HypUnLcj|?5q?$P7^EI3@<vO6U^DR%cFx=4F=AFCE@3}4FT>ili
zd}DE{b!%V)TM!OsPcKvf)R!G#K-OVj;qGHGB%Mfs@g{m%ea_f!Bc$T5CQJ=epn!~u
zO-?XvGM|gJFy51HG#pF98D-R2ba!VSN|Q2SoHe_z=kx2nyRC;AtULnJJ3?VnVyL6s
z>wow|6`$)u>cIpR#2!|aZo^)pVc8dfsoA*+Pn6~yYh3imzZVo6xo*stzfG~}LXMm?
zRT!+)OJupBfxwZPH9V~#5hcCh9r+T)7WPqAVhJ!~x<9Vp8!m2_ab{&@pK-NAHi5zt
z>gkW8Q_N{G&fE``{xZCqz-!#r>`Zt{U8*Ube{w|P@KjKYRZy2T<qRPWTU_o={_L}X
zB!wN>4|?~(yr!FwE!(&JQV#?nlW2>QrwXTX!a5;0AWVxT`0j_b(2z;fn!@IUB@|RH
z9_{%-sYi+#f+Bre_h$&|ixG4sY?9>rSwNPENC>I%?tu$rY*M7n7i>EytG<Y`(zueG
z^Ov2RJ)k@MJmZEXrT7}mZtQo-b)s_ODU<kupYo5#xIY<fX1LxLt+f7to~g2TH07pV
za95<g??hs7T&%BIxreGP=YMk2;$7-g%o`K(S|L#g1wfHNhT&1)bb2L|%?!a5cuSAj
zx)G706J<>M;pt=~sAjpUd^_SNL>-fJOw*_JLQD<Nj62YKw<b^~kfW03pH^@oHZVm`
z-inYR0DyDi%X|n;RWeQeDc=GRNOx|52u?Se01QF%z7@SI3QVCkAh|bp!aAW>@iSVu
zZxRCc)8==l^+gmokPhKhr)WOs1#1OteLoZODpv|ooy2tNb86^>k;_W!K7AraBd3?W
zcTf3JAzrN};AzM%acFY@FlDV0@O_ZM>p0r^Zd?I$SylL#mY4v;rr?4|4I^^nU^(8=
zD44QBZgGuzOMvzdZQVB70@89)W5h(ZuM@du1<6x{ZObUm!{IS`vtnH?so#)tIL~;7
zyDNwdB=;>EnFU;}UWQZH2n_DrW&$iTH!zutI{dAo)|IIisd$EkSkiRLp-6L4f*bU-
z-{Sea?&Z!K8&;nYsw-19Ch%Td=zu{bmNCfN%?>Pf3Y)1sGB<z(-$3b)5J~c~dnH<?
z&X-`Px~n|A11N)Iw|f)#@{8@Q!YI>qaL~-Ug3uq~1QJQw2~{0Wh*YR4ImOeV4kf<`
z)08AX*c>rAiPqjR;t^u!T5sX&{V&ns>P>rbZWVEEfHbBr9bXa)xi?1dB{J9H+F6Gl
zFEBsDF$1opNNP1xY-MI!T~y&{-EV;Z0Ha&~0A+s9PfMwujYD4#NzV@T+C&5S+nPJy
zhynM235NhmrW#l&)~FknGP-1eBf4X4#qHp$XCzF%=u63Hcx~4viM}rM9{8_?1^u9>
zxdag6(FZ^SWO&S()bj6$A-N>zAcD|jjY%#u;T}o~zb`j0BSxxBPh8xMVpx^E5lcb)
znmpI)NGmqA40ay9XM-fSON}>yI_?stX;s5kr%kZ{A9ubK`L|`w-JV?YNYXPlRJ9A}
zxbb*ZevD~IRMEwkNqW&gY8ptbFdB_r)eKmJ`_v(P@T^ZwR%lpet5TeF*{Nb$gxgIM
zybZ~NgnPYSHy3<(dYN3zMlgx(YV(PJCt0~KfKHwZ(^3Vmqlf$y=gioDr1-9!!t@M5
zRm4>nDz$2WNtDz2((?62?`NgHCj_+}>dSyz&NpaM;uW!3t2ylj&yXq%z*(S|@2Lbe
ztt1;dMe}Mu2Tp+$!%4~@mKY?-Fxap$AeT4;5*E*Hi-oXcth}t3X=}8RXi7*<Om>O)
zJ%}fBi|sC6A{f!^a2N_;w7t2s0q-t(<N$QI`@r0^Dt)dm80@YMc2T=om7|#?Lr2f4
zQ4<>~q~a=r724xY%ha0_(CigAHQBq?bH=|zclDg&v$$OhYt3w14Lrbz4kgBQFL5Lv
z;T_$aT)PZ)Tr))~2Ruc{hWItG5Az%H0Zp|)rBFecdc!UyEuaRa=fa~j3^qOxG1j$@
zk5Bb{qqt=(JW<AuKs-6&hmdg140h2u53+rg%(|s{eiVz`wi-`T4qo$r%rzkDyv7@+
zVVamoOiL2siKj#)%%<NTlAr$ovacTqALP9Y=KI@#Ver^0!i84%9$}K_46fOSy%|rj
z*SLL>PTICEk*L>qo9hT~cwSsB>kn*RM|f_nhleP7ZjZ^SX%7&S&hh@#S@c{P8C^^n
z=3VtGhooX;I8-+1KSinbK_@9KiT2plK3E%+YZ45*OtLluLJapLXK<!crHRL~i$g%D
z00O`O9bxZQ@veLDxJOkoOGzZ^O|@Iv(azD%<o%e~S?eO%XMZzDkOUi~41`@Axu@Q+
z88<<&m77-XecSut0Vx4DD+&pg^ylbUb_Nhi!h-rZk$ct%2<TEs6;!x#6y@I!qGHxb
zK~-G6z0h46A={UFF7EY&Z?Z4C@`9pepWc~6ktmTchM<GLkuQF5rcog1hE1X-e$R|A
z6xO}kY0Ni-rW=)2?GH`LvLZ7k#|E|M#NQb333SCQc|L&Vv`m)nM)Rn?jV9ORtl<Fk
z#*$(IzmXgGg|FJ!ZX8J|y6stnGc=Qzb(cKwIrLP?cb_7UC&n;UFm0t>t{}Q+yhuJ|
zxS98KTi6UkF$Sd~-6AX}!$A=Wdu7riLwgiqW+v3+B-|2dwL<egan|=KcW{2pYki(|
zV~49r=Fryx$eh96KBzu?<xt|Srq{9KUhpTF5J${~wi*IaZ#=h#<yiDKi7O=~Lw^4N
zy<ma@5EMN<>$@l~0ur&ZN{&c5_j_R`u%nt>nZSgC0#w{R{)k9kvwfqU?&vP3VvvAV
zkMCHD5L9wa+wxt?kRc$0BI1L=Fkv@P$#;IJN0~(Kf~l7+-jql*5)&FH^x6xR`izA=
zsW5jZJ5$ZNqi%yfed6NvS^{6Q8Qe|usc@>VtufsOE?~W*>k8HkJ^0oNaW~haO8X1s
zfkH}Gk{$$>5XqFR)=88=%J#Keu>I?~=kG_3aAPPpvojzQ5o5S9xH_${Txy)dO*}R}
zVN|j7vi6Fq8gXki{{VzBLB{ICZe?<BmPXBL?(M2dUY#f|J~%>!S!p#aJtk%59(=v;
zG6Y{@E^tcNc6LUqjq3ma4T;XyjK=PRF%sLBm~aJ0h|{X(yi)?y$udmFQcMjA9N^kP
zBGbnN0)vtdn?&51WgkKU7nVHGRkh1BUhUA)B%u)8LzwgKgT*A3ZBYqV>r{aVN*4}j
zzX&<P5EL%!P8|8=lq-OQ1x&eq=r&0RhbIeq`xBfOgk1>16>?8RBy>U{Meolfho#d<
zAyiHt?&wh{LCG@SnP=ae2ty!5Ppfuo0yj<Y{T^Dkh^aF5VsxZVwIs^2t8-2_s?E*i
z=-#MNg^a`2;#F(WbKC<q&D|CT%o^z#iHV*|OAl6_5|t{AXqRq9^*+`0@X$w9W%8og
zwjaQoYpze=d0>?~VT{a~QyB?SVdmzY9bvbCc+o#ZeKCiFq+a)SWP;F0hhMGkYz?QB
zEj&w_8qnYfaRA7-OLh1vv$&$YjnDAc6jfcB#MN*0okn(Vd-Ib|ak}YyNYz$7mKe3z
zIxZHLx8xU|tv-`bIQvKDnwX+)&-h1MlDVejiJp0JFD=l_kpT#5o|I2)D-mSH(3KNC
z{V=SZf(nLf!k=0!2tsieJc~Pf)0_xRi5JuHfeE=K-20$H2@dP6`qBg@#1`~@&>;gX
z`b#2C<%Dj80ut%p^~6x4stoH$RN4IGoXBVX<l?40X5n?f{3qysipCb$+Fm~}LYTod
zw!0kyt1T#vQK^*9(IG6SsCtCnn>Q}&$edBRea}@RIkan@9>%%NClQf_&`Ha<7Xhtn
z1G}O@xg@#Lt(ug5Da16z9L@sKHn}%4=SBPgfT1bVleD>wd3l5(R`gwD+qs#hyiF2#
z(7*a^D^@h>P&L{cTA|i#8rUO33=#;o{W+>_s@i0_n*gyFB*;2~EHB`p@#D1S=Jssm
z2plG!a(-A@oOg;<!{gFZ{Gty#xvF??x-#kt^mvcoG}SO7=3ODdyo+uwG#4Vu`%5a|
ze(dCfds(_$d9=Yi@70S`>0m*5=8?(O>UF@o?oSGPBIP_4B28RtTJr6qy556wdAn1z
z4Qf9XP(v`zAPx6EJcms=%8^5qcuTYaHs@BO98LXU{z4z49kOC5{{YK8xB(enN%uH+
zhk*_qLg${1V>R0o?KkFHd#B118Hq<*Z?q+)ES*j@#3S+X588lNrW`5P9ho~jL*U^p
zKKbSlX*vt39OIbOc(mMnM^u1T=SQ<SkkpgrNRNVGe>T3|+$;7=!)-U`!Eq-(lVm-z
z;5D_&nm0C=4`@jj7qZYX7!cZq3rqrf=n2?Cw=V?)R;0UE!<fdH(q#J6oODsevd0>9
z5BMw?_c7Qib?NuJ%^-{GB4=P{ec)=I5IMM)Qgyqo1RpKA@8NXax6*1^nWRsgZ&xU~
z$BML@R01;qZ>p-`X}Gc5S$!x$9Js!1BUMo7iU35s4fBBtw~gu!Q1EUotu*Q*(zMx!
zgX!7B9C0sx9|cykBf8Q=%zP*0x>9Ru*JU>tk_EL~xrV5XGqhcXVOg1RIVwF3>F3)a
zzgTW`%QX85x;)kit*rhXMzcW&Jwi1E7d_HP@PHuAuD3BHQ)V)=9S|D9v>0n(dAN{y
zt)+l#)8!{F-4Vw71j)B7HED)VvQwQhKQx?Qj45O&9Rr^MA9QHos@21_xOzkuoULo!
zC9HF#_Gj^%0K*_$G0vEpc$!q}&02thCB>uGcx{V^;n%z_eWr+8=97mc=?gP5H@R1T
zhHiRf@0pnQReUM#g{YpaZBnAjB-Oi|Y+UbH^wPs#=m2=?vma5_8YzGp28|kcoyqEN
z@>UBs@eerd;JK7Q5Awor4{2k>$bt>Q2VyxL!ZQ|`8*)FmJwmNNLn%3Vi8_3kzd~`c
zY5F_FHfjF=My1j84~NwzqNC`#jVDoSds=P;8h+3mm}E%g!ek#&)&W>;QB~rjjz^6j
zu|Jp#c(@PZ2IUVA)fv|c6x8WCZv?c=*8|Vl^NbpG3$1vr=9anZ!G)|GMj7wVu1DTp
z(X)gIs>z480f?y=F!2<i979`|@eP_zj&C+N=s_SGa$3(}%)0CQQ=B)6Tz?Zx6aMp~
zcM^J?+oF#M?TuV>s<=DvHW7aGok)!!ZO-b~5bWDS2-dWY=Q*{yYd4u&_*fAh5z02e
z=CaPIXy3aHC6NRrxdXp?!Z%FwShW#yP2a8@LQ4t@ZL8_-g4(A8bd&^C9MkH78mw40
z^G~T@MA;!gRS8i(xDy6e6|~QP*PICOL$!G-=j`PLlx?~~j{Ne=K6$`wrC^3l=DbPI
z+nh$_8i`u(XbSzfVp)lGH(aO_mVpX%GGpTUX%#VLz*1{lnq8-n8~CdjXw35C9oYDf
z>Zd4@mHCRqjI&HK7-*^h+<~};WXt8$KCh=uZ7<4uvE}&V>Sbd<?z9}Y+HO1n5IS@L
zZSdB#W~k#N#Q4RhtQ+;~DxH8Fc4DcCiAfYdv`~Yj{{S1C;{E>sqZ)h*)vP%Nd}QC5
zxt>-ErxLlQPUYVS9z&NsR?2L15guE&#`?+$Cf;B1hNF3akfFZh?oMk9#WB?khc9ki
z;DDM!wZC*l2}<5c_3wcQWEn4Yz1cvdWF#wDF7K;+BO9iMp_D}kW;w4)a^JcHA(I)U
zqneMx1ohE3AoO&qZ-i4CExWL$C}*`?E*PFk=dksY{{SdUNxa)L6B6BmU|azwMO$pF
zRkcWR&N6t@YE|G>=~C6%0JQ5-n@F4W9xC?_EoTs>_idQ}0F~Z9CH8M)xCajG_sQD?
z;Y=Zj=cn-ewCzM&nyk~ClAP7l5Ot##o;FED)2WwS74Q^z=926%bOz1fPUMTpCT)4Y
zKp}0qJ7)l9z<;%(JU%6Jo;Q?rx!ge8lS<$u<5_0dhM$;g&G7oa<%&)+G+kIWD`<IU
z(y-<o#}aNF-B|IRPhrpeoh|Z9EYV~bYMW5)Ca)3dO_PE7Hpd_>H%R-692XB>y;@kn
z<~6kOG)sg{M;BELZ!CebNOkiGZ-tH_RkWJ=Bobdnpax;07)iIaBi*X_X!CqTDw%JO
zs%R{{QA~}oLCI6QoITJCEGQ+}sc|aug6Kg}-<tY+_d%wJLov$i{XP4lCdfcBLQK(q
z<%T&S0Yc^6pQ;{$5QIUM%aZ&vgl(cap<8JYC+dZ5(1NY5S!d7CkP|3DbOe)uc(a2`
zq-+ut6WeNgAWa0(M02?ivge#5p+MCNjN^?xP$n17H$SQ<hiM6hml7|gGwT=2%@A`3
zkkcLB@NBoOrNnIj?5<M{0llFSkP;^134_w`I>}pO%DtL{Sk9JnF0ykj(zLW&ZPr9>
zeUD$fb6x(BE3L{r-UZ)$=fq_b8khu3{{Sf_O@X{<web<zhK=pZU+W7%;p4uGB!~)v
z1nYJf+ik9#dfYJw4#>1f_c(=tTGid}_kg8I5lb{k?_sLvWC9da6h-gvgbcb+fT@z^
zT{(BabtsrZA+<NBcIZga2~cw1_2#fCSltM-PD{(v0uyrtE!lm`0uyd`67+gtLJFpR
zTm6tB0EdfzcYFv<Tme<${ZLqhCgz-0mRCe}gd@|X9i6DlNwCVBj@}kFHxYTVikRZo
z@Y?0(JE!Ae@m=u2#~DaB@dXovS?pUZrS`T|YI82JD`NNpSUl(LnEr%Q!MI`$7&OSE
zOt}956OKSUfO4^jX`(B-BeZY>r~@(hmiv$qi|=vG%%_PuqWwyunGB{g6L2)Q;&Jv`
zIqFTDB~)cR<Ft8zH9dI_Wb)8z>nL1bD=zH^2b{xVpw1w$lhxK_f6ExJ6)JIPuJY@K
zNVUXf%O0*$;0E#uCRUpZS;O7g04<s|v6z|EkPQ4FCZMh|K`%1p*L>US<M&tF`{P-a
zDpRO1xTACGyFmW{pm^P*SeEVvbC}AFHK=Xt55jr%Z0>Z>>`4rN@>uR=a~BYTB<l`~
zjX0v{^qXbV5g<9Gt^m%)<J^HQA<Z%KSg5|-+?akWsZ_gF$$_!H4~3j-y?6frYBKB5
z#9^^?s$m=&Jx~cG0!{2^D1&UhZDpZGr7HMpjIIeH<^TkZw23k`AtA}2l$4vPPtT64
zGS2Hbp`N391l@|8x9{ineORxF^xQoPtTez6f<?<_K`sTMxk>h#M^Sxx(e!$i80I=l
zblB;kl5<AH5_T$)hMv`;Nd~57)y*j?f6t{OO9<mr2Eulc?kc7v0g8C@_%EpqumWC_
zjlpuV{ise*KB8D5Aa1n|+^Z7wEuGY+rr9Hva&}3BXiD8s?Ev8pe?(R@_`8cZMNr>r
z8}I)BWpe=B+(MZ*i9f$J%hGs?mtMp;uM^c+Ud+xMl#-Q}vC`Lkm}o>7h<9E@yj#6!
z^OUMms`kFObTWu&5=_`JpurQG?mKL4D{1X}8sSU4Us6fZ1eX|)bch??L@u<zlMmGJ
zz*9d5QfY_f7zbI@Yyrmf0Bp#dUE7z{k3om3aX4pCZ6?(YXqL!rpa>wBNRiF45+chY
zy*5#-D!QYo!7$<`;M~Xp{zvAm*hfy6Pozq^@i{3x%Pz`KJ@d0Crao!Udeq@x6~4b!
zj4eZlEyLEoY9^b_i>)~yB>mt5$J=Q9K(2U(isjX*+q01N@)H@_Bh)7D9?3kSFP<+(
zaiFN3Q3bv${4kR6L@s6l){xN>sF>lZ0}ZzP?B<>;LH__(Xn#9(;NoZyM~KqzyVe-Y
z;w`~aXKz-wV)F%rUCwiE-sZ0}U`KcZ69k2S1ngU#*->X`+vW$=4KbLPXwkV7cpJjk
z9!>64XRw|rpRBNr)ut=64$Ze^wmo28Kcul+fVt1YSE})MRKJc<B0Ng%#e-Q^gBhS<
zj||pb60Prz<(1%+gTn$A6f;rFsst>fd9Lq9?4UyOH<It~@6H5mDJb4!?yLw{QT4M;
z@F5Lb&3EU|l0bxlZsgyWE>IySM>XH|CU7ThP{??<_d%*#q$p)?$8zTakbt!w(5e#b
zVjLe!c6Azk=4MP>KbmNDWv+34?jQMB%g5BqiK9a>ZeZyrKqaz$)*VjKD()4T=+%5o
zMTVrof$!9vm5BDYvrHbx;H>3c;;5A4Q%*Z0F5`~1!ajE39%E`R4=ZEOamV1DRg|by
zX@)`TGpPb1Ezm#^wfJjs0gruVo4EAlr2hcC-*p#&-C}vNpJK!rnI>jkR)W#hIhmt9
z-&}0)flZ<Pp7iM-IjcWQ@`atm3y5ic&VRZYA60)w(%$Z(RgIOKeHWAy&I%@ZCTju^
zbd?u0U)KT<&{I6qc3Zya4v>q5<ps~5HRh|od__qj#6;eSx&$PfXt!I`pL{K#C(2ON
z6jMJJE>LWwhM-HuzjE`4>V%CkRNe5qI1|>$YTD<TFEr0A;7dZirYc!3;>+U*-AX}Q
zhso}0enV`S9)6f+l@{H7W!0ANJw=yrQ+JD;Wbvl8{0_4pxo7_X$lcZM97}QZ%YQfe
zO8X12YL!Bx<ja6xF(>6ht6XAT(YMXZjNw=xVLrI>HG4L-!kCMh{w&VT@3Pq$%ciF$
z$AAI{lII*}^5A0(PfebdZMkp7r)9y$b2-nOX>3s?5=(O4eeF6vtY=ko8{KXGL&tSy
zm@YmGuROC#uSu!Yf>QS>_>*C_%u1UAEz;zqsZ}ecm3?<-2pbNpk-4j_%wv}ar$V#v
z*qWWwM-fOjS#f<<$Ue?ivQE<C-b%(b&!<kUCf_|Llc#hwgh(38odckey8P?BJVpBR
zkE@L!uMl-i)KwPbPF(U>XidmHoR_);Au10nk`a_3n|tm^H5W+*Q#HtU<eQcV0W{f0
z&B^HJ1SE!HE4waB-?|ai8t8=RmPClFxp(e{6G0(W_UF%h1~ft$0;=bqe&k43(JfHa
z7cKstS-_fX>LP*e@b{;t7Rl;DkC#+Q96~HD&-RIj96|h+LBpTdzw}~*UfQ_2ta%L#
zd!&*C<_)l$&QTIH1n#XCxwv*cgE4bN2%RV4H|juJzsnD47$a!gSeL5ciRYF?<t74W
zw8@;fJvgmP-g!C3H)OlWM|j7b;k79@99YVYG9RqLfNdr<69)d}t&D4(hHhq&1-?Eu
zoera4HO^RlK;-HO4k2Yc{`5?&&xuFxSoF9T=fgO=*?Wh&FY&)w)nj5?6~P;|r>2`J
za!j~Yd!pr)bTtgOqJ2<UAZmc2ni3-S?u27x$0Q;}P3!3N=LsDU6j3)KK!IR@yU?;u
zC4n?DL{<3mND!NlD(+1A@9%*Lv5KM+D*RwV3z;D1iI!Z|AVO+Fsl<y4K|%qdFAcfo
zkdH(u3_9S;;!Hg8{Mb*inyP<DF~brwVCuJ<UGgif6BwFj^Ec)dWx2Y0a@4#(F7i3@
zY_^=rG`uam=@aBR{j^WB8GJd_9fxtDa_1YJ7}nT$lX4mlQvqJa3Szi5tQs^LxSO8<
zED85XUMuEZX|*0DTLe?X8Zk!$LG|2cC@yh2x&b7gAz`4<0I7ZEF0(0VFwh2qPJn!O
zom3VmOK^Qj`;|&wtxlUoHOz!{Tp!&Tb(s7;92=@@B)j;WTK9ZO_bb$#+Q$;sTtEZC
zptO7;74dA9=ExA6D}_YUsAFEEDOaZIH3WnD$RO-NxFX?aG0&#v)n%jhe2(Kp#DLjM
zoF!IOcpNSE`UJnnzv)K?RDCL=3yL!~WB&jbL$G+<3pic7fz{1p+G|iB#NFG}b8lU=
zI}#@md(-2(EUqb=$RL@xXe}ZY#zl~i5BSAAT%0y>?2vwA^j2`k?{b~N3as;w;uf6K
z1quO&&UCpu^nW%!-fQ9X#GD<CX=U2iZgp@xr$N-*>>}v`FwV>_N4S2c_jGn?(zGhx
zC!tA>QgZLM?6>s>-v{i)UL51tzKiu~{+}?rV&J`%8|^px{a4h1A6L3PEN&Jo-pHyD
zBHFHM@{X8w*?pX0_8L=mcMvjq0r^30!T$VW@w$wshos7y7ZQ8`{;N-hHP6IUdv<Jl
z&-7Y0LXBF)xPKg1DaJ=TKZ@FtzqaX%t{aY9;N#-Ae|M@;g34IDKz&R!wqC|qw=9`-
z5J5IN5JVkC*0HHZkYzg8?!DX02o@m4<mwFhE{el6Y--HA^tqaJl*LKLNy%wx+1Z(C
zhuY$`=FEKCiNE5WD`QU%&OCEa#8IVI2Xr~colK%bOGp_=-a(16g-WydS&Fo(RAif$
zv^k*aK?dX9r!~7}T&7D=q^Py&7nzg^MrB;>vWDA7V`4pFn|<EAZMoP=m`_8;G0YIr
z-vy%41U<6QLxi-4NH9Xo4aAH~u~<ik87I0yH=SBx#BCW!*y^QeQ*XB8ROz{mP(S7~
z{{T7ii2ACb`9b}CW8=0|E4w(tSAKFqEO>~h(XN*58*J?w`pvN+v8_p^rx8J!K(m>q
zL$<fbPpWo3)3GX;uE_Sorbz>nHG$zN8hGlwA4z^~pSD9CABdq!r(oH8zXJwmdF<a8
z5EL0gvzgmE7(Ww^R{2fO@`9>KRK%E+-I2r3HWICwH1qOr-zg|ws=fJg_hm$V=YlGl
zyh9fAPY(=>9HwKVszGdz!%41iI=!mEr-H*URIe}GNQvqhA<_A&I5+YTt+-nS(eTvA
zYLg90)5JRS?+>1L{E6)SA~@}a!C|mf4kQfEI=uSbK$1&GnJ0G@%DyiVif3Zum7_+D
zJmMtz05$<@O@J6oVN6LdC7!M%sU6>FVwV9Q@HL&zq5l9&e^qOW{{VkPa;87jR`#=W
zZLGDC4j2%<rIK-XcVz+=2r}pUphC``Yu)Y21SIKXzaLBpNz%<XeQ+Tm8_2i!_dtX-
zZ-;l>sR9x#;@#8mfe9LU7w*mk#gY^=UN1)|5xlJ%jVjE`<GFC{8gufJlAW3LZ&OLr
zDTg^-aSr4{{{ZQXd{t(%iU1rMLt6)3OWmL#kzfwP$12go(=Wl2M28b1V!#3KypC{H
z9%miVF+9~icMwQ0?`NF8UR-HnP7erhDLOWK-Yl1dqcXo1;gy`)n8Md(IfFwan+qF9
zIlaska-IhlNIAH0M2IB1J=X+mbdhy#*!Y|CxVCJEG9Z)!a_+lan*PptMXm(y(YS|=
z%YJWx<)<r3<HZ4Jq7Pq_etO+*(QmKmZ|~>6H^7yZb&WJa+&;vFge7cqAJo0jAYBbx
zL`;x-`9Y)tnX$}LNwl*i-Q5w~l+rR)UH*<q0umLy*(ZMe&><$pOS4ao_!Q9WL0r)}
zdLkr35>!(x{TsPLQjjbV)DcT8ySE}P_!2O5NwVe3#m~@?CNx`nVA6DIhSf0arY5J8
z*nP$)#(G(^u%2Ng-ck>`F}Ua_fY)nnC7VzEM(nR}=>F2CNZH@%C4H3gS*8<yNjor?
z`7|ZuLZ^sZNpLHyeJZm$xh1xM=aq;t;~wtgz!45~cuDy!!3Drf*nC3P8&amQwa}?F
z*(of$63T=&+mn|s>&x_GIpxHh-@7X;v<0yIu`#9^nA+tx3Z@#F{3%Z+`nzdSN7_kO
zS?sk5fyp|{gi$zq-0R(xUEtV!LEzN!y`T*k+90+LY3&%$XG`XvEewJlI;A$_UxzKb
zu=NNtwnu6?g4ok=bSvN+fT_!VNgqx+mE$g&7}$hX<~esHBp{P%=;6zB7ea4hrN}AC
zU_uIpd$LXZ*C>h*NEH_?-pu`w#ex%Z6<zAQJ<71<El>b+`@7;NR0)frs9wLvoEs*F
zLa3XOZTP^16%E8HpIjQD3B*-h>7UmE6ITThGfwsHfibF$r?7_2YGgr38=jceTX;kg
z-fI}otxE}By?TMM+BiEf6Qgd~9<b1f?+v6%zS?fDMx7=T>H;Eq%=Ge+=Bb#TVw*1P
zyxk6AmnK28%w-xC0TKcs6<aNHi#5n@jHmG&9ZLDbq}-F2yO_jU0U$(a7PmEdxN6O&
z8tX(_ZyqxM@3BF^6OB&OC!PjuGnm(OQFwI>cmC+x;g6?QjZ)3r4Y`qjmxm>riv_h<
z4=C$nWp-5W%kNmwRN$Z*WVtG5mwYm0qUcM#NfA4tCo-F=_hp~51n;UFh=`e@@cL7H
z2!tXPTT;xj_WB?~YM`pQE`0l-LJ6j6nRI&~LT*8lW}Dzb3Zky%Uy35Ygwe#^!k=6t
z0>KHfqUX9IglP<hrA{Iq6q#$E^@E&3ulFx+HoN3h*g&end5@S>?!rt?R_A2iAaQ42
zZf(1D0JOZ=yO+>Jy5CmsS}^b(z8eea416`jgCavn+`*93B3wOF>jKjwPZTKD#F{Q?
z%mtvqBXB0tHV0jZDsG(V%Y@|o)Vk?1-wTZS!MylG-V}>?y1CfQ9x1(_X!e3;S(T?V
z;oU<+d%$|!uBy<c8eNH^;QWm+E*+TISorD*xUyr5DXT4j@xv1i{{XCe19K?ACE5bp
z;nnJ{{$HgP@zstgjIBgVL+t+m#NrRAH)xByM|&~>D#j9*VNulT9`SFh$O!e0fSy7~
z8o{!N&T+vPlEVT^^Gk!D-x@>HAQ6%C(ak>boO}YLeHwC^9$@_7{L@!nwFq5glH|2h
zg#IxKuFXu+#!(;`n}Rv(EpRh%K(S7nNO*o8pd@r4_-IFzs7i%$n@d}!o+Or3^JS#r
z7qmKJwJdfSjx=i44~-7zqle=h#0Lx5g1f&9S;MrYPcJdB>JQ=gjV`k<N|>r$J6doX
zl4H-eCcbm0i8D{jm$Q<7xOLp;&@ZYcXDda;ubJdp!U@z303{S!*3}H8MYVCX&dmf!
zowquY0+bHM?86>lo0R#xiOq9!ir;4Q-MTqa#2R=(QlrR#FO+2-&Cak0)@1eCbtwm8
zGSeE#i8+q?X(SyYJ}RX@G$)II_xb8zG*^E`!aqxaA~;9RE$9B=&1s3T-s5`}EK5Uh
z%>wxjUoO3LUsW4xb+PINwOBhMt9!a~iD`5$U)uXKPvbYMlJ40pg>b9&O>HaeoS1Nh
zpV$H~u^v;8R24C}d?ZYXjV&f`wSb#ww>Vo%aMf#zXeZ11pE9)$0mGA%xF=`NN>)r5
zn5@lCl(w_YwD8I^sehU~^WBzy<Z;2$uZW_`*^4t&bE7eX*|XLePH7rk1IWddj73T`
zY8pfbLmPM*hIbej5e7NOF>t!otQrhR$x7lJ7d-nix=(6Tu_VmkG~;9Qwm9gx-M;TA
z-Sl^cwYaP`S7o?!D>?9DEDM_9<PMvQEWX)}bDf8;Ois_F8zYo};C#rHNjlyOq=Azq
zH9to-$-thNtJGwUSI$~O;@#)aeQ9wu$T-8kDmNE_qD>~f^`uyul^8|i?+cs-zY_sA
zy9TPk!{aJLoEuIe*;wFYNGt)R*Zj~YM{K2Pn>E1`GPL>1h0eWEa`|R3vyV7)Wo_IX
zWOF6IsO=F;h^&a#z+h@uq#j|<@5K`59HVgJOP=C;tuWS(^NoUR#$uaE;etagH3LAC
zwpQh9S}sVJIM_^;Zlqw0LgbX56|L0CibHI)a<o3oB*gPV-<;GYNOgF+zzZ+yL1NAu
z4~`5xHxW~_$^priTT#2Xxuvc$Z1x9|*I1V{iQ!(p5yWGtv{aKB<buPqLtGr)F4q=-
zMSu<xwsT*EROZ-}e+i+<v>I*vkm&ig{Zwt+ZTJZkY7a}I*0I=dV#BjK4vcVbxDzUY
z7;5e>Wn(Atf`821)8+zFS1Ca2ZA`vW^Q+L!NrxoJi~a6H`Y?J<8Zy*&Lwh_HAM*eY
zkv=HwF%g-K&N;M)4}_n9T4TwYS=RPq1mR{*Ri{t;5ud=L`zg-nQQs!3yjIx84oq*&
z%c2plhUnQ@i^$gE_CST3E?wR0yC@K`r=QV9+^`{gNB6*mi*n`lz=V}+B;WhsLPm~U
z;6fU+`b)_(fFUJY6<c?FAVL`iN$+RekR~~#DCKuAY#SycWuP&QF9lcY7`SS5rXIwt
zJTk33<Mufy!*a}(+4vIj@q67_WLP-LTsIY3xW!XzoD2d?7PYKx=0r=00(y}oLe#@M
z8Ag>OMZ`7<Ff2=)U~e`PZkhs{)pi?(De{ul8h>lKv)7+#nVmwIV7lx_v)&xV3z`kR
zTjP!+fN-HrwHhB|FAa6_opQ|jwjg^=9Mq}tOk=7RoVTzf4Sz8a>l+dyDN_3*qYWr3
zWSiWYk*3rQ#9kM5yGxtC%A|JS)ER^|>9Edb(RrBm!_A^C^-<zA9bL5C?VVA+?5l66
ziW{kKG|g5wG+CF`B*9mK7l*$!$`V^7t7wao?|}49ZiGOcIrrzfBXuN;6<2yEKTHTi
zvn0zeYzRm;_uPryxpUtFU`iIAYr>uX00o646(x%Yr=I@+Od~0XI;?5_zW7P%5L7Zu
z(0p6qK(I)+G~S$(lB6PGHMq74%~8#$;ozc5vh8{1=H%(p(^uI!IP9~4!z*ubIg)h@
zVSS*h+3-q?KG9*Fn-v*-4BUtwI_R``TMbNji@-MLEp@&Gue2Uc)fxjHOM5-88?CU^
z?jpnznjtjv%sDpgQOj*Z8^-<HoFi0D9YT*A5A0kVW(%*Uv}gCfAhylwg3DR4t%QF0
zt$npdA8)q0GDMmVaHYV#r!HB2d$_$=&w06(-Ih03-zXKQx?hBp{2NM`n>~xwr0TUv
znuDWO!Hgv?Xu_Tn;3CDw`HCCblRKsyXo6dWkGBi2hh-`pI;BT6yAwstXb#Hr+c#HT
zr@9>A)0D^rO6iKHP`a)mjRZE8KuIE6J2x8*M|uR@L0_J?8C--_S8(q{kE_b@GoUG<
zHo2Bd#oU>}WQg4b*3lG{kv=e`39^C;<)7TJN)X*2N_WEWLT+M;S1R}OgJ?nmkBcwY
z0uYEYE&h$m0uyx{)0cmy1SAk5??L9fzl;dg1i)_1Q}un&CkQGEoL=7eK+Tb0n-J&A
zytgC<!2wfNm@6?+R~~g=c4}u$pqa=5?i-WTTY8o-s__0FkH$W=cDT5``Z$0<vqVLK
z)yI=g<OQV6Rx1rbtft@qYjy`vN%12<I`k?oX9mkjQfFnQ9dzDw!-!tA#FV_@C)NH~
z<6DE^aPOvR2Rkv3TX523ctnA?vAh-1mO8y)?9d=`=0Lx`<bJAVizOo@2Hc|Os2Zl}
z75d(;{{VvJ7w{DD_*##{;id~(=_HxVU2=CYWBG0+R>Z!oW;Ug5iEfb)1-=vJrHL|J
z(IZJx5^j*KWSNp>>_`*UH&&M-L{5KH2tgqvPx(j?ghQU}x%yy2YC@Ujo}W%oAq2-g
z{GdW?NTit}{m>x@o6CB4a#%-H_D#rg-irvCLTX`f<hgR)lq3-esoQS<0KH&I2uPc8
zmkByt65s47iSGgr?#YV~Q(0MA*}PSZDY7!=bAyOy8)3G0pH=af!=Flz3w28F4Gktq
z&<Gm)^xr@ZivBUKe-I6(36LNSchke+tgGbZ2CYh>$UBDWERA+>j(GxIyS15W_S`1D
z;XAxPohG9?qj5OANi@bvbfKU>lfZO8aM%s*J3v6G0G5kvbs98)aSnY;%Xzj$wCG)}
zz`e_FbE?)*XJm0msAMluCdEy+MMdoF$i4ON_+1@wG^^ogXQ&-_HqFV|?{e0FMXU?}
zW;rO~>Qk#*xfW^HHf^mYXW!4Nf~3!hVfst&gInbrxP&e}!S8i_nCVKc70hABwxFHM
zs+ReHKphCoI<9YJR=D~pWwwR~U#uDUADSRdQ$=-In`=}@gG_)={{WZ%ei32c*;K&n
zDbsm6dZc)lI%D5G7Z&<`Ni|;Pf#dK3C*+&F+_cc&JV5bxi=nCmm$0+%i8Z()4{GMT
zn^+hFlS4s|9eRY1q{s|uWL!xAlgyGt_-W*(<Qb5C$0`)ra%57{b*5Ui$wIqvf|p3=
z;_&#mcf|}-a#HNg68n(XMndTh87dn~g^d3I)QAH;jj9AXp?7E-H!Yme-WP7RK0VWY
z7OkZnR;49Np5q6s2QF@8m?1yPbdA~GD;#Lyu~@t_n8!FzH;lZ;b15eBbsJ1Bm1|MM
z)&!1b%`?kcjR4sap-FyFXf?McQCenu<nkV}5p-_3SMQ4rZG0aPpxVS45D4*uHUruO
zM}p#APZ5NwX~=>N<HzW}vuST081;+Kol^AeOZ^aQssw7ddu%nCvX}CcsJT4v$xMwV
zfXf2u3uQCrwGS^6*s7nU9!rU3t;R7G&;1ra@i|%baq?b+12p(;Dqr_s@Dk(l1KgP4
z8A*pFc6m*bls7vi4q@Q)Q^BU}%)qu;BYEoc<{CNl_e5}Sd$820(WdH<$A8>t1@$5|
zIsX8YC^{22U2zR)s`FByWR{O*#D=nW05_&-i@QrM$7=yn=4Ga*v28)=xl#>x8K0VU
zsVQOVyG)$4el8KOy&jBUXw_+UInJq=Nog)?0M_j;Cfnz6>kAdwy0qRKglky_1AuFu
zPe%rUeoSZy$x~%99H>iCso79ef1ghi?zG^4M%ka7`bWd+>ckBUJ~o)4gro~}&iuuZ
z{{Xha=zl|I56z;NX<G{Z(-%<Y{4Id*=Wr6q<#(OKRAo0~SVIpq=it<$gu!*SJu>t1
z(>HMy$RwI~WSiQzq<9aMHgLQy5ASdwOhFGO9lT3pDZPt(qe9rxJ0D*W+{GJBvg*_4
zIfp#g<mpqa;rul|6BI+Eb~^HgwwF__+xd3arSbNU(QrdNqE5r`?07T{KFRI49@`Hw
zB!@R+1hwP3NYEi)8-;gnAKEn7-@5LHcAh}NX@Vt4(kWDVk$Og{hb7ytFc&=7Gv?h7
z&j-3+sd!7?oeFq5?sW`B1~u%A&4~mZ9?8{?i*C~_bU0o$onD(;R)tp<yFyLD-LB;3
z`<x(+;%o4C4>klRK$cdAjDV#ZA&&Jh*cRdHeak&pdmxPt#4b&(iZvVej?`{>twrS6
z(f<IVDig($YE90$)BQA2;37CPN!{=I52Clli9ALQ&2QCj#0}l?u(MLM_dtc^Z!UfC
zA$v#D=;s0#m95+TFd-*Tb*uM4gn|sd?@S0tR=4iy&zeAloi9N0_CSP`Z;O9&@_`5>
z*)Kl+=q(CEM3)-PEgt^-paS_)0EMS1rWCmxM_PDC<l5FCnl~RI`G*E?v$c1+K-wp<
zewPkwa?xc<&H^h`B%4_0%Gp1JhQL0_JdLiN^O;N^?$|ARg51VB$-gnwLD>_81mx3m
zMP({cZk=$}+rOg!06*w@(Q2cE%bN2IXRdnwAFGa{Lq*9|+6V#KmuNNqZ~i`KV70Ex
z9GQAa_T~pI8eJu%0$wolF8=_J53r**gqOLGFWHnw!T$gaU~C|13g$dCIkEVvAAVn;
z0QC!Lt12!KuW*j__C~s)vhfB*pNBm9U`4k<AVp4D<+**>6C%k32%P)VxgsZ21S1QR
zdV5nOfi!|`O$BoH<pL0t5qtg!uqF&4s34A4bM|^-B}rjHP;y@7-Ja-4j%hk;ox-0L
zfj#97Tgf-Cvp5q%P|b4RzjyJ7p=>RaEk;{twAz%8y7pQ{Vr9<6Ca1bnho`)kG!JP;
z4<B<|fNNfLI&1CH+r)epmkkaz+~Kg-`Ch)iYzt}ga(gNAbi?z&^t8QFYOyspBI6B?
zfK5$I_E6z21QyYS%ez$W;<+QnJ>jf%KBUiypWnG{v}SN|*Z@3&oZ3;+P-=4(C(R^T
zY3JWrWx2-N7KpcU_{A-1j}_F&5)~V0UQm?kj?Hk^&gN$we-+ktRap62c3NSE^j%^?
zdp0f8)o>jEz0uUmQMngGrOWK_%p>1i<EfDp@f08sO`Tem+GYU+fFOFM;G2k^D!Y3Y
z(RQw-P#Cns9|Z1q_ceS+AyTz{DZMj~sp)0n?qLB#B+vE2PAv#ITDIliyLM0_+NP%1
z{hw?JL~5E4ExXsBHG<HUf*ArT=DYn6AvRG?9KTX!0wAFiY3cn?Avd)bdMpUl2}iwG
zmvjgSWF>DT-3xe3%-J^~R88KE(6k9At|Dd&mw%!W(J{?eWSkaestWjIFHSnpJMRfw
zrm;$aZ5s7l@%`mT!><wy1|CMz-JD6dF*d5^sZD?%QH_z(B)jN$8r({J<2|9N=ao{>
z0n;U1thK=_zW6asZE)ARJ#ZV#c|AdN4*}H_Nu*0g?Ig#P4-h2tpM`pqgE2lcjY}Xh
zvVzesA~Wm#-`b8?OotH4w40PZoIvK%4uZnR#Ogp@F&VozG47nN_z*|TD+&%<vikFl
zB}fPfDm%A*@F4)jc(Ptx@<<SyxEU^Jocr>D39^p-vq%t|k`{ZD;JlzhZs4W4z=YjJ
zGu)6N2$3z}&v$%AZy^b>hY2eBJusA@ChiI*<nLYG@TN*PgxS<3E0@2TX9{3h5EB=X
zQMVL(htcSOY7&DIRMl14sVcrNkfTe8%}-RG?G7T`^wAefKIr9c?)B=74j)PNT>7<K
zB9o<{f?CrTd`Jb*AQL%9ssSXed|g)4BAZ%e;7Jj2e(k{^PQoPXBFW}FbfxhZ$nsOF
ztWld{mni#7VFrg;C;R)>FJS4{sfF~KkIUPV_W;)WdI=!=0ire*x~YhLHdpLg+3dh*
zIi2k5q<d0rx@6C0ttCQq9T2W|IlkD++S_yDtR8QVs(+jw{{Wm18b=IWDwtd&nhTuJ
z1Ygx4ix36cc6YLDL_+U|sLI%uxQ4aTX5O$Mi3a2zW7_eV(s({$AH!7E2Z-7dWz{sy
z7_`gu*DSYx<V6OIP93ypvSQA2q=9_^HYcWoQ>u+a*;aGvKf$5_@h4dKlC3Qavj}C`
zfbQn5rN{KM@WrO!u(W_QL1_clZH@JxWR75qt|t>&<o1y1*y-=oc@$)qlY9fRPN+YZ
z6H|mBKT(el>Dlz6S@CoS-nX<I*=KpM*#7`2{#3Hzs$QAMc{=0C){o{MX_L)Q(q|bW
zY}%a;Hpx=oyVgv;sA1S_YaL7w`LP@L7MLTzfOqPmTU!p`=K|hG^ddp`sM%>r8NjI(
zA%tz&Q!<*8Z@?wzHN)NF;!h9I>7hmA1C6PKs!S3xxSuc{PUkc<=yWi3wN<ARhh=t)
zfMj@B$GS-6ab+J@m42awF>JK6WU~zftm77bb+@hl$t)a2%%mt)d5C;RITLR%5Bnno
z3(Z<%cKv^<`lbzTx;;9>h}9Ju8-z<OHGszG5G8!Z+I4N(hR3Osc78RAaaq@0BI+x<
zK{gqBG?BIYV;&l%YSpl~%u~P}rACVrt6Jv}Cgcp@$-cx|SE;8-l<MJVG@j!eZU*J8
zlN-Py+e+!WAITUG8&qjB(#|||xoHVEns#IPovmwcA{$x_V&+{b>-cslJ)9dB>^p&b
zSQs_$FouFdrox$fy<-^;vIruLM!Sko?L;)VyP$9+v_xv~&@4S+q!rO@;kT#VYHt+G
z1jC?ujL9}1h{_kNf?qyf^S@vljnRjFe)c?vN3cJ;9y9&no!#BsFlw@UX3~*_1-#sS
zbIc3F2SMKE?UlTQ$geEai;JtOEo*gJ+a;jc?aBZp2vd~V8jQbYpO!;q4^_S*5@pZM
z3Xc<I?H%z7v}s~MZBJ(4i#ey8NVq>e!3u1tQj3eJ@2J!f$b+c;h$;qgiiJ6#Dx1da
zl)s&qtF*=QRPTAf51t=gS{>=u*mXq=Lu=D(Ux;)Xex+b1EZB;y8Zr7jx42?RV}MrZ
zx{HZzCi?!hB5Xj=({c8iZqR$fNpK`4b$Z#(;GA1KBtqR<UYRoGpn9M_WbANPrO<OO
z0{-$8P56gmIBp=1yL%Yy&GoX9(m@+JjsjTi9Md|qfdNEe@svUDwayTISpeK0!sh7a
z>O2gpmy??V{)e!U4tfQyO^I<>I-Z{taNt6x_AqnZJ@>_;25mQ+f@a#nvBk0;CEXF+
zWZzqB($VwXZuM_>Ol&N~q-q2#>2BTiZh;Fb-%X<eE>cyzm#ZxKB0;J(SV6eAdVLTf
zAww=Ja^2jJAthVQdXw7%5)>|8E!p~31R<br{rjLyazjw!_`(R1vH~r-%cbElxqaAv
zsSf04-E<Qo7J-%*!)F*8O&)Q{x|34QtfDpra7+=nW0#KW%W-qG{Sl1d@9g>%n^2z6
z1C{}^q?-txOaf%y@Ch1C%jtC-SO!Z>k#PWC@ew450@jsflcGOJm<nA*!C6FQ!lap*
zJEm5dX%d!qn2$>LrQReZD7J-cJx<24*~Xb9nbhSuvjb704DJMjA;h;;r&YbBFm)NA
z4uBCENNF~WMuyN6eV5A%bjRII+OpCQ<f-8%+#%2G<qLxX%2irKe|AOW!<y$`t@gHD
zqh$Dr1=Pqtj2|Kg)8@8OG?!+cu35&;@Uv1?rOwpq<h!908X#(gq_^9edlx8YOQt?(
zWYXuoIDH)AGiXmlg$?;CbKe3I40tnD?%s%yC#{oo1zp+q<(_c0PAv&4370?N2$@Y`
zLp3B+<-0tgDOzDAK}WxrE&HM~s%R<-f=a8&JrGIV2xuFvF8*r?3wHTFOB35hv$bim
zG&<3nn%Y`Ik}hY(;N9OC+<PhDk8AgRA^!l}3jY8OpX~<5&42nz`u_mrVdV3GQ^~(&
z+d`g`o*YOOtV|NJ>gcg5%X0PEtpm4DjR`mU7XBFUbz0}R)N90Ez;f&KUV)Agl#WVk
zY1D}XJWSO>QA(crc6p|mb4=U<m(yMT6}5TAH<yp@%tG$a*FYg~cGU3wcLZamDBC<r
zh-`fJZ(J#(sDmyu+TLnwCN(tEQF27w-u$k+hBbIYu+GCc;$YM@=w*qR6=_}tfQh@H
zO@z5{#tO}Idl}2w)SV)7J}z}V2Z^%zq=J^VtLplCqv=Odyok^tD_Ji807rZYpqpyF
z*>B&v6_K)RVcw~h{%~7Fk0j8=6PWV!Lqu9o2vvFK_rR8+1zS|%&%2&bCw&lH1wzX{
zYq=5)P+<szCCE{CF8GM%Py{B*D2Z|>!zgpHO#sI<x7)cO6E{ZaCJCNsu@s|{Xy9B!
zm#ev8Pe7Pbl2s+AQes^XIqAI7#mp_%F6f|JTB9ma4gfv%1V!U9E^#;YVnF~Ii|DO`
zO_dt_q~&W*oLB*EO``jvE4EA2My9BAz-)|La4icHiEj0|EK_|t^cdEX3qzVgGbCmL
zNp{T|XibR%L0#JVRUBhd9UuV{1ll)017J)QS)#P^x2oWvnodmWG`sA~+!WvNM2^@|
zT8&zCo^*Xr9`JG3p=1zxUuDj}WcI$JcMUx)1K$S4PbOo%bKU8URYfAzRm{J13Ec$3
z)X2NwNj5Q5POsMj6IVBP^7L<k3AqX?5+~HKAq0gp$$Py!;L!=Hf~N1(s|rGB3b}Li
z#CoD5R007+yZ1sWK{q5tERi`crWS%xblig=&3|@KZiLEV2SwD~&{^iN2$3k-72cBj
zbls-cYBTa)Zj9j7$+wFwcN<~58}IzRsUC8LT6OT)oGuyiYdkPm=<?>1a5QLf=WQ0f
zW)@bVh^j43nUC^-K=UxQ+o<#InL2k&5o#NY`9fZz`{U3TQl!G-Lt5W(8V`gHXJE2g
zDpq)Fd&K<8lf%-DP1Y%5+nfV?&ziG!fFu4t+Kk+7O0K|Fs_<CYhPOvNohR^(^6EhE
z5u&+;Z8l<<)||u{CuV`Hb2}1w+nUuFB2n|FRagR2m~N*s9iHDuJAVFOqa5&cs$uZG
zwNCRS$r`f5z15m#fi2eUn3g7+YGMl+XIciwr$ftQrHm^IfD&E>o@vcD-5R-p0JMWY
zzlx<0V5Y^-FBhRADflb|%dEHALCobZxmh&1tG*G=r7Bg3bwP2VxNz`rzr{_JA4$1b
zb<|joaSAQ%DQcvxMvCb(QqpbLO2V<1MZ<4#U3X2fn^a<{V}lDtQ1Kn&0U|CAfS(v7
z0A*ZZ!_>k|<j7{zXdsBzCQh0uiP>|QrPU_o#g@(4DV8T0`L;-H`PS{v#E7IF-X0yn
zR=s-Eu>Of%yNtPvJ+O3eGnu;w2N{@gm?yO=RLySQleRMhx@Uc$?m+`d0ew0<TlSjA
zvtDycl}6fKVaz?V%y(Q8$>~Zo>&;vF%@Pp(rAXn4c%g%<Ta#V#f$=I-L8*wRNN(Uh
zKhUq7CtH=;ZY|hm1f|g&mZeJ68<7BVU`gT-2`OFn^eBUWv~l97vDN9}xNCv$R&<bW
zE^|TfhO|bj)R%TItBs#ln71sU&iIq$I<EQIRJ4=zTw6IH!X-J{l||5RuIBG)k<Q~=
z*ROd;uE4R4H5|cbIp*?aU6f8@NhC<VyV-gCQ{Gs>^97-T)`J<004;d1TOp>42B)V5
zA$0!$_!s$_{>Sxw9RVCB8uEKBiN9n{$NvCpk-4U9dyqs3{?ZaHQ{hrRRRG&Z^mUEg
z8{9tFzNsey7NiJ7Mz@a)JrIZ&1&6%~?i*maDCo34*rp|@G&<|pNzwc?S!s=GQvs?k
z%en>H955f6;o`f4J8t7G@#lEEFt`?%;Oh2oa4j^&^D)E+_eI#zIi}1GxB!!zXNSKL
ziSA+<nV^H5+bf5we`wEkE=VLB2vfTNl<fN*z%2&x?1q2$&PDDDBZ5&Ua`XOoD%?=d
z6n{X${!q6nwY%XO%RMDqwO~RbH*$ZB2v|n4Pw{~XBs=X-xgbKpH@kLA=z$3))?fKR
zjTS7)cVwQJ5Q3q@)7`s#8eI!2*=6bU?}Je(i6~0eN!{#)sZUj0#nXQ!VN6p~9;PXC
zkB`}4<DF-Acv_V?HOJi2N5x8;bJk;dkUk+%u)Z7YUxKYpRNXq{qaxh!H~<qt?7l7e
zry60P(96I)z9=GGS;Z_h4$0Ju>K7J>a3P`#LtB&pECA98kp@UA-Fmr&K-K^Wv6Fcu
z$-T+&wu;?}rCXk()v1fV=~o;#KkWP@LVq7tI^x4?;i_SW;Vvie0VFzq6prZB-1?xd
zSaNCAZDY&`=Ojpv!_UYIn^y2l(^K=6%Ly-TQi=|?Bl9QN{{YKx_F}gVcWV08qBAyy
zro-;G{m9VVF8Ide6xlx2TmJw*>pnV#dN%lhbK%b={qd)|qXrUe?_TcYT%bB2YKLhm
zjjQa4sU#ChiIVrGN#6oIP{=oKND~4=5w=OmKU@ijL9;Y0y&dp|keV3+s&49C>zpIW
z0_GAlb4<|_e{ifLbs!*0!UFe>ef#1%CX!8xuTOWv5-kWS87^KNy$~U8y@5(IGmqz(
zn5QJfy5YL&IV@p#%!<gkiWzvm5sUVNDd1HDza1}kXw`p)%ea7#!v6sDmG|Gs+M1HS
zNE<C=nQTQd1~uT9Y(=K?#8bQ|og$IqD!`^kir%Ls-sM8t?2({i0#N4Y_kTnYnBxvs
z+6WT9X>GF*;g@!h7<zvJ%Q=jh!4(FiF%V3fGgRhXmYO4Egm@+sa@lpNisx5xZg84-
zk4AlJb%<|mLo+gh97Sp{a~rgRc91}Wt*K*QNIk4`0XZr|q}#apwuXA?uDbl}K;RV-
z@b29oM>>_|Mu7~2gf2|Ccf@o}1wz!fb)KG3rt*Z`(%wj$`}2nzq9rmWlb7y>gc5Hw
zx6qx-oG%n1C=OQL&%Y=nAvaJZT$U7#kpTlsCAAU~5Q()#9vr=yK!nhn-aGrd))~1%
z5~a&-eejT*kY-tza6p8`MVBUKTWnDUvbjJ~+wS60`~LvF6mZS$s@Bi!z2h$DzD%Bk
z=r_td6{^J3WiZjJTYo5l)D3)Xs%KHuRwkvJ%KZ|#O+QEWq2p!vcT->nA;xsGhEe>)
zNAQPqQNuv29ODxr-9eM&FXIXTj3r)|CY<wARbQI3GH+9ezh8V+?=uxfqg*4kFd_)f
z4Ye`>=b^Uh!H3YO(l3(dP)IR7By!uqPpD<T+s$KLXX=uvzb|I(&?j+pB|*<LUyLA3
z-8hh%5SCdkT>UUgP@9-4rsd(wbQ8%5vxu4eVM$H|Chh`fn(inyK;)YmdK2!1kOUM0
zMDJesQdE<2L{+s%mwG}d(EyV>(1m_5AvRRYntjkB`&~46GrbXcq$|zQb!(iQq|+&w
zu@Ybwzg&KqDAQT-jI6YvOgvxv!figmN1elv&NEXqoQ32xYUjDdq9)d9c_hFRF1Bkl
zyN9rgNN6PJ8UQD(k^})Hs)^0E5`B(JTG2G~&XVTq-tW4e!HtH|d+H2EEQb;5w0f-Y
zpIP6S+<PEub6rubVWmJ@a&>n(<PYH%=Q~MK-)5DE>h%J7K2n@{nHC2=A!JWg_digN
z+9>5+rxlK;hS`mC-Qq!G2fVr7N_4*G4Em?qA)~+$=YQfBqouRiYuiAOBug5s86q#f
zAMlF&DE|P%IlPGZLhX+S(g8bz^Zg385+zLXBSEPu?ju}=Yd=c(OWWV#^u>bL;APl$
z=4PdXR_;p&E^Hnwy^jD!AZ4gw+5@w>2X)vx$=ekO`)VXjkg7T*VF@L+<;;HgvK+^-
zRAG>GBSJh^BG)y=ju9G!)y>P9eg3Qxh_VZfaJk-EXZt>L`>|6LLt4t;fx!F!0NHoL
zjK%c-025PrhC*qxX8!=iPlSwoUk~x_j&(zLg*Rd7i38f&LH0(cNIMX9Dr_eT;P;#I
z{eFw;z_V!Z_s65rVp=70=al_aDEk}4W60;GWNE|p<^YTKMPoaP)a<xbPt00AGxS$5
z3@_pg5u~4RKT;Pg;gap%P#8t&6IA>i!tzhwFQx~qGF_yL&H{lhqn<JS9&zSrytj){
zGz@k@#2b47#GBanro_qEg9~-Q92H>kJ){V}jxB6hgX;3q<_*H?O^f3B3MLvq!iiai
zB^qJ3K2_{<4oJ8zzVo}weuDYI+*`VMqjv$sFw`i9W18w#XbvIHX_+IuBm!hu^tqLp
zi^5i`Q#KTb+}c2PfM{Ty4Z`UPRd$6{J1JGx9hpzaw)krg#Crb#W;R({4Tw2t)N8hY
zLGh4p<r-?U9ZL~Qy-|g|V8@e%yyuiAeVp)h$E2x*lc~FZ>TieiUe5>he>YbA80B~O
zTTK1U&5ozCB)~uP_xroIybdSo$7$p{PoSyrc>LPj8gA<avCPZW;*AY--BPesik2;+
zQ--6U>OTCZ^S;jGV#JU2ju~jN_KQDqD^<ne@a4=B^4&vYyVdR_HG}xvIs&Xy!&j;r
z;xdO=7f6%&L>E8#T70!tV3x;N-8iSpore$dlC7Il6Cv*EWcS)JWj~^FDj)4mC?T)(
zb{*R##C}&8C-}0lySF~7%y8TQyL)#Iag=z0Bg$NP<xVkX9HC&nlx-GXflo$QN)<rd
zKBp!Rv!4+cQ;QMBVKAIKf?{dm8yTD4al4#kUCuW?6CDzziN$01l>>-$!LZS-vGqoS
z=65O1#pQIiPqWas{{Tie{iBY&AOXQ}_tZH50G-KkIXqUe<(K^pTS!~od})mp1?6w`
zLqf_A!*XuTAk_;h-w)O3feQ-N-W9cb;NB31jqcsw#heIU)7)2eCnZP_kgVnP`yfIZ
zw~~JKg)?L?X_j04J<uCMb^;0tH<FFHrtWD5(19B&KGECN!F9<tMM$RBT15mrdecm`
z{pE_dx%4WH9@IC)xxulvvaDgF(wI)teZZfhh{W@CM8r<jsMLEZ?a7&|Q-xzQEZ$^5
zY8lvr%uBoLD6fdES-7s&g(419XJ|kq!=hj`Xz78Z-F2f*o2X=|P!T$fL5)Pj&V{jk
zc2UjZ2P-&osWDENK$YzC8t$uS+Dc<jj2BZaKS%aBcg1f?b$k^{l$->$#Jc&5vYGXR
za_a`x5Evroxx==WA&q!8A_l-njR~F3ux?4&3uf)MQ@Fk6vPKoj0017A#kNhmQ||q!
zp9vw<>j7+~jUX8SwXP96*aJJs-9sN|Q!-5EAf0YOza!GnwrRDz5R=80<c(ddscJ5j
zr<(Nek9RmyjS@7j+^hHRfe1?0Ma%VXh?9FK2nGtgHuFd&OGJfbNpeN_X1k#iD1vlB
zCU@oS$#a1?goOo7+4J9*I4p=Dq^Kh9%ZegyRFW04MbFW>Glg=9w4m6lxl29yZ%yza
zVOvb|x0fuE1TDd^d19Hzld}DwgIf+)8Psh!+Y^v<7cXv082oU&aC6yf+ER@*?FTz6
z+&d}8!|&aE{{V%4h&+#I<0IsWvbivM#~W_HjX++RE1hst5>Mny?aDk8$uw`e^enwc
zc`7E?<p5d4NUbOHf!~+(Eyin%x=OSxzAkOS*H_9mr7S9*Zv?qxYM!*5-g!r+;Qs*q
zE#oOOT9xEL`|29Woua_%voWx7O~0`J0Q5h}auPO#&=P07c}KsruQoJIh&)O6qzFxk
zywsHudHNti0=K(0ABJ$ZLIJg!?)3A{2|{e4l0qkBAh-mJnJ0I4_((z~)fAU6hUgH0
zp~aded3zv20X8bXz62(&Hl@vZruEJQCgusB4@z>$Gl2=Q1;%Ql$v41+!e>iS#|`UV
z0^CU%hNTwMbJ|FA9a7+E3(btcEjI4SgMC%d*SYO$fFZ$>C-RGp!5Z4!?3gz-Ch5)7
z(Ek7*ElbdoFAoWPGL987Pj(u$D!?_)1d%hK=Kuz@2$?6$%IMk*6{)_NE&zZnxjIJf
zlP67dva5-sRSJ^KIE3p!*`uunVXAiDyTiNbC!8)QLaSR^xu|a}cq2k{0)K~1G#RdL
zq*&t0O<+1Xr@%DFk5YAoZZT0@^5njZ<7~hrEZ(WPf-2|j^h9PtcAKW^DW6XaB&4Aj
zE_vn71n!y`q9#d}{{U`qBUIeee-D2pPuBthY(tW7UjFogLTC;qOV8?6gn1!0AzbqG
z^X|416L%31xqML|LTp^F%i#iJRM^Kc-MfC}Sc$Pus7;j<yT8*3J=E02Hj`3WR_51A
zqyZ}Yq|&{n*%OkwX4&1tX<D3J%hkZ8Z=Y*xw~XQ{P(G$&U`dGpz&1MFwj^jliPUkV
z?jJ?9W4szJU{2C@@%D9yRoR&*899cW5=pj<f%#`cMfQw(hc1#>)@65whgEK1bbBt?
z=3DE5*SsdvWkuAQT@=h63teM8gS}JcLET8=noH6(DGWHox~*Y}78!Y__|LRDI^-F8
zc>e%qIN@tieLCDuqrhvNOp*X=8QpA?q``In01&dBW>gNPgJ2}JqQd04(&G`zK=RfE
zt#>~vQX5Cd(ywGB;z~{AtDi99>Gs7wA6b;mK+xv{rdE?|#x1UgkydfE-r~YY9wZHg
zf!u`oWgc*{Xi&F!3`#V;D+f7UP&cE2Y5xE)3W=j$EZOjb@{~!EH^5t|0R-3xgHO1Y
z+$qQl>hJFNho_cPP0qw%Vd2+v0UhJ`n+6*0VA=b_pqYXm(Mr9d!NeO5N0{Zf+C`KD
zh7Gdx%bBYT@VH_(Tc3xA+Z`Q<!&9O$4txWrGItTIjl^4cNK&lohLmZqLC?dNFt8vT
z1p7n-`L7?>{vW02igjCv#?f=@cK{L_x$5cC(*O_ESf*)V>NSk-9nQnU_o9z%^_gYT
zE;|P3xlD&rV}~HCJVin$a=x*MT0Zn_&6ZS15V>>8I}c4`+1-)s&Q0Y2saD^goqPVt
zB`Gg+uYHj}zSzL~K~z$}(_@aVpKyWkD_evxoKxxnIY-RT(Q^*X>o}sViT3ri3S$h*
zWguW}J|8zI?TmwAf2(MuV_?Q9m8(?47M!%pTuJ<H0!8({`z=cSgQ!wvdGEYNy0m7$
zK?}6-Z{gJs53wmxt<B3a-{R&$f|;Fb%`^-a_8wnTE+CIZPOc@C4LnX}10emzcEE<;
zBi(2b5I$=m?83gNbsVlced%MaNw9N7bF;pQz|nqHmVTc_+O;}t(ZO(N)fyatXt&~R
zw3*DMy*YIM096sSc8WM`O*+tjyK@*`2jOOSO}1;Gav@4zNyepYJL&#*x$sXnWDt6h
zeea?c9;eS$oe%VPv>Lh88seCgl#rE>mXa%oeCDOibt3%8yPdo+Q)#Ml@XSR=k*iqN
zwW4hgFD_u4hyll;3BAUc4#&}BLD0Op&OKTi4?N_SiPVCk<z(??Dc-^u;kvdImeu)E
zo(w-FO{0-MJSI$W*aLFdJijz8^wFgLBZsC~+TyJyk_Z`^#)4vBQIO$INq{yISSgVf
z+%Q$1DjHpz+yoPpxyC(~x<P)IGjo(Ux>o9>(5kf8BqnQ^eYwY_O8F0yCYT92Z>yxk
zmWzK%=Lvmy+NVbmM1zT-DrGX!P0S2$BHB*Lc=pM2j-Z8@`m}1uqgc}}5aRIc&gLCa
zrNj<R0i&2YtAF7oDWd-XS!>H0{{Z-ui_8z{Tc`VRXyY)gy>jY*#@9Fe&|CgxSJCUf
z^m4!WT+{H7?}DIhx^!OHp3&y$(*WodFGo>_1;G$ETN5rHbyh3!7t*c4Yrd0%#0#{7
zJz=<&+Xl*%_>~q_X>If%jgK&`TP&6)Hf?}{-~AOY{g{^u5y0@hSw8Yl)f_~h?OZ_o
z(e3~%{+|B$(CVorkZ$)<Bn+Wt8qb%flmaveUeL8S?%xQJvKBP+UEfw~3dZS(-FaEL
zL{q=I9N8}fJt9@zp7r!WGE)LpRjtLJa={=YphG6IyZFKpH$WSdHE5Ik6(A)z0by9r
znEZD_MV7nT6yV59;PEL}94GN5J+We4yWN-|)O<k`=>kPOZBas<69v;%mI6tMbOFqk
zfzw?9R<Nw?YE-xo$K2Qn5hvb}K%`F9D%F|%Ni3-FthQ2wO`AL;c<&#av(Q_!<Kivx
z3JiT}udRz=2fN)d7Lo;~LuNqIPHUYdQO%`Il<@A-Z+X|OK{`&bBTZm7vn=&D31dt_
zOL<a$W|uQBG{P72;m|)j^x7Ty`UcYQ$37rC936Z`FKw<4X*P6^SZ$et<i*<{=m?e&
z=fqQ^P)1h*58=2m;O!%w#_DGXQsrsYJXa?oU+UmWZ)P#+b*{(N`{Gz2%6N=jI$8rt
z2+)H_5g#9VF(uEXLGRo?e4ih_O6YRxo72`6%Q(HFsFJLXe-z%36A8iHV@p1j;a+I4
zr0SeP5w)KD{yC=TNuU#1j{g9y6>@<K3eq6w=}1yEQUVrIRo5GN@EqC&8;b8tM7wTk
zS!KDCJuUg=!Jb)iSP>V5f^X1S@6R+$pi1EjMR^i?GFTHDLK=?z)gVL(UeQ~*cXG}Y
z@=i7jWo&|eew~fwn5I%vW~We<n50Wf>*y<vz9Nadg?H}tV;7IS(co^x0XZK@$FG0H
zZbwyrhJfsMXEB+t9|ihd@>QnF$oVF0l2V^g7ufkr7Lq>Qw(2e*p41)^Q7aEL%u6M*
zJO*=93d)WdXJzPG(8}*_u(f8`WkhAM0_FK}&;I~)+-JDUm4Tx8Z{;^QG+n+^s*<$B
zc756lbn`ty78z~ECOWOmg?(FubM>QlgyL#brpyR~+SRB0LoG^fk)R;J1j!-|f}@M1
z9^+p9#y^x9V}`uMpL*pzI{}^iId0t_NhV72qd<T`#nZ|Sr3fb4cPtW6o4DGcZcm0u
z2??TtXPU1x<<0~qra3N2)7=6PiL^w`3%Ud$6KErre;5#(m=d8Preog%6Jiz5FWitJ
zHXtQTlfDG{rsN2n{qQ836KYE8XV2LMk+QEfQ96^aHzeL<4?T71n73<$2ut^8mwyar
z(5UMF0Cdk~?<ZIXJF?G!xDp95HWDsWXs6zHwAO*tWNgQYwCDuxVJdE%3d7_ot21E|
znQtd}cW3_q%gQB-q{m=t<}z}d+?N<210LW9tFVG*G%7#RDUPDs3;yPH*H|i~+|tu1
z&D9@sQM9Iv$vD_#QyK9e(Ov%lvlMF8YjCCH&T3K!1{@tSCI|;ZxbBFBX^eBK28(y@
z3|NSZ8^`PBt*sh9%)80sY)^9fM|5n09MJA?efL?mY0JqZg$JcMq&-tj$Pkx*-m3{K
zAQ&S500=ivRN27ao_x{*VF|GYNq48cU<O2#CdQd37VZ0@JJ|`b0xIubUX_DXCZ(Qy
z`7d(7gxG~M$v;#Hfhn_=s&#krg`!#}r<oyh=ev9(3PVH!$PeJYm^MU!Q<|RkshiV|
zMqhSbsXRZzG3cD1Xe&=AAM_QWR1?B!6Scq86}t)1AH`47X6mwe)8!v&vz>9k2IiP=
zoU-0YsG^CGsPaw&;nTFo;y<&)*@N#sp#JPkhdROb1Ts#b#&dSwQHG`IL9VVKFplxu
z4Oz7GKFBw;EY=QsefosWI$?{Sb>f>RC)+BM*B2XQ#w5o700VryL{XVYIJK2=cyipK
zxes_#Xm6@PWjcZGwC8;EYjjqe<}|vU8k&=sKrSTe47s^S<!oD0R*=LJF3D{m8F7UU
z(TA8VH7`5ntOfUDwi>n}xQdl2Ij$}yO~^82mdRlzbP^`^n<ZYI7t*Jk7Jw`_w#Ea1
z{mLWElDuH*ns4C)0YkU+V^xmDVk9>Z)65Ebc!xO0m{c4n(LL%W`V#*Dk!S<rrX}TP
zSd?O-plwrJQU+Xc_dNZ*R;B5G$C7x8GjS}SYWg)I>aQn8ifsd{iy1xwXTU3TO+if7
zcnojmE+qH|I{yGg2;DB3P_Hpiaa;zK?LvGdJ57gWqFGycC+qbGx+E}FaMZ>j)w|<j
z;3t@uNX?`YB0vY2fv(EVG}_2%wSo7*+rdawY2PPRK1Qok8fte-v;hqElWsHNcvSNf
zA87voYjkxOhQAJLGZY9mA<b!UIu@LTrg`nW#ngKzikVFkH;_zt_om<N3b`;eHXM19
znVAPf{{ZaZ{{UAXgDu<Q7;H`>(ZI;pM2Km4JojUU{{V?P)LGjSOiUZZe8KoHtES%X
zgnFfdxq}(v8lxUiG+IoBNSdQaVA;94u{CAJ!M}Z(knr>Z`*h>X@rE-U#W2`6=m;mJ
z%ba|g-R;qA@Zxaz8u+2sC!hd(+(PCpganNC%j(Y6aSoL2V-wQl$Q_zwKNr69b2mrU
z_rE6`ID05m6>Aq|FQ~@A$~&T7O~KH{-~+D8)+yCkDN_y18rQpX-m^picM>$;qVJqH
zGgPS2>7Rn%C_L+ePn`VRz{topP8m8`E{3c9Uy%&G_}!-G)nIN3)hd2$!IZ%!ePYQv
zpHZ&gMUwWlk0%x>vLHrW&mai(b3=5oc^(9(mAo}dCTZ+#BNI+lD$hIGDfGb`wG4kY
z(EQ>LtnnHrtPobiQK?O>;`QHJOtqtOHfR#}0nNuRFcD6@M!cRGMYNz@#9RZeW9l|}
z0X(Hv+}JHTyIH5=iK)4&T>Q85MyVA$rrP2Y8=U#Rq|U~I-+jp9-Vs6Mzr8}l2C?M2
z`pNYhKZSxhxRue|Myp$pkz1K_Ko-&mtYh&!&nAJb`bQ0p5|)>isa0wvM3x^lQkShx
z080<7<6!5R`fc1kBhiZ#*-rsUxLjFtS}Y#dyi<(;b3yyEW(>;-wr!Uu)OIqVt%ri=
zT(!+<cRuMLX9I6&awyNzWw40^?{1l0pPOJ`i`8aCXkXH%o19G-KX<F<`U66j3%iT^
zD}U5~-FuySN1e?xd$UF=)W`Q>y`-I5-@NH0TA=qlTF%z)-W{ipM9kHT{{Z!w-~RxO
z7ls~y(7*ox+^PQn%~Un<6L8ypMo9kvEUhoMd>w`4Hchl@^K?;K%T(VrGV89o;2A#1
zb2IkFPZq)O`aC`Os&s&GEg7ZL>t5x?eCMLSfW%{KV;Mp932B)Y0>H+AExohpZ9ETy
zKzC(Hl>Y$27}4Rj;Mhk00PVBoKUG&BekX{J#_xrd#T(t|i;`yQnj$d{(J8@R(a(q3
zx&%RqSr)SY0E7v^UPg0Em+H<52}d=Z5ibh-Gl4h>I$7|idhgu_iWilf{26)^C{iU5
z0v3;>J>0W`d8V<Wl5X_!NF<=!RE;02(+Wi0I4fA=g^?QW1!|(wtgA;qJZc?HmWy)D
zo_-EDlYCLdnq9sfwPe9_z5#&F&|e_zRje~ySyZs>(eccF4M3ve>U_J^wktD;FQq~G
zd3T(#zm#<C8&ku7E#GgZ>my4SNv)~d7=qvphHS7ca5HIR1954I)Wg&{J0UUwF(xOT
zw$=^K-Z1#?a+@`SF{5iVFV$u#POxJpTEQU0p78L3dwfgq8+l^t(ti)c9O2G&GCQV2
zTJEC2UQ4Z*=m-EOhh;|MsX)pa;BR8$4=rTuHs5t6fd%ue?JHEX{{UJ_hH24vVs86=
znwNgjcf;YwGw~HQ%vq*~?!YnAFg-~m?*Zb(3QRj^@%6Xs5#eF7*e*1r(W<8;&)XU#
zlBt0SNXC1Ixp$&Kr5Av(O)7KmPH;+a6Lo2AACC1{QjMVvTHnK;rzM3Y2ujCtC&PRs
z&~(<^#TU_@P$mVX30p;5SEUznszD(>t1n8opwoipZW8ehT@e{GqILifQynx#>E@n@
zHBpPCC@QXbXa4{@;6M<tEvl|)o!s+*G&D)Z>-#zyb!Ev(lf{%)`)dwZZ09%`m%=gl
z<DE~065qzphGX+rZ~!|S2k@VY{ayJd;i1ExNZSvhja6o}UBudxWY`C2bgR?m;c)Pe
zoK=sX7iYsY!>_=R^j?q*n1?sPYi+Sf5lUh;Ma$2I{{Z(*#%s@f1>7&qpO;1-sIJ|r
zRar{<w<#+jh6-jxTpz{(W^GTRnW}ta)9`0k7#M4dK&Z*EhZ9nrvwFab6R{8m>pvRC
zm*;>^QZ+Imt}_}QK=`UO!zzwADx64B^dsvP<<xD7o_^igLPANhA{VK7-1>5Y^AMe~
z%%;TJp%zKsyEqZiHgE?m<CmdeNi;UsmyLO60%!)+$T@G%zbGU`q71?VjjAqLd^6t$
z)<hK0!4-0|pMN+ukccpV;O4peJK+N;HciY`H)Y??I3*BMQw;nZ`{7AWHwmf+MVcpy
z3(5|KR4FMm#G$_1ce3Qv)}vyaU|Q$Kp0?!W`Z1k`bxbt_QKK_l`h~!138>+ia~1%!
z+-loMR;y*JTO*7R)`2=6=&|K}4*C(gohBAxZA9QP;YHogoOU%R;vFHtfib+8Adm&%
zPEcc9+rs7<R+!6{35lCP)C=12C!__eBqy4w(y5f0XVn{eMY&@nPA%XXo4y&Yahp!9
zDwukBGVflP21nuz4gemYnT_GR)z{Og*S@@H@^$1)N$`tu3eSe!NsK#(-vw2=Nsj#a
z2l%9GVVr7UsJ+{@zbmR0L}Drr%zkP}!y(Jt-jUfuWD-l0HR{W`As$Hq3A>4s=8$R-
zo040^o}N&XO-%_4G+(|Bh)quHu0>c7n-UVPd;b9EoCvwmHFGpvm2&hXBte)>9L;wp
z)dYm(fFU;|5nGlO${?GR+0j)@(N~{1Nx|7QG+eoN_d&FwNi{V@m7Sr?)uf5{eoru$
zlUMYZbdqJ;oI7~jKt<^rGrO)ZYqu8eIOgGBl10woWfB@la0JGst|MShnv~4}!y4>B
z%?C)|G1lTl9VAM=%+@ED1Xpon?LpECwAbLXY?Ey<_=o__Og7pFn`AT6Lr;I{M@CVo
z+-*D^2C==6?MauuUQR}r6Ap9zatS2bbbzAjtacgIo?PtRz&ovgJH#{z0B8sUeboU@
zM-b`D_9gjzL|b5=n9t=g@U~04k1fEyE`s8*W4UfG1`Sr!8A+E0v(@J2g6CN3;kDe*
zE*Bu`7>Hw-WC#V$^`k)J0^>%V><Q6nTFpvuwwHrtg|S&@X|qzTu(JIm-%q`s=*@LJ
zJ9y1Qh^R;~^5RFd*_gXykb^T|u7^}SHNBf4wB?KJ;x~iw7gxIFK9=hZ%gj#eSAp$E
zk%hWanduc@%eY>sa`7Ll)6r>SD`lX^HrtXMM>hd#Y~7+eN9@t6<=eWAGuk*G00{6B
zW#?jjQOuht<$oaunVEQJn{LE;$FC=?>EHQ1-6%XM>cytuMjg3Y#?>61VmZOPP4x~E
zcXHx!F05|nGpK?*#f`dqm#2=nGQGNXVdAHf+uA!xIL+R!cUmj#D`TZG!^_1M(rX>~
zT*ilZi4ft{q^@uDyt&KV;mkOZ=X-b%5vjGQUd!?weUo;CMW`BI&MD59<(s?q$B}!d
zYQ*Z%e`)DBG8pk|`mY~_@wX=q^9{s;T4LIbEi|hP!27}G)B6X0X_h0ROj0Ik?v%hg
zBRMUrD>*B6gy^^C?oZWtmh@hz;e)CwR;@&W($`68HxoRpLAF;5I9x8(*aIolX$C|{
z-a2jJv0qgTEg1A<z%a1a3T+;F%5=w-o;q}qTxOT}Gbdquz_l#PTmJyw%qL8FYV}=L
zj^e82BeJJZYArEMq>y))N5&TI76HPj(+>Xe?DK+NcGv;#RDF%%$!cB*z(0gih7`%3
z+goM%iT?n^iI!M&2O-o=64oAk)inBsD_|(nZ8jbt58fBlU@kcbXvzoxfF27;y~rX4
zs?I8<M>vcjy8~M212D{ISSB)&rKA~OKxs|Xres*8;EI1>!tITLRi0pt>7RL~p6>Ui
zH+rocdWuphHir&wMvWXnkK|#zb#yj#X;^g&O+!~4?Cw6uBHzT}G1LinG8Nj3hN)B$
zn>iMVXQqFh&docmH9PdoxR2q<JR&h-ojP<$aFSiCiI*P}^bpao3ybPhtTMU)8UbdH
z0ne$DcO_4@&8RWkRmLE~IE#&&ZnSDH4;91VgJKTo(i9TV9H+SK&KV}zA}iddyg!Ic
zmy2+JHX`)2B#Yx{;=f~82s2;^jp>sO{zH77L7HcFw^iI!r(9*l-}8@He-QS7=Vx^`
zRa#{HrW17Ot;ssg6}I)BrM@>kL2vWnZ_ms7TDb$=4b9c1h^2vZnAsHD>oL}K4zYQn
zc5_J-=ESh~Y*eQFV?)W89UMJcy(em;k%ImE+Uvh73ctwv5wDmUETq)!K2hs<h}t*A
z_<TMT`l1|d&Tj{KXdbg>OmjAgoBR{ujRT1^Iz$J1^t`R*f*L=-{3EY!ugP+K9X~J;
zdjXxCFAT?lBAX^AJyE3J_EmglQ^YH|zm4BNJ=MEIcY4LED$o&7H^Pvd6r^V^ef;6d
za8@+4Uw1eXf>N}1f5sGz(}28<Z@cf^1lbT}VOrh)00>iSrjo5%Tes!uNOFh?8hIzx
z1cajL2G>t6UWii`Qb;DPP5ZaPKoYf$ew&BuIOL@96o5NSlf#8D@IKwm`T5Wv{{Yn#
zF$}|EalM`?p}~~NGpBl>k1`u==S!=Ujapb@`X&pX0yXZmljVE#Ropj&>GfJw(v;$s
zq&)ddHlRcV*B#z(-rSsgxO~;z%Z(g0Dzpfu(@Z#$AlylrY|nbY7kZ;c0wkH$tx}y?
zCpiEO1>>3JbKTVHVOMdR4q?n4iYc|CH>zO4%u}XZw%XU74j*;TSv&f}&>S&T{Tl;X
zontPEVIW0L@R%S1ANZO{<~y<nS}rxQ6yC@JKp;<c#v)IvxG`g6%-(Pv9okd)!kqip
z%*#|7mvH^yV1HCv;ndC@_^dq6N`bt9hKmk@ZQ{7%-`h5gP$M%*_{5&Jy1pArYltJw
zGRr8@kg4wpNZJwtigVwZ{-{z^plrRMiks8Bo=}PAni;aG(u=uZn;`>5P_j=?)dYZq
zBLqPqMDELcM43WvO{Y#-?}AE|(DYFk_oN6w(@vk?0uys@d1jv-&>;&+Q-o3y7qo{w
zl|7r56!Jn_48HCq*|Rdz`jZ8V=s?>`xgIs0Z4iF=#pB2S0Lj0dzb7mFJN|T^%I~86
zi0rYEEraZ!vR&a>EY)u3RG828$IqKShQNO}`ik_9k4b!V`326KU6VDE#H5MM@#eqL
zHyN)WRFJy>7v{Um)tPB+wc;Q}M0v_~6wGew%6E6_MUPj&>Noo#>?c~3<NpBMq_Z*2
zCm6y9vhVb{YdN+@Roh3#TfPzWBXU=pfD}=*qC}{l7)XdHWZEiPh3>9!rV1iT%*?7V
zT0;1q?x9GFT{j^}jHV`3Yl63B>4q-pNi%R2S8{Ota)%}*C=)0|7b#ktxhg`Enr>iA
zaw=p?$`tH`%)vR^-Ong@2u#c~OtR*^@M>iVyMQUX+uX2{H%(jx;Y`0=38q%&xuSTz
z5XY<%Iu#8f$+s*(B%(_(Gn!fHeE$H1WzqrIVfTl;>YI$M0;8qT7mY7H;sUzz;atlF
zgLX3Gm9RYeYzjltj-=mfY_ltd_KkCG(oaf4^ft4<hn6&vA{%kc_hq6YW4*hCSuuz&
zhE?wop_3EHUgzN!7rLg)Hr&%x*2Zbh#UkUfEh0Il<9QKWXrexmMEy9SOASV?3K)|2
zG_-*1q%^>pg9OA2T(rT3W%XX#wP|GrK_f7cc!`@$@3_-tF+!3V3S7NB93>fJc3c$K
zG9edFFV!0uGZ;oPXD(*XgPbqc-4=c#eId>wCJ(#<pBU!5hvKnj@VKDbikB~XdSF6s
zSt?|{o$8Px3`kAgRqs!_1SW)vtCFYHoCt_b3`AEXUye|<M}lm5qUV=1=7|WScqyx#
z%a-n2;Mx-hVw(~YrbwKZbR_SxYJQBkeh?;`9ocSP6;5i9Au}}g_i~8Mp$VOFrBA?_
zml|CV?ZTy`OK%qK{{UK~WWDs+_KlSD2_OiOGFnEHy}j!}q2Ji##)O}761Eo6oobvo
zLj_Ar+}#qE-)pz0<vE|mCjS6MR0ID2XO8$kzZJNIe|+g5z~19M;1_kxD~=<%+2Mcf
zv`_Inyy@_e4}#YXNOrL^alsiOPdM|v8~*^c@Q(E=wu^-}g6AC?H9b$^Z#Nrqofjrk
zH9>hHplVOZb3F%~1+8|iHglNAW|fI?c?o8YK&EZ{YM^R1k*|j9NfY#LxZ?(^9fRU*
zbw>dpJGu!B8HgdEf=Cx2z_}y<AT3ru(s*;j0Xbw|B*YmfN%!?7S=3db(Ezj>O9i1!
zwDRUZd~eefTTB`k4!q3F`%}F0y2*KYdF?j0gl!u_XsW20C)Tkuq?I2PiFw)S2m3^X
z*|#vGX%j_l!|nB>o7qmg4LZ#Z9Eb#bevnfRbvI{Wj;N(IN}D3>(wa{Gex3q9dwn0E
zO~(^Xy%GNa3i*#^^5Nv~S-yG4Mc7oQP{nrnwg<dueg#;hW@d2xA-Q^lw{F1IDZe=>
zE<`R9mRs5zLLEjNZ!E5`g?bNAV=}dBKiY>o`v-=L$Z!OU2S;_yFgv7&UPe+R+-)~{
zA9Qwrle8GziO`N>M>!T>QNRNBK({uh#qP(fn7LkhEg^i9+HTXZ{bOX<&O(PY>y>)7
zT2f|p=H;BxXyzy9XI=#&@32`PZiz>Y?LnuG#JXmqngy+tW+qOTBwKw~rEMcPN+9}y
z#EHD+4Hn)c_<63Wz&k&dcE5tK?RnQGTBhMDJfJq{=?sC+JP(+KgTlH<d{<qW5nMc@
zTMvX~T2!$3&iEAR*oX}-IgVJ-z)U;V2I4FgD%gjy^=i=WobI^XFb3Xj1cBWGInCEq
zN>wTtZf;VQi|5Jd35Nx$eR1E)x?L<i;RE##k^8Q7X|}a&HBm<zWH`fW9sd9;`9G7r
zSB_4ypz3t-EH!Y0DIT*1pTy2>+_zTkEz@I*9vgE0>|2sGRb!Gip8n{NC`3TJ(Y_p%
zx}zgQ_d}i0yr#4CnHr?K64UHl4Lh#MY8!4!yZO){<7B;{QOd0<c$R?i>dsOu=Fj0f
zbr%lMgDaydRisOpb<`V!;y3w?uDDam{pYZD&86^5lVbfUtv<}mh`24X{Xa#^jBO4R
z6VJdPncc!Xqs#GOjN#ZZ!`|k;o6c%yM>i-Ww=KbGnQU3N0t7T|;0_-hiS)L#wal}(
z4wC@S;(@)$xs!F=m}*zE&}eL1(NC2R`;d=OL%D^d-b`rsG%Q-@J|?S<?sxi%v7m8u
z$5{)MMu?=>i_rx0O)Xluw}-+BMYIVb&i&tVBooOfy7DvX?(}B_-8kmF4R7E2;F%Uq
z0?wXeym*tAP^Xe0A)6U4=yS;s5(%96SEsrZltdIIXMX<pV2#t>R!y%=DMqWGYElw&
zn@ZX)A+FQbV-3=&w7PqYrRCY&(0XX52tO<%sGknION!jel<3#u6|oqRoKtHAUIB+X
zNZu!ZR%R8}KBYV(DA74f+~eXqt$=ou{wRrCYY)mZDM6(*FPCLWU@;*(g3hb6^hNqW
zQ=-ak*bm;nUpQ527e1?;>)(Igt5|Jq{@b&<^;JhIQw~WzE~=bUCEFOa?k<tv!RhMh
z>d)dXHF#}WqekT;n}~CJ*U$~W7N@%a+QJN&+H5+R2ap6A9ZBDonyU^Fn)yc%Z9U`?
z;H|XKl7)i*0OBl`sv&ft=8Ked!Mipm8vb<~a3X%`mxHA1=A&Fi3^NYQcb9y~@;{-t
zSH8@uwx;~|Wt?3kMU`L#*$f%UuqE|jNlqfjg4RpTH*#1Jcta+~y%X<&IJBS}W1e1^
z5f@F|Z4pUR>w!H`Nv-#0o}O^0k}jH%ExVK9n&z-tLJ(9t$Lik!J0SsE%{Tl}d*G3}
z5R?%JH!m!=!k$Q8Mzc-h%{S(|;UP}pRzcmP>}VaDCfh7;Ln3vv6gIM(1-pxfpRE|5
zXe4lJem4A^v^a14=_h8NqWy<#u1&DmKFHCh6x-0N5SbO%iqz&~=glY3nLOR|E7BaN
z6!_`#3!XN=CcTFdWGVCE{{V#4X4L-huC3(1G~db&(@kxs5HMzlM@YhkfuB}W{{Zlh
zSHQbJ>}sUM_O#<k{_>ui={hXjc43TQ6Y4v|)pn~;O@gbh=AJ#-r>B&C3P?9zZ$JT2
zWQ&<9^&)yAB<_TR9ERleU?voR-2qcHEdKy}DdvC(OkHUZx#TI&x&UP&6Gd+H$veF%
z4@3bd1PX(ieyCE2Iwp=bg1M@v-2pt3(J^r{O~c>lf@TU6RUFqO^mABfO`Jr_D|lDA
zS2&J^2nGnM8Tw$7P?-uMqa>-#Is9RbKtTo|Dy<RlRvVUiRVJ=^z!LnZYaVC!{ok%K
z@Blr(dT%Z*5gpL?Lx<usGrLZsLat)7H2v6t4a|3v(PPL&0nAPH3RI0n*3eZo33q*5
z;``j@!zt?RH;!)d_?QZ_S`50TgI=?s9y|2vsnqlaB_C<+!p%kf6G6(rS%b_|>tEW8
z+HK2)YaHU<&Y{ZT6PL7eVA#wUkr$9RD)Wn-UC(Pnv$SRxjM-Y@xQ`v6rub3idQPdB
zk^oeuH1$_9(pyM>QpJuQ3oy^5eVUIK@saQ<zAg=Ryn9V)=hNn_W+<Kc<<BLH%Cc-~
zNW0U&dccI;%+U!ESD&U5cZAf_ep&K@LP@FTo<zJ|%LdY&hyW(effqmd^uh^mDWN9r
ze;lBbqH&t(vaU(>ZiI*t37Mt4*VDf!69Nzk5jcG-3T~QYe64_)MyZ*iqHkY<tR)~8
z!NKCx)VpChFr!YB3WY`GQp^My1|D7CNfY#=&v7j6;*?!Pcb)@5>f+|i^agyErVlB=
zygb>iB=`Uw9&o4=QWJ7)vpH8hr{DAAF5>z3GvZeXJP1qo{<zX*PT8Xm!+lnFdrxP(
z-HKyJkOkdl4sJG7+|a^m;_h91y-nrEVbfF4C^@H%<jup(8x7eGT%yG~RH%a5W7aZB
zxhI)Alkgh=xm2k7rd)Io2EcS5vUgSey;B^aJAx@jD6<qv`hRK>T<n;8&pbb6OE0nP
zk3Ci0t}d4opYKbu{{RCBYdzo>e=^plGT|y<<J1@aCjwpZf#j$4W#`kfR9;5rfox`U
zxANo;dp+)H?+;GXsno+g)o@}%i7uz4l6_`u9r=A%{#SRFV0&u;PPvG%chTC*M#on5
z3KD)&YMn6boa+h`A#Pw>`SLG$GVdO7)}2aNibAauAltz12-Fdw2H|saT}{~@BloX|
zRT_KLk$&vrYALzFq7d4<@uFxv;I*l6^$msp02rr<hSZ}~+ZZwihAr-&C1!zfr&g0r
zqDPnPFRL3(tqv90?mfhWTyhjy*^v5L2XF0V8n}v`mJY36O+NUt&!o=QyD9S^e1Apq
z8y8_pU*Ol1?i{95#Bm-YsLf4Hn|Jx<0mkH|_-?nWQ`^PW9!9GJu(aq?v8ZQ#;jOlt
zIL^0Tj)5MK>W<Nk9{afbBcY40q}URxEZ)`e%ea8sXDZLhY)rTO!22YjON(M>#M+TK
zoG$&6IP{ciH`5;4(9&E${3j^N16{eD0RYC!8ly8|TKZDy6C3xPGDj$xBFm|!sj}0u
z(^T4&khv`A6EhO?4O|ZosHqwbYaY`|ppa}q79)}8^XjPr<~g+(VlD`=9tBgu`HTo6
z%~D|9z^TI2^Ui8LMfV(E$~Euv9+x<HqM&_F8w{-CY3LqxXCU*tc5Op(VEbKIEM_=n
zqyT>H&_AnrJnr2c?7_0w&D)<(3dQLPq)CM60AbH0qUyPdF^(6(a>X-<>Y&k}YUp*$
zI3BgVS$lbV(aN?b4}xZVQ^R^{PlH9zx;n0usA6$X$%kQg$as@rsIpZ}r0oGUDAEf=
z!UfY%o@8CB&WEhqWx&FFT>CLewfLVBk3?^yu={TS)IY+>%HEEA_!Vf<z+k`TUE4eG
zw#*;I?M~Kgd1|@u%KHmYu(rn~LYXG3Qk}$70X07~=V+Xq;C#4@_{*0rQJ2Q>_-r2y
zYuF2%=UwhOp6Fw>aoh5UXxIgJhE}hM$HU+Ti#4MwM7U@Nx=fy{xuhk9mSvi^p|->J
zBhxk;O|9eyzy*x{^;aHb{K}FB;mU{*jty?_f_WygEnK_$yOIe)R*qa=-sOTySJfd}
z$#{2V$^dz$l#N`UToMtquOmdLLuWpD!93G~R#l>2E_V80ltC6iR(Nee&)WbgL6}KK
zb6%eANe)6_N<)(QU1;$lZptE0GMojZsl`_j+a*SyB*`^1i6<q2*Cxrb^Aggse606Z
z#S}xU&MEYo47`ahjT+YSI$rbQEUl4D3JrUR?#`2|obG-bD2e5-QyL<o(#gy$LE6rb
zDI_xv81WH41FoKEi@;W)b%6L><5H$3n?CKO%r`6z%`ham4CC&})EEa)2=1A+v|MHG
zf-Raz1ED!jd~+IT<Dn2cLK@v`1~FHtOzTQNKTDWl#>Uw|Z?`a7nKyQ~3NH~CE4W5D
z;@^m?gB39*&NSj|%e}X-c5f17+yN?d8%da+&X~|TZ!YysuMy57H!9n*L3z^0^HWWb
zP0UtUn0Q=HEmNmTk^a)hei-vsop--d=FHmSn*s9W@w|l$YeOGRt>ozN&STG?S&J>W
zjG=KNeQ~5rDnnE-71I#(E?K9rppc}HP4J2${{RS5gf&aIDs$ZdH?mMrND*%5Z>AWC
zO*bG^-ES`L=u%|~v4ZEleUP<688>$2UQ50NCdwJEYv1XER3_#-^2;Jb%bW;7QC{@!
zfeQec7De$$5Um(4oi9?_X2$8WRKS>)n8s9TcO{<O6998U1R3~lMduhFXk_q^MEc`T
z2XC^|;g-gj{3qrY?6(3^r761{*<_TIpfxTswJTX^u>l<C-zD5UJk~yaM7h`sQ`cYO
zE786NE~l5tb4J$X2e4u^MCS1GPw1P96+LpgH(6hrALS9JG}hkg?t{K?u;v%Nt8Ub9
zw>&>*i`9=!z?)IrCapl=BCR<YL7%*4j?{Py9!onM=)}`xYRB$>3EM;TS{i6kUV7ZO
zkz|qdb7vV|YmiaODus}bB;621b4}dIH=@9tMu=iL^YlWL(`Whb^zMRXE#)ylRUFlR
zwcQM!Xow{s!7|=G@T4Y~Jd4ZG?amEUNZcj{i=KIMhtVrd(^oCeHD8Z>2=Gmhi+)|(
zvq%vF6Gt;$;=Gf536KTTKundP6#AeZu<=a95*2PqiF2ef+|VkvChMrV^m)_Bqb`xA
z?j9LZneLYy(EuFo?*nsfEIKQd8b&OZxL+uZN$!hAk>-T#%F0X-I>Uvi7dGjqO}-NK
zcjvw#=@{w+zY^QHS{mZ$8Vw*s_nbrnbD%0+mJ+BPi?lbm0#D9B?unT3Dy>_QeTHYz
zhu5_~Hf1gd1R%>~0A}hJ(eL8$^u<!_M`cp1PD+nE$dVlRhJi5PWRP?tuY9z{R<c0M
z#_PVJvnQUwMa<5r{V-4}HTFVOnr%v1z-WgnM@vupk*$YtHSrB=NV^OXpezI3-Pd(O
zvq+~%k`0_E!NmLSrHwrym(>=fSGs8GDs^}I76}NvChNa>3h(dFD2jK2n;ob-*NIjH
z0Xn9nDa$XdRs`TC=8Kp3z=#Q<DhhkPo^Tk1Xqpy6nf|DZ!fC0lW0$v*z=*V_X1n(%
z4qf}9Ckd;on%s-sxpILxgv_@jE14-?V5>GB!?JjyO1DRvsptvQ>Q*|fY5*EaiSJJr
zPlxuoMh6#Zb|(zh37dbK2Tf0&i&FqvC{+$S{Xe3s%~R)2dSEp#o+G^4BgyATxVLy+
zxi{tozBH-g3u%K{kOi|e644eR&5X>5h=S416-UvRkxkNh5+u$0uF=9S41>~Dk7m@F
zIqY2A&3T1d4H1W{wRXsIF3&7FnrffYIbi<)yoh5f7hp2s90{&PHV{F(>`afs+|Ke&
zg)-buYV9MlmpBmy8*#5Q1I=iitxeC@aLy{MRZU8)?9kMv51*bVNYD0`D)B08rAH1a
z2n;2opvz(IYO#ZJpH{JXlHdvV{{U5CeaoC870WW+oNBb@V_r7{;wS2}$SrXSLQbzJ
z<g0Xg>ts}cU&Y^PE|{lh&fi>f>JDD5g%#>jcWdc5QUi{&>bwmDx{VtBM)`%8RyaAG
zm3xi3x2XRB<DuwoqN?<%nnRN0`2oqkI<m8D73@v|b0J@kQWo^_M5$J%UO*X0^_q2Q
z>(J~)h`34g-AswNIvtOPQMfk>46Z9jddE^g1;){@PaaAErNjYwSKK#+sn^X-VmgM2
zgu}G6lqO7Ve#>;peLP()+`K$`A-*zkn1|wNniwYLh;eqna|4m)U>{^TUBl88Y6sUk
z*f+<SA0Z`n7nT)c<V}F%CP0-PpL$A|KkpwBFW3c*J}FSr;gveT_|MT&!8blRgOs0e
z{;P!;l%vS`LhOSL;>yyuOT+Z35UIG^P7zaH)Ou;wQ3HzLGDP_NBh1i!D!8mW@o&i7
z%`<xjkAAZe&2KzCn0noVa}WpvVn~70eb-=YJ@zN56IEcv_0n#8wm&e$h8cRLEiu|A
zj_a?HF*}8^yNiF-8`xGhtsebqgg6pLQudtAI_6`!bJum1T)B>p$(AAv<u)9{+t2zf
zw9Q)*;OQCg!gKX0<}MSKqt-AjN?oaD{{S@GjI^3*>LHE4)r?usjlxqS4N%np`md(o
z;nA)DOmcU<am$9px@)jA^ERVTy{3KO`_+O&Yt^^rCS<?Sok8lXyDhypDJM>xW@c5l
z2gDOBApLrTp7cm#YPqfgfqq*!H!vbb?Ry5JTltWpaI}Oww#zybl#e3UN&f%{Jroe_
zPFrWwl2p&YO4BrHhRyGgW~pJ?=;76>+6$}Zz0<cZ$(6h#RPhRYB(b$<)N3AW(%$n7
zdmZDNnfyAm`lO2s0}6P99QM^?oZVarhi%`GA^XVG4)FGqz}npQHwu5?)m4cqeB*k6
z`_p}&Onvv<4Ip1~P#!PoS?3);w0$fzVX8KmaMYN0iF1ZV-)4%A{uKj(^m2P<nwfSN
z9PH8FcZ-V~&o=v01Z>X>NVpZI=9;Njl=Zt{ki64Q`rtQ~{{VT@bpHUPi<o>i7#~QH
zRmS<fZ*|;B>f%o*RqNs_$6{J(4I|aJEd+Qp0phl)uwuS<G%tQGoWb4YQ&DBa9$apJ
z(2bi?^*sFED`vp;#29z{Qat|v+8oZoQE%rz7RS|U@PGy}<Z>T1+<HO|z8tMGk3vWU
zj!PthO!H2~o0sDRq7sxif)teCA^;7y7kBMLl%nzy^t~Te_)=5a5;U?~Q_<eA<tQuK
zd$(@Sp6Di1gj8$O+?=5kdnS~O{`e%K3IVl}TAqj{2a*$Q58d3U1fUI)!d~~-Z`W`~
zazF&Q1mXcJmvF@0XBc3~a;8(nFx2gKeaV@ZCYvWF%&JbAWF#xNa{Q-1RE9DryAzAT
z#Wwd}QN*~G%d=D@!2B%$12%D+M!FAeDmAk?(ptvY1888>{vu4<LDh5|wLOU-a3w)f
z<J54>h>h1ha#G?o_Qt4fHi8&6GdrH7?s<G6XOiccKhO?Q%XOu;jZ2r1DM>nWh4MvC
zwN4tvIi|P+R%AaI8-8e=<VoY|w&cvSF`<-$Y17g5DVMYYVrl{$*!6P}1-H#5zyl=R
z!xvO}H8pErl#_o@*v_C04%!nI7OWR8&{Z~%ILpXw%~#4{vu~o_$t}t|V4&5;5J4TW
zA~_Z?kAYnA1MaZk{{Y*+h;xFySeb5BPew_*bY4~TLk99fRPHk00uUSoBHW9)GEMuS
zLK!!^JpHMh2tg*rE#2@~5P)Ht+=<COFd;TJ*F2Fsy$~TcB+HUz>D|5rCibB(3NN+<
zz?20PRo<!S=L$y1h(T~yE$QEvEYb-Gg@i>_<@I;Mo2L-2+crmA+P=m+Jfb-2<4P!_
zMplX4Gy=Q)JC-nh(HP+a_}lVw(&2!5Vg3{I3-)J#7#DUwvSrPG>iOK1w|}<3A3vod
zhf@cWztmTxsRtQQj-Md8$7`~G>fA)5pU)2e0QYUiby5S|RNK*hX#W6|W=)xGw^p4I
zUeM+p!vtycXCKNvF9E2*!SNK_b*V}}{l*frvFvfiQX}xbf5KNV%?cCF5+-7c-5*GI
ztb9++d8ihY#M5^V-<EM3CY!hjq**6?DJViAOp7#4+?O~3Ap+M;TuvRspSLI`_CbkA
zvmv>9;avbus5a%l+07tMFLfI-B&1GzdBX-0QEoO!L@253R9JEpV9chb3u)v$ppcwE
zFq=3Co40*%r3ss(n4W3f=>V}nPQQX|gd_~y1V!e%k^$=h)d$rot19BZDJ;yRDu;81
zs_NzpdBe9qU2@(j8F)87z8aeu2O)HrwruVob;}^c-O(fltChf|Rqu57qB+J;Jf`>j
zI$WnkB+a@+O0JUJuPJ2Z$;h+t#)oFW-t?K@Lpu;e0UH5fc?R=Xy~K#yLAUjt`o~36
zmXkG|#NZE!5TLe=D`MCP8vsj}tH1j(nMy##_C{cnmLs}a930X4u3{jKg_YWqnucyd
z^SE&Y9}jrHP@PK6GVM*-B@#!1R2;a+E8D+^_o_#1Biig`Ml))3^#mIiMxyW?m&#<W
zc!oKYSWn_`)JB4I@RcB<nOceQ-5Ls<CZwJfNF)msCd8TAxqZnBc_B1BkuzVvI3`4<
zn;Hr)3MKyl7y-S}5Sg0p%~zi!M9vh!kc8CKc_MRONU#$qh)mr}M<QPR$&@7FG<6Yp
zNO)B@rcg*u28o#<j^ys;x)j|Bo1Dpij1qQ2Xy#swy%3PLKFeuR_3HT1-*HD(iS07a
z#+3PjAJ&nH#mlKw!&Pq1{{YQ}XyAL)t5fItIcrH{W~V1BPMc9lgm3XjTwI5X@&)eZ
zyer{fbxJg<3Y2P`O&k6ujYt!5`N#Q(cTftb!rH79DwzBy@cvWzN&f%|zLK{!HiL-S
z*(-Bu{l!|UJtaETLGYKCr&8qfi3!I{kfraNUp6;y?8m@yl~%V8e-+NhWi;9gdoeG4
zN6hzNhh=4*m;f`woDP@i(#5w2khW#UbC?Zldd;QR8Fhd_D@JVshlPIu@r0!!dTM5%
zvfn(i#?+WLmh>q%d)xcm7Cq27<BF}0uTGf7c5MN|!Ob-$;1>Ym*4WNgXV&HyN)+(4
zp6<grBt(!yBUovGNw%(wX(~FACplE4(x;ii?>xh@j2ODoIv3{9XTPVMT&Ibq$}OB@
zK>$l!8c6qLcf8Mvs_GAQLQWi>L=pk{gzab&>N(BQ=NZqamFZLd*$h8Sa4^@hr-LgG
zlU@?+2ej8W#IXHBwPE}|%F0a0jw4~rH!|?{i8_S2Tdi*iipv?3Zd=|R+$V5uiq-L%
z>JPYkUeIC9<+yNrSu8t;e=XN)bPiJwYk|}src!&urklt|mBTE{NKVk@Db&@Jn7qDI
zo2{_K?NPmNG{P*^p-)2GrKiS8(026lZZX*UjWx5jF9C(jY-&JmV^Hmx{l@+ox*bWl
za77>v(nyogn+E>?F1*xK=E}CsQg&W?OJ~z_3`>vV!$V*vrR4(atpt-YwJ!VmisIiF
zy(fb*gRuRxTbLQnOI%4BxV!>SkCm*ec6h$jR4gVUA`iPPuBGyk#gxoHh&CBcsL%#=
z=WEjy8Cf?SwV|7rm;7DxP5PkAu3vA9NJrL($DU3Sod%eaPgsB;bu$yRqlK_L6zYdy
z2fY6PMaFv{rPb)0Tiez%amMCqbb57kXFF4*-h-~0F@AO~qUsu_hkq1M!c(t99QKw2
z*;9PUurZUVHzWb)V6N7EDx5qJ=+5TEYHypvL3uh2F1y!d`#Egzl#tHgnj4R@)U2^h
z8NAcCgsVn-<=FMjd|$dZDPi-oF{4bx3GV}-+zA@q{{SVMR<=iw+K;@Qh}40pGwSkI
zI^ey9C)%OO){VSM<9D$`qStEk(~b|Vr09?|KJ(gR??ZcELBXNrr0CH4jwivc4w2Nr
zQEP>9oy-mcQs&cW9xf4nf=@-Radj_h^|FwoRB@0sj~M)Rn?$5awLM#>kLM0yp>4S!
zK1X?4zPqtrtv)4GfNKW0XaEkV2k|sDtp|MBfY{3GP52sfl9?lOCA9s-lJ1ZPgd<RP
zd9@rdD9pamf|lkE0M?swi6=Vdv)ng5Hj?fP#Nh7g^P`>Dg^-(v#L{zLjtqeWk(`3)
ziKs!$ZD2_(6SS#25kjCX!)6570$Lzz44G?ZQ=u`c-HcN)?i`YP6W|)9HggV7**4`C
znI!Sb%n#)zrT+jLvhQ#V1pU`~u7=}qjO{R`u4%qTwa_`agEU`uf*S2b^{icmu0=dh
za3aCvi5!Ewx=xOQ;nZn6OQ1_*T5OIWr_Io2Df2^fGi_m40E@y|uWLsDgtrnF^j;E!
zu&0V>uF*-;p%lnI{J?|EY7L4zO1M{=MAU3;+F0^FsRSKCxFJb@qwGs8Hq$ZLO8)@)
zDS693Md|Vn*@+A_R<H_H%|GGK9_LhW_#?Nb@9!xGc;)$q1Mysk<x@b!GB0c<PQQ%G
zpySxLNsUTOys0oYH+J^A`=c9+Hx*3bsbVX3vg=D4-PkT$uDg#lp~Ia;^=dwqAjlgz
z#&y}o%jlW6bi~IFUUdHewH~0xV;suE69d6HX_!(Jp(KXm^uv&pBLI7_NJJ+VZWt^R
zST<m>(H1Kp-xaV~5D^KTGwaS2$~0cn>L_+X0S8nFNEYC-6kS?yXtcVeVa&AUB_EqF
z_iYoO!XoFkn;lc|f_=hM4KE>sdwcA(h8V+d(6BzrI|Hq?S_)M=wl^^5ixm<HF)lw;
zxVTT6d*Qk%(qW~8!xq3fT1Yko<(u+1vG>ZeQNj3{(j-ZXTz8`0>#I|^_Ychx)8#8H
zF;)_?Y$u&=`dNF)cbqR#2Xgh@XIDpKFu?Y0sMekH5#8|{o<l+|NE*SkD-oL^BWMQx
z9J-ScrmJ7$x%zEe4^X5q+v-&3F$Fo|tUY#<uWH|^pXIoPE?!(V`v#6Sr5czqPYmz+
zj3xSs24+p1WDY=B4nocvokn5{>kS(2>|P=^(_0Np&g#Xo^6d?60(5Rx<}ozXz-u3z
zAUSzs5?^TwqVVqHi&zpIaWx0ERJ!}i7-!C9am0erMUL5#?sJFqza_Oa?LmvO6;#nc
zu7955f_bL`B{xvbM~8+U(vm`kNJ9!j5J-1HAvh&BRQ+&?oe>2#G0*CPN^qMN^tK5^
zAUFY+!VPYS39&9|-3vs}4ZYk@NGMoA_hCvF)7VZX!<b*n8_Djqi0TbDMW^GMB~GGm
z@ixSxQ(#UwN$8I~_L0s0%w`?hxxmD1BpKy$0OzvP;Z1SGx}!(3N15CA`Y+jr2&z$W
zF3mQ5U&FNqU&GUl)+S`8q|97mVV{*|?J-g6Nbrmqa7=G0{{T^5fy4eG?j1fsb7t5W
z{6iD*bi6<KEyiVUSb)2>;4jTD@{?&w?RC>5Byr3)L$rTCr#ks=^`p@6D%g5lBNI-e
zgu10XqgNPvNr4=Jy}2qF>UCX;rBDL_u9L8kK={JsZCPp7Ms`KSIX8UkPB+qFA4|+C
zH??KKxGT%SX~^xhOvpD7+Y7}AC1=@!Ku#nkrON11h?^i6KXx6EMeLJq@feKqObyW7
z^{s+-Oo+OSa@?WHi55|(&K!iGrlv3<G(J!vGJA*|g(GAj_{54NBQWtu=v78QCYYzo
zJiy)anyeR#jJbPbGYBhDz|*UT?*o{Y__&f<c0Y-x#N0twaT#i2Udpy20C#{EJ{ui&
zPW6I+fxOYNw^vYY`w`vM=vMo(@CUUp5Aepf-}2kqw=qhsF8)$U(mp1B1FX%PdX&`g
z$yaMZsGXquGoLwYc4*pPxfd>RTeV|xb==FrVKZTrum?#op#j#piwnx_*|_29HN#`n
z$aRLh+!Az~ru4JapQqH^C0AypoMoSRrtSXL_4r#7!-6q1s*^eCxxP&;jW*Ugd(>*f
zsJ)F5mEivXh#C4%Rb(z@8WK*bq$Y>c2Db^>G;^FJSY0+gcR?^=5DI4ZC?p{@a@Zzi
z390h<!jm#leN!XlbSWmtO<XPTAu)5G90*Mu5zD`H2u#itge|49X?ePa(J;*>n<#G7
zCnrKlhFj`0b%}<*y%>CP!I(;QUF?#7(tq(QM~6e7QkXFyYvO-Y#)pNda_>^8ly;?S
zIOb`_B&6AxpX~J4CY-gV+D6CmnWZNc6OTbX!;ZBvc&f~PCa``Pjh7s(CZ<s&lJYWB
z5IvCha&SqsaP(Xl%(Yx>CBQOH*+I<XHiB9YB>w<yvQm`Ip;fT{8OE$lQmC_KN0KUg
z=H96@YcIMW%V^VgJdRnX%xn%W7^8*5QDz;uoL&*fv<1)10U}9ohZp8_0t2&5k(fc1
zX0>XyYW7s?ft15w>lPq@14N%xi2N&81f;z5r0Z{Ety4KUi)EbcWDfbF?|aS1d|Mks
z0_wD=IitE~H8MS65*wM()EL~Zcy~G2pxTu|oYsefq1?cpkt<fyEOQIt$I7OfDJTu;
zIWF7HjfJOjE~mThtWJ+_TMODU8W|us%Lu<P1ETYCM%Ph)XHO7mx=w9?4hPC&d;*nf
zMpdJdH%wja98Jl~H|K8fw!H09=a&+~HTc(J=l4Kwwqz1Hv=BUr0(RC@6!>>&0iWRl
z-f=$yRmz)*=+p^^Bq~(GB3A7w3S{6g44=H?t)FTyXAmd9;N?#WaZZBp4x~Ie&UYh}
zxv<_*bGZRY^}8V<ravq|=`e5OrzI<ec8`ZJq}5V>UjjQa{OCBu^EF&Z^U`%Wa{P;}
zg9)yk{X)6%1%<vYR>9(NY;7mk4ajAsl)!Tc-HtCgPS&(of-I{xwHidZw9BT(T_BD`
zWfSjK8ipUibxa*m$6F`BGR_4#p3{gWrzfW7D)?M({&1<PjZHCf<Sz?+JR<nEc9Wd5
zV_RHcxTYn|&S8#gSQ%6Dla!X!i<U?axu#9p80<Sf-wj9@)Dq%HPyvG8M0&a&YiaF2
zvSN=by_Qj@5B*bz=)JaY)YNs;e23%N+lk|iJXaaR;haH&2_M5RBj7EnSWMJ7ZX*!c
z(8szb<9?T1cd_hUKT*P)St)8d&SDH{I?TUjWE?l$)ST@FynNe8M5vj$?%U#un0n52
zd`WY27nX?M)-Ey7k-T&^D&1Nkv@q-@10>C&%?%en4qTVgYAt6IP2r%cZ4#*V9K%zS
z&-OY^HO+d(Y>ddX#n8>tUP)}DH*kFAvA0g8XVNdkO#vnavpd1Uea#uUoz`7OQEg1L
zVyTPT8aLA0l4O|ZulRQe*IkiznL(s}4d9m?)OKD{!>?6K^Nu;hGOO*m_dld&I<bz3
zq~6FIn1c6pCh_K6d>1FY5)D^&k)ZYmx+ls9irSca3{Y^bKjfS_O9muK!hOMOsrP?8
zWL$3-yx|hZb~c?LTTQX`JjpjFsg=pjeMe?_pWee^(OFyI<7~2XsD0R+#)>MBIf9*%
zo1Cm*I)F&lUvpY&RN2+0RZrG&vFI=97_hx<UQxi)G|7Gt<(rnDQ27s3^=xw>uA-)3
z8C+(K541cE<Dpbe-%zJw$0w-Np24)4^VFZiQ?p_;{ApnP>koX<JR{<?3&8tbBpmZO
zwTtLCBb|3sCeP~EXkGb+x;zb6P#S55Rd!~L908hr3ubcj7ZPQrY<G?u#1x!eOZifg
zN5J7rOv}u*qcb5g9dhmgeA$Pi`$D<Jigjt%r`l0JLJr`KK^%(U`gJI~Bm6-Z1ak**
z{@SiF^05<{+8>iWC>kytM03aQhNQ4=^~=~h;|q+x{xb&CCL9fcI=#9;=(RXlIWZ2A
zq+Ikrs_u<GH+e4^N24T3m5?q1ls2^5;fR9bWjzx+dVAth5=64p?^rDogzd1oVG>cQ
zai#iDvlS>o5VOM$dMK>hx*Y6;n|F6x3Q&b+;^z-G>A_3U;kww2nYxewi>+I>$(LmI
zO|)Lj3)+s2H7eD=oOpCc`*B|u{G2|suj-L)Z*QvYL817%Z2ZIEWk6SCC=xiPn}#S9
z7KW!P)4HWCC<1bQ0qh(1c=sn!_)z)OZXTUxP@zmn8i8@AfSBe$9@ir_x~u8BrB~RU
z&5oz7yaAhm%t_B+8ETxU48-fK%mDe8<)>yedhfN)G&!q_d0lV2Y%_&rMnQ2M1Dr;)
z-Zn3t&4APe4&Bt>Eda+bCrGxTusqJ?QL%*%nKn3jp4~YE61!KIWya=DwrT+bzI6%T
zoF8wy#5YAORSYwY76t(4RA4!w-P1P&`ohzx7J>$f%?g!Uo!P7y@*-Pt1_+xTo?ybY
zZ2nc5K*sXYw`etAeXtTctg>2EZT(SWhQlaUWX@WJy?d-<VE9Mw7c6YLr8h>cn{{)K
z*XDOy1cQ8NNmY7;(ZR^fVS~*h)d?HTV33FiWEmtWD1d-tn3pIdrhs8`!6gZ+sh`yV
zF{%?&9FSb}L3Ad>8F0Zo(Q&G7TbHRf!jOa_T#_0Si3yUDo|K%>X~Sv3b4q1_eV0Ae
z66;T7bEwAFr0Py2W)UEee29V%QE;Zi+0_*r1`OOtJ|l3tOJo?zt&Dcfu-ZjrWP1u@
zic7^$96cm5F5K;x6JXjl<+m>5+_%S#<D4@MfW!Ju7&zyKlIo9n2Uzbo$Z!*^w6yHy
zy~FW!aW6Eh5Hf;n(9wHsp&uCp+%MOr0bqa8Ud#Ug7Hzxfb8l%&OM?wGM>zSGLL4BQ
z`3!k_1iOdQAnG0dr!88dF;5;JAi2YAfu~Ny=lv-3{{U3nXIA!t=+9+-ZNHR;O_`0R
z(!y(ah99OIS2#nU;Z4jAr|5kz)f+q;4NATT6N`NP+nHc);O8~DdK;J`^SBDGI;-i|
zr9r;XK2ey@!`!&8WF=X7_84Wd0eRAG{?09*U%USR%^v>%1;cQhI5F(Z2$Dx=^4ra_
zmt`PKNfKi=tIt=(R>XqC1Re6WZM?+X-&Gi6zq&N1WmMfYA@{R^I3Y7Web+)l71tCw
zi=YWB7DP*=!k8S8pB@kagbwH+DB~Ud(7_ROpe8O3_;M6rfhmK7y4#@eN;gdmcSDkh
zOq}KE;)IBV+`0SE$E*rO2u&&SfuQEO5X!TDrMx~e;rNthPz^jjkz>g&1|-WwILRYG
z*6MtQ>g`hI;_6F44I*t0HV`M7+oj?ZHBz24^NrMpF`Ey~Ij4AjZ~p+Q5y4S$6)@no
zJ+AI&b0S+4yto-Lqz5K36&p0`-=*Qb3>oPF{X&X#Tb2fNW~}YZ_D5T7p=<%RPxe=j
zFN{{Dhr$eN=+ywrSp<zBxH@D*Tw)_xh;g#(Ql)Q3ZARJ)fga*CHyw11LWgEfURq{A
z8O#NVml)BzfJB7Ha(H6J5ZV=})69F1dtjSEYl}$h-GkIFbqT9ZqvO_V9p(rJ=RRt~
zWy|iyoZ1yAFq*q<+&jh;z=Rtp^9}q^$>f!^O<WD$^eH7h6K6i?5vpeI7lj~3yCzO@
z;>&ahf|!@=z>;cZ^}>-7i%gW!&QMu6DT|)$B6%VS(KI52h$lt8yqG4OmfQZv-)+3i
zF;y4;0Nwz9dNFwL&BBT0Kj?D5!~3<|4=?<JlG$SpU<_ro?X9_xyMP;?dX;~R5X?<E
zAxjVUZz;X-kSOblBNs;v!@vIkAhf&x0MlR(spRhKKba~+va44T@A`8e#7P6{e3vSi
zgDtkhWV%8Y1?B3-oa%H`^$?J<GxKRXQfZZ09Zfl>OQp9<+u~nbM}LpjjI3o^SUMGO
zjqeAw-O+PScRz-iI%bGA5)H56+U!kQc!yYr3;sQ=&A@9Ki<|4Oc5)rN53wY~vJ~X3
z?C|P1@AGii>s9G#IJy4-YlS184AJy(`oz5E?z3CKG8?(nlIu5HT&IW+6^k1Uk#}z#
zr<Xyhyt_2SlWxn>jaAnGn|^eL!*qW(DfuD&W#i%D9R|{=QlPJkGDX8h+TOpJ-<av_
z^HohowY{A&9M0$T9X;v<*=)do+eL5Hj<2T5&C6Ax-aB|5N0ABv@1|y%gyx+sOm*^R
zqfvF4V(D}x`$)I(8<6t)CH7;HW-5*F?hpL0H~v?FBl&~8LY`_>f4pv9t@^i5aUKfA
zMV!UB0>3Ma@Ktcer%683EnK3wW)hq>?^`Zu2vtzU@Dx<I3`k-Qse{E+I6QMQ=doo-
z8M+5fR}gLsPHn&iFxs?_YeCtv?%lk#9Ec(hg5R4$uRSlc4=WsgMx@;X3r^Ez55mEA
zW-0A`8=rCS(TXzKoNE(^fO{r74XjKwd^BB}*A5v*kNwxQc})Elnb?gJ6=B_-?CN6=
z%D910sy|dQn{G_lnksh_ZZ;-pZYX)nk=q4U%Wf;9jt#hqbjB)!XFG{A5JLsB+!92(
z#u{{)vbBnJ7u7VYIx@6D6Cl3FH%m4nyfuhlP<B^C!E91w@g^IZIl9kh8f@KOphR^L
zHpE2b3h_dfv-3g?qI#Y2zlRiD!CH~cw4FiSv23xpf#<5m#CFW7*t})j0ig#+P4iDO
zPm<e$HjfTbVtFdy4beg#WI`yH=|hl+PDF>}1cX9iSff#~9WHK%U!}D(M}Oy~W?Iy3
z&wJzT=AIaf+e)(yI=~;^qmdg$#>!1KDL6ivh~@kEXtd>P2W<-n)mz2eJf1p?+boL*
zCs2Fw`MF@}l{5bUZxQ^GyKAA~b>X-QomzOR6+I&ibM!v31806PJ(K<8e>k$YO0FKD
z>@g!O?lcFO{3P!k*KQ{3n)bDdX<JC!{{WJF{Vlgzd8&(y&(dd^(wwDC_QnXCG5Wdp
z#!Y->41W{Awv5YM1DbVd%q-#0exQY+Lk(94hQ${_1+Fb{(ZeYcJ^n>~C8jcS54HkJ
z9+A$<(pDoHT=3NXsc&l1<O2f?2~MDgz}YYxx_nV7P`Y$1?z#!QA_+_3Y`UP6pi+oV
zTjE=}LeVfJq`cOj7lr`gqKleAK<~GdIZ7ob?7=9|*<#Dq*i#o$0(4S(Hh^+ShG0tV
zmeN**+iaZ0Zw3AiO3c|eSFF<MZM$XqdXr#RW$xOs!x|k^3D~g4cQLcdTrcGcSZHn}
zJ;HK_A9G4p5u@SCO@5<KreVaAtFa*8X_q&R*IyEJ%U<&SaD-kydG**^qEu<orCGw0
zGMK%|CP^|$Cp7*qmWZh8)hL4ZW|CsS0RRYrr-1mY?Ay%AGp#aejr8!1#6{|&4^sOE
zEII9CbJQ{Vod;5O8kNDVbKRA{zu!Fq);Ql$nVbz*t6bQ(T(yvE&J>`N4vpMDI9#_5
z?++I^&c->x_-fTjWL0cLOF;(a#$BX=APE;TRm?M5dumk1X<?br!4NjK)2{XwQn&(Y
zVxZP9nVFPUpZp=7V4ME0Oh1P%?kJuBn8i8GiLCkB@ZWca+H9kW4LyKqpSy3(-v0m{
z^^2{5fs%LHv3D^lxNVestr1Ke-mEY}i4sthcwx#Q!fs=`i=|+c0nGsVh5&h{M1<AV
z5Kl}JQXw=n%}7QGDIi&x-Qs|h;^;~lr@D}nf^|(yV33Fjxd5SE1eE$L3UBI#Aq%3s
zm|3MwZDNB?oTDqX_uiBm1ov(^cIZ_QoMF+NAs^=nV367UE#1-PzSO#_I7KRy&XonD
zUhAzpUh7+fwXVj!uHA>`K8yB2g&5P>=FW#k-}PhsH(h7rPe+KtfH-9V`A^KRL5M%`
zwMggm3z&A!(G4dQkab=^_)SF$+L`92!C(UX<av1=9%F5VFI198T4M;vYBSnNpE%)8
z=(v{ga2g`z$!xYa_$hmMHY_&M`g=~5DeGv{+^OTh^vWlbUH+FY1)1uPr4!zFiC$Vc
zrW`}zSVX`?36uvd@WMc{X6F%c7H~-d6E{Es;`Bl#Hi0q`^zI0ODquy^2a9{94tXe>
z-7qh@9P&yEZtcBXt%6YmDA##=u;fjWh!C0(Z|;XBD817MzW5OqP2C;NKo58%L@=aH
z%<D`|%kpWboj;{<b^iblXZEtVepg}GE1Sf%q(I6E%NskH<W81X@ma-JnnDEJjbwo{
z%cnkRi@YtL7IC8IhBVZ1{qh}7!*x(K&LN}s`{tQWeB6)b{QGWu{{S--L!VEHnw0?i
z%40CG00+Pao>>9J>n1D$vkiU_GM}>J9ecO%m?U^}RLRphdD&LvjJPL+UTiuAo)RzK
zk;bD<ELAcdNNrIdfdD(G19PD!Jnzw`w8OX%25NIVn4P@5;ZK#>My`QS>X{!d_$3L8
zh3fRdB^stq7wE&1gvrCZ#EArwqAXx0&L0m_!8T4H*ro<oFMJYEW-OXITi~178kkI6
zFOATUL<0$%h5Fz@FcWg&NK<5>YMD4&tsazuRF#VaVo97aityD^R#l^Ux~(|>02u?c
zKOAR?EHp3LjBYv{?hf|{5tQs>R=8i@uW<4PD71iOr{)UZzELxe;rSMJdJ+KMA;eg+
zt&giR(o*(uKQjECCR9LpxMk7jE$9BS80W&C%GhzmYQ7szl{!JrItI0_s^Uy)^Okuq
zDPo%DKhYR2=Q-pJcMTrv`L1R;3Uv(s0O<}-REPfnM6mt3U)Gp(eFnI_V`eXQn7n9=
z9MWn30H{ZK!~Xznr_~>w<7sOSs=it{fA68>dowK`#CPy-nNgs1OjB2+(?cbA-3eJX
zT5nC6ON*pLcw0~TDCp*4a2PtI2oOt-rN#~1P25j1N!r>Z*H*;hStd+~*bD6z0CZZ9
z0csNCN=C4boo&Jq>C`f8ZT>d>BJ{>r45(;PbyqTMGZE`F3Djy^_@d)d2$gpl!1}?B
zF*<TS9aXi_akFu^0a4U0AukMWw6*OIVVTg4JeFYvq=yLTnioGzG>|*^bAbuWjY^|N
zotCOm>JH1)+4-r7SIL)K`p1eQ_Oi1LDy<F&P(*{|Y@`dAL)c<8BT?X^wv{DVsbF29
z?P8}>Zep2SuQNuJW@OzLrJZ^5sT+WU=8D$y9xhtNY<cn`m26X^y#uU~ntX+?&1%#2
z=wa_^HqzF-^Cr>BcDBn450iTx!pve8ztvbQ8k-U`V@_gvbOq~>#Vz1)x0dL=gmoWA
zpvJfeW|N_Dlkh6OA+?SotJ(>dTGM7UY3`1F7h2+aZ*IjB7^7#E**3Af)9tZ7TH5T?
z<!mz%cp5J_ftLyO4Z~qux-V22>WbJ`G&tDaLx(3u!|`g}r$tJcEMcXJVJ!k;AcM`b
zr0zH0nzmtVXDwctpP^Kx3R4<MHZD{tPyFBkf3yPhzR~zhK8e_Sqoo?gt8&$DR}Wdg
zR-M~AJKqIDtszhED4fkE&iWCvfYfQqLcM*_+E)-cqk6DPa1r$Qu7fu;qUhq;4gTod
zUhF*6qGUi?^Zg)vy)gWS&MAK>I@xJIQ73dM5{2z|^2_eaiTxpp&c(K&G_XzWNU=IG
zr>pcNa&BjMF5?X50Q^zx!O%pw_&0ACK=E2)ud`a6bIQYEPSopTH;ShlDucQ*eVhZc
znGTo@%eur)vL)dt^H|dSAKJVfdhX5I7UqEs8FL1KyZ|*OO{Yj(creUY2gFgJ(lQ4`
znI=a13Fs^r(QFfgr~3vj#1nGLl5~l0I9C2px?|AS#WOt|I}DnK<g;-2YJruj)NFQQ
zJk<hLMt-XEy`7RT5tpeK4N6p+GZ<9GI+dJqw4*%4ypIE}T<ihvagc{Gl;jgCHOZY%
z4nq?Kb2mx9b<pPNK~9-F9Q1@I8)~F1rpjMguSs}gn0W<s%Pmm{$~tBRa*bg%D0!mj
z$`gr3$=916aThzfx-qiw6vXEPTtU|y!Mv`n8C!MF+gOpR)aGyn7Kv}l&eZ1S*dlvO
zK-i~D*m?V|5Ug><x}(h+RCO*SUOiSggFOic$vOe8?H>J`e>XIIM1@Yn^4OQ~gA>bQ
z8bf1M5|AozQ|kKfOFAtW6QRJCsawb0kn!9x%9E?MtHowtyOcG-#$qM9YZutwY`VNs
za~av237HTL#=7}$$yn*Q*J?C(rX-q^Vd=V%e4JBKvW3`m`t*kBgt)IxG#+;Fm8ooU
zd+`;k(x~cfd8STN4$9L69Hcr6qCwQ`2oe`ceV!hbN^GcD9y`KBkPO~hOilcRjny|4
zeiX8_x`jfUom4wpjLc2BZQqNz=FnOnHk`K)eYbjJjw$%#*}AJP3!+aZI=}&Fwv%ut
zQd6MZJUml%plKh7T#{~m+s2Dx?7p4ZDR_;W7lOS#^PNOqZu0n%=CN}HQLd+1?cSr0
z`B^KDC55l1)3XMiE?Xtg5N!Q!n7AD!S0Vz2bqx$Xu{jD5o0{eILzIZ!HE_4Xl9UwO
z*FG2|rv)}8-P6tq=7=e|aUI<9f=-D<-336A0u+Lr1l-d<stEulM>Gc$bP^#Dn~?Dz
zE0zh7B^gLb$_WJ#WxqU`YafGKbJ?CCbeu^kD#WD3+l(=+V!*sQnGcJDIP*VhD$!>P
zr%t1QmxJU4PlS*b#|UHJi+ib$W<1P&1o$u67R@TM)!mhB#-%3Ph|83zjtAw<hZ_88
z?iCj@AYY{(FAZ)RMhHFAr-_xfUVSpDHowp=b=y7v03_m`Yv+Am=$nc)Zb@BfgCTxt
zKa^wwm$nLhcJIbs!W8CPAI~=2%-J}9V|;CJd_@`@D;0;TNX+&TV-|Z2&Vs=B?JB-H
zl}eb3HCR1XnZ5KP!&7s%o319D3k>Zfv%?=r&e!+aW^NA9ZWom!Er|X908JE8+ij%^
zWXezzR||x$${uT!KpG}z$_ct5WwE_tB&jqf54EJl<q-x<DZqsa@opH9FuCZOcf%wB
zst}F6*7JoVBI+^z6^AM@36q?T?a-wWWgh2HT@G(#CgpKi%II_62}FdaB*tLhbml-0
zUD6*r=L3wrSj@w^oA_-SPGg+{#5N$_;l?yqs)~)p)O%b6o?`wZ^-b#1jnc<fal&&;
z=ldffS|{yG?)?!c=jk}czqEZk%^s;FiJxi%Np-FS%o1)*rf5+uEo<QWKmoSWNCTPL
z*MOK$Q@do8m*Zf-l)>=1F{t5&pGj#Co4X{4I=&|nierT=*_!sn+aoY0?zkTjFhG?d
zN|h=uYoDP8Y+}T0Zc5fJ_Q`gudKAJVHrVW?EZGnfBmIt9{)BNg>zZ{1RybwINewwf
z5J4sftVGOrqGfdKsmntg4CWf#k_?EQr^#HVIk_9h)-bYm1Sq#05rCNp)TRy|{;c7H
zvPqMtz8W=5+=U<^G;#?VAv18dd|;3enV;=JH?k8#D^$IZq@go1++Nsngw4)<$qQ2`
zOpLmrAwkw2nU|zg<|JG}o^7pW57cx>NDK6%KZ=0v7W~({ec7P=3b%qJV+3+{$#kCH
zvDIF?wSA;w3SAA!8QJN=8I%72YhqJHw%L~j0$nDw1P~Rjq~dVPlxFaXcvoOB;~f0?
zKf7tQF|T=|!Sya=xR*5gplYn+4RKzlVA!8nk68R<H<Sn?Rn0|b4ZDFO(Ri%Ed&;=s
z*6K7$6>}X)kbyA!ba_`C>erZQ+wJy)NUQ9+DPjw87`i+>+<$og0O4U7LH;NE!+NVU
zzUqZcOHbt<F7=O-errKtrn@dD$I@aM3Y?U{kGo@`6!>h@#a5zrF87CjRbv?sqe`za
z^AfgGGg6aa5|v|7w}q*&8;zfj=LhwqUNq`qo@(HRy0Ty%t=X^X>brO_RZ!+t0-bHN
zcP}1e;u8v`E^?OhOwBfyy56fVHg~l!->xg-NbO3sEK55|Mq<Pp7M@NUA02j;&CGQt
zWp?^<BbX{A`HyhK<fy>i0tD1X(drFP?DJq46_>4e=W10st~*-uotcKoSLlb_QRXXD
z9){Kk+@BXf*OUddNrYluQ0xHsxcye}*>hA3Ls7*6iK|q_oUyulZco*;w7lfNa1m#4
z?jH$?r-s<EPqQN_tl0tqF6D1MH=S(-ad>g>icY(vh|*;8w!IC!mwiwwPSP<x9D1it
zq|DRgq$-n>HJleHFKD(z#JgtmY&UahX)hC3r}S+KMx{D|l^faySUbH;TVdV+cSpQm
zYpM@ubvI=$6O?6n^RXrlo;KF&Z`bhV9>>($HAT9XQw;t$M6S~2gS1GQepR<Oi6MBz
zt`g}ni%(-Kv>DRvBT==s`W@AFm(g(z8tMqK@gKO@%7>xj8nqs;Pr<uMWwxucE@)D#
z60Paeu@U-%{{Urs`XcAl(K*Os*S)S3tJs|gn0JRKboVJbn;s2X8*0<5d^xmp)Njpg
zX51$@132w{7bqI7O;p=0C*?_uFSGh%wX9+vJ1a|YdnAYTuYA*Aw6Rh6x<a9$h6b~@
z4bXoRv}}2up;%@a<-M(PDUNR6!M?HgM^yGl0@ra#1m@Rkp*1gdpkdl=?9v$eUzD6*
z>o<F`4@a$w8Mu?=ANp*N&*nCB`pFIas$#t?F^mf%^STSalzOc4lN+vqq|8!UZdzH|
zaEzlvrqgW=fY5rVgVPpEOBs@kHUt0wfzS;C<_li(>cE-Q%%3x<R{f0ZuLSKI%a>rz
zVf;|*c*>tcc4}Ik*|qys^ACrFuE=@E%6{JQ{v-BJ65b;GFKpUu+GjBj0>EA;F}N@`
zUf=B20|AetPzt@ANi!M<CvnOz&3>}>+vE+fYzO51UvCr2IJG(Hmg+N76vwnXd%<aT
zdy9(A+~T*TeW>6=hVc$Hi)gQj%EnM-GT}+3rPUtiIl<2Chphaeq2;ZDNJ^D#9~#8g
zz+v#fQWEDlk{n}6i+8L8XTYz_Cfa9lJhf(aizkEHrQx<rfr%L|A-CFIIQoYU^mov0
zXI3>_McnOthJksWAaM<<>MW)yG-wA-XIWhA=LvDfpfIuvPN_EV+Y-!lN(v_rS12<=
zP*aw<!kN`LD$PbcRnYOo9CD1x#Xj+b&a)}l=6%BA2pI78%7f{aIH-;+xT1NAgcV}L
zRN|VNc<t#ARyfwFk-F-vP^Mn#iG-T5-#Y7j!M+%D<ggNl-Ox5#9~jO|F|{lL#GN(i
znU-jkX@HiNux~d_makPll-#@Gt$L-LLW3IDKf6%FV|W)QK@)$e3$;2^g{t=u*)0-z
zNj_(yCK01f)C&;S@l3U|b906TC?*8fm={W@lIjlrT)kl~an_S+RnJ!uV&Y^?qI9vi
z>0osru5}z38rnMvo&H+&Gtw5oo0ejAhh-+^gry8DHjHi(A66_Jg@uj)gR$lN(Qp}q
zrKM?nVJ)MNwntB>&Ks1rC+ZWo6~y4_$?&#o<4#MTKU>oo*p~+J93v{bFk_tQi6R(m
z<I3|o#m`%mnob(IrXpGdc(`1C_0eCj6sywm7=#Sbf7g@TUbjq30V0)lJK3vHd$Ny=
zCy4>hrs2C9ekA=>mUZquG4723RY%yu9ECh;3p+CO(1e_;B>VR`0^^<$0Ydr3xl?zd
z;_ig{hL%v#IgAS!(2`7O00!rH3RLNKq0e!E!6xEPzyN(?CryN%lq#O#x;h6=i;M__
z)UEczk*YHCDX}g~1cadMo2d$Tcwi>!!f$=cz#QR9BXk0DD~H68rpd<PHZrmvWFVR4
z3Q7}WE|>{INEOTQ>i1)vTHS*{12J`)*^I|sc4*bk&<<pPX3AymE+zof^8HEJZ{WKd
z$cmRLK2cj1#Wf6?rc2{0D-Dy6Y3sZ%8~pzOs%+H0%y_;&T~iy`6Q8ncUs2l(&S(G)
zn%QV;dXh}a-{H~d<bjv8C9E3g(>^4@C&&vz{e58r?YL74)J=mKxtTecm$iU_!oFd{
z9MjWcq+sEI95#uFmXq-+{bJ*as2pYp9{{;?Z7u%*sxe`1`O*IX!fGhiBslo}7iNoq
zzd0YuCnrUzZ6^ic$!ij%SY|+8Cs@?c)SIceMxS9a%sp#kPleI+o*!M!Xf8Dv)*T2X
z%@8lR=kr`KP82aL1i;Ek<zVZto^+`vW+~DEEXtt}>2`0-yX^k}2=<&i6e;l9Sce|r
z5byoB3lsbf)x+2)Fz%8DE-H^{ihBs}ZTW$`e-P4l7B{+eNW=9S{KVv~Fk04PPGWi9
z(KOQ@WstqkoFutz@Q(HHiAJqj)UeEJ>d$%2Xn<XJ7;L~cCs!Wm)Q~Qy)1y+m%{p;l
zO~@8#ldkc$r(mmv#--mKWpjzqAh@8iabTWFwR1(omlSh*gG@^5#?8`dU?4!(sy7rS
zI&Oym3<X$lt}Y&^>bVgmGb7!1I4q45wxxvW$3NOy;eSAw)bsrpYGvr%m&PJr6+0+8
zOzo^h^WRN6spOhnu3@Z1_q3Uh!?823MHuMrznMdo0^Uj`(cj$;+bF?N?sdS3DW`u=
z3_UIp8z<~L*~I5r{(kF@6CcnHq50MJ`k$rh#!eEc#Na5<Xw3~7pxBu((smoWayz4P
zFV|qDim6%Fl4GAKj(>+p)_K>A$*0BE^G64!*wGBl-l}sSm+<oOySjx;xX*Imz5f7B
zIwH<lF+?j`AeAZRfx)-U-K6E#uhGkQbWyIz;&^t*k{&Z|K^i!2Y($Hj?iH$T{{Vxk
zV8AUgw9elBHnOXH-Bx&B{$T8g5qZ0<AtwO8Wl}lgDt2zH*F#I`M&^BJA_m3?8h}qi
zr9q&!o<libGq}=Zb?Lk&ZM~}}Ea;whGzE99!y1ftX-Ty?fPLXnaLD&Fks&m5lu1By
zOdKng9nqjoP79@dlSeqAC!%6y#6GMECZ>L`R0vEA_j~XmHF44S3P6O(*DXj8nVEcd
zz?>MB9_-U1L?q%Th6F)Tb#FOHtB){QN4kHjid;lb3#>T+f6NuEFZ|mFIUev<+r+qG
z>?MeQ(cN;WpD9YpNjpobYIJMLwJ_UCzM)_CSKU&K{u=RLgHgv*WQPJC$eh8Cb8`W^
zng*J<bpdy-BN+TU-5a&jq=0ghu2Zi=eF<9Db=Mh~nPRUdVK~VUJ80P87qvW5q}F6$
ztAVHh)B<xdHIRK~`4}ICRW6rh#J1qXhz^}EVh_4{t(mndd$@2g&dg;pC#Vvd)rP~b
z<%z2um=#axThnji)3HwlS&wy2A&Lc^xfx0Idp3w*&9vR!;3h)vUYUkrD4gAxGB=a!
z9I<`0C36g+X=Cd5^z>$&{{Y=g2l=8HZs!}U!R<NPeqbH+T`PCEZ6}?8_*q&@m!>v6
zR(hVI;sK|!?JBKK?KoxUH6D}H(l(Ft&6ng)+sS-lLlx^$FL=aIc1Mc`Y6$!<9@Bm!
zR*}C(KPc?#2?uUnM<eh$6#oDcV6vvs{{ReVRRo*B2JVl_mS~kZ@B1_3{>psq@Rw6Y
zd{)MeBvq(zwcFXeT()LLnV4;vz5-jLfHx!QmX6k$%nto|5wYpIpO)E);d-VWcWw++
zXw(VmX-cHLpUC8O%FRAt%vp9_nTtS?%T~gC+a8;Ba|Awb#tMP$;qd&D9P(nD2oiHR
zYCzOBPf(=VBxN<NIt^TWwjEBYiBh4{D&A9DHo%xpc0i=4Al#R3u3ptXb>^IK&SBKM
zfE=+nM}A)PV9;ag8Z2=Hz0EN1R76<w5wDuoFAb$utQ!NSyXn?+T?ewil&7TXwTY@;
z4R%6)Zg!wj^$x-gHLCc9h!Zb`mg%jD+=cCpDlHBfoko+hNj|RRY+~H5slHkR80vx`
zT27rt{MN^p94|(g?Q*h|YJBgCV-3qw8gr`da(IYLxTCljRzJALx2s4W>8ckw#JQPV
z_c}IqWAmm$WX*34Qf6vX6dx$6ohqR?@k~s1Trm8@Gm{o|5_ZQ#z2yz-5OeVKWwbE{
zpFr0Z$@6!?N-ETYQj1^W;186wzZL0iu2ZoE8SEX3w$il;ihz`{GgP6{n4;9cTPF$h
zeqpg75V!MpiCQ2ecE>Y~rpiUND;lTphA_{Z4f(+ww-Ok47Y8zkWN(?kk8v~PW14BB
zj4D!ia)q=VzJ$gMv{`;lomU-igl!3(0vE`6p%$ZxHu2k@Va_;_pvL3tjHgg98pd+i
zaU94eF7{nN;SEPS3rD;(CIBb$%!%X(6XLo<4`bRc7NbZ}u*~ilN)T~0#W3p9fPVUZ
zDtG*DDAA`;xOz>eKvRyHfP;A**H4K)YMa^GyuQA9>y-IKwmdAXjcuuoPCPh{fBJ1n
z$NvCRvY`|7u+!dWm~{Lyg%O7i5ZA|ukCTeJP<FkA-k59Uj5S?iU&L+Usfn-F&T+-u
zL5H}?H4lRc9`?`gW!1Pd>r|}4{Lp?YZtP#mF1pv1T@wy$Qx440SQ=FxYN3i|XvP&8
znNHC$s@`)VhbV`dh%LL+3*#Oaj_^zu+CQ{(aXe<Ds)n@1BIi?VwT0)qV)+&z$8?Lu
zm#N{H_i(-$R}X<>TO<PLYveh>w<bUviysXa*e$U6eq#N%)UnnNuGSp+IK0g{!h6(<
zI4`pFICXk_ZmD%@kAnBqbDrqgptzD6J(h@Q)Cs!e{waoVtHbbg+el5afG4SomGb(t
ziEI6>`FYBbxFTLMk8e($41vsSFEduIBQXzj-Bli)7x<oKBEx&O3xPSVLSlL*Jy9tt
z5Dutz23=h+<rzl79ijT6%2`U2mwQ3ca5q~KBoPHj0t~I2HAXB)gq^L(4s$Y`Pmik<
zTU24OLXAsi(P%oCq~S$&4_9BR)pl?66kf~~9J?*ih3NkPC2aDmi)I>B19wJW-Y@z7
z%x2cD+hdu8Y1&WCXwal-U>%8e`Ykp@SKK{OpbeD>y4TV&lC|Bc;?S76N<xK8Ww%@4
za#_b-n0oPF3hqdv;MegtJo^l69l>*nvG5;5Woe|h2pT=78R`Q-A2OX#%6^q0OP0he
zTExfmT8Ou7yY3(4kMk${CE+W$W3?*fw8Lueq0kxU$j_14bA24{E@1NgIgXn6smaNv
znr3P3bZVZ@&(0qR{Y(6LeJJVLI;bwDU=3Vt?we0jVbtjifwpzY&<p$j08LLtq;W-P
zUD`g$ik+J>6608yDT5Kd&d$)2mX{Fo&o?5LAc2g+b|)xld7mbNVBFY}ONGdrs}zLM
ztk>Q{M0J~AYlwlQtG3D|fxIU^?`Brp`D&(GM0b#R{`hzj)>Ep+2>aLk!Z9JlvBdig
z`l~C;nh}?LYHMXhz|5u&4T1AWq(#;0Le%PjEpVE;VrE7FgP(`hhN4817EKI^cwxy(
z6GJF2o>U_dC@!8`AUK&9q6?)7u`R392^&IaXL5XCNXl$$x*mvvhJbGsjfyW1*%;+k
zqcH3aAek`Mja==3`jcjk@v`X3W~-nWjY-fBf6!`R1%fR902gePMN5a@h_RiXY)&YG
z;yc7>8ZVWzqei^kd&iKIoN5tDkbyQEl5H;LenP`}`eo$5Nu}Z4oB2mqo1sii<C+E;
zrcBdZm6sO=k`F%fubFz<ot?Dy>f7sG^Uf&xn5+@T)Ud_ZzgldIAImQz#h~mLslqX|
zu??cr{lz<UbF}k@&xM!iE6Kukt39zzl)qoDtmSF^Hzz+gJn4DKho&XleC}Tw++0dM
z_e2~8BKjZ>>*@g-Mt49LFaWR;Vl^okd27Ho4Vl(*iTCT~wQkvl`aUnW^W*;jhTLS;
zi9RaRZwvE><>@WjcEKT79Kv0_o=P|+b^id2e4{$o*|<Ste}KbXhoCQ-xF^AlE}wi4
za-+lxoGMfzBAkDS$LhG4WaM6y?J|Ias^!xnADu5`Z(v+-Tvr*wo^2wlKZWodJc$8P
ztj|foP{oc#j`41vs;<>#&ZO}<rJYmqpZeFo^uK>pY0`i8cnl}*&4>Qn`y~GW+a_Xv
zyC*5`!4#@9x0ey#8vbseek*@AqVq7-38~tl%Q-ap8;we_C<I@dX|Zev^id+)11RQW
zL`iWzB*{KzO62!Gu@8Be0C_Y(_W%Hu3-(ebTmY2_D%(wy^9)_|iYmSR;;mDsEd;8)
zOz-Uv7ID>wWx@NESu*M&j?%e6O&l%vL1KbAAl=327$z*9(4PqEjj;k?-4QAk_R8;m
zCLESn)kPoATfJO|F?2wPO<o?MFSIa7iJcG|C3->Trzp~sE^eA8uDinNe~e^cJ8B#|
z4*Ff-S_y98nIwa_okOk;h865jX8cuF)UghLdHDXSJQ&s|9GEq3Nxz?V-k*o8zN$u_
zY{Jpl_leMa$&CQikv$}kspL71djYBaI-Q4%*$QHK=Ow2cZIl@%N?S>rt`-{9{iWZ}
zy<(;&n^{v1YCBh%bU~8Cevtzx+QU+3wHG<<tQb9!0CaGX=O#}2s<i14&Py^w^MGxA
z9+J8n%}i!mP8o9tHQx!hvNsnbh=@tKC<6AG%0q`O;`^R^T5n|>(CN#v$day~oBbXb
zv4bVtK8Zhs4|Tqajd7O~P!4C~bOYrU7bMBcvVjm&M=iLZL_%m~ZO|bxanA+NAuw?F
zz><$LD|&grlwCJ)w_9upz)>?2%{Yf}Wddn}%gfnq%ldGn3k34HM7<{}`q_R(3tJ<+
z5YAMz9&5BJfi&RT{_ZN{Z^LVk$bJ`Hu+rRJCw^eJmfNuPZciFvc`1jEZl^th@V!Jd
zrTER4=`0WqUqUlD9wLJ{0IgNRm}N@Mmk<T29n4?iyW?rpem^BJTwK-|0F$I>JQb^C
zn030YG{W_Ka>TH|M17g3{!yM0YHS1fvv1jRji#-rOKA}g1($D*IP7a_;opbhfz=yc
zM|2i|b$Mf)14u2IgO!BlBupz5Fz%^PKNv$zK4cdk!W_{g5wydI8d_BF>;_j7EZ!iJ
z@`c&;TNF*qIX5@*me*uyFFi8hYB=tkz=3fZs_{Jt<M_l1Pr+88z#s~Ag+r7J4CJ|>
zMZ2`mxj<9E$$M%xp(bU|$S--y44*dzqN_U=+N771rq8x=#UfIBA7UC?%+jSWono6=
zVyCLnn?MIB@vvp)i*X)&JY3?VUqQ7WQj=W32qZkX8Hgp_);I}1$O1H0982gowjk;|
zpaB5d1ex=>TdNRxEqPX8mC(i>O10s8h_oj*Nj~$0->!<oe17$X7Dhh?=xZ?`b0ZXD
zc@XBfj*<%D=<aaN7+uiPx$^{1!@+C(BTb`XY+3jfP}uytb#ZXmp1M}OAmuXNvNq-3
zH^oQLGUUD+B6YDi+%+EAnAVxlW~TY&e#wG5x#Q@f%2;Zy%4nXpE&l+_uG`piICEy3
zPMX6TDe_*#vpDv$@UcLyyzJj=eo{=^n|#}qsYx8GLBfjZ&T6Y-H(e}CoMz(c@hol-
zR>tC}ma^nVP#o(H`(D=<Y#;`^Lm_si%qAVhtZT2<XQV;++-kOW{{Uq?En1%1-qUbx
zOKOu?@ZV~$Q^bL?S1NUBZqROo%aV&wn__OFHz?mQ&F@)dqR?wBav1zuokI`AJsHgw
zYddhp5-vMH#=W7!<TcSwZe)M}Y0+t@>IKqnPFWyyB<Zw)AXqPycoLY!C&~}XCl*T!
zF0qDbRo(}J@Ecj|tdd>STB@C1u|~x<c6>lJpdQVx_%|r`k{35PRpI5%Bxi8=x?{ir
z?q5ZA4~OMw*gBad*a2Q*LYXXA%X!=TFK8_Nw>=uPKx;CAuo`*&RVvcwl1}I7K8n$b
zHm`&*pUM)gSu<6ZJt>)aGHhFtbHWZdUTKB*c#V8597*-5(y+X3Yk19vpOTAcz3}ZP
zLU-^Rs(4a1skG5A!b*`+uzQ_SGL4&WOpnk%I)53f6aN5JjznUyLMj}9>@0s@eo=b@
zx_`rA&v0iv&YoL)vmT*ricZXM1ty;TF{~*p>kkE!E>tDkoO){CSe+l#Hllb$W3y<`
zV{$cv%uM-}$@OZ~c6T`b{1V2Wi?FEw0MP2AlQa+;enF|!9P>l!vhn6keFyai)i$7-
z)ICF(awg-zokvG|)9zLN^hf3c<~Q*d@m2aIYGRW$K3E8*R#}fdR>3nT*{7-3eqE6w
zv_6QDHeZu4F0Mh(q)V`JK=rBUTRXRx@cO6OA$5k-Xc&(G@Mp=~<DR2=7F%kIi{vY`
z=4#YRxjQ1^5Q*MnVUJJ}7Z4Emgtw$vZAR8$sZG)k&ygB}HwNKxbvMy=SBUzKVCX^E
zixnVsDThyvV3+O0r!J~DCh2=CseFxzyrXe7sjm27JU+{X#buJU%7NwT_#?*Ro9SQr
zi-Vy*-O$B6j-|M2L*1D9t?V<Mhx;vFq<YmUx6Eg7{b5ZrR&lwG&@Oq~qt=)$%CSS7
z*Je+uRUn(TYscb*&gXSa-Tv~0I+HE@<KmImG13<(9F;DwjwB=}8kd%sWQNNCr9fK0
zoExC+2(e5s9TaU(X_8cQmnn){#7=lyA>bdi3$<?P_HHD?C~cDmGS;h)Jca6L@kFz<
z*``t~dv#47+r|;?4Kl+I81)wfM7SmxLg-H?pD3N5$YZuCn5ps%89wO9&YQoPKg#>I
z#%~j+Qnm@zaQ?F{-x0T()8UV8ELf$8^<QUs?yTE(qRH%EX|=2sh&3B$`Q1KeO7^Ga
zjBjnhQT2sW^rMFh;`m-S!<&qt>Dk>JEI}O%g1Te)Wquuvb{)gQFC+Kyvg&Ur+&>|+
zydAO)kufwWUTTWXTB|R28*aMZrvCt=bZ78vbuKN7#qnl1Ii=;TI}i+f!o9@s&cNXC
z+&DG}Y`e#QisdRs6`RHnf6<0e?BRDyjyIi)qlOV(!DBNAUlQq}*i-?gU@RF|+TBkP
z;xS6MZ*EjrmY;A8g^3_+bJH#^HCW7TYMA^+D!WMmrJ`lS)n_0?+?$c$ul7{1bZFCY
z1eq}a>U!v@^+~)xj3j#n1*CS7p`z<<%#mx4iuF9o-*k7ciouICDblLm#!L3OFm1Uq
zW1yEE;8|C$=d<r@-R3?**W_$By3v(NMhU{z8ChC|NZ*v<XGxaY)TMnNHn$J5>yqCW
zDC+DTI&KvVybE_g%-;(fz;c|>EKZ8MYrdc~#5cs4J~t40=uWDsiDxi0<|@L{6sokt
z<=JOK(<eAD(&A6Dixsgnu+GF$r$CnnvY`e^(=L}9Z(+!V-h-<VQO_A*v>QdH^JTp-
zmMrYjuTfr>L>bzY<mRpXdXGM5y(7O2d#cfWXPKIhGtNl=0B<C*%oK}Q1b0A>50AZC
zXDUN3E}vyGCjDO&YUYr$pz5nrn5spl#GPru^}OHB#DR@cH%;HIWQ@2H_cG^Au^?1q
zntIIRGb3|wP2y95y28wYbY%)r0Wx-aU9}<9OVh)2IVV&=P10~3OWJW4=)f$I39`%z
zfpdo`01cBa(!-=c%gJ!{d$39}@<=1y@Pcm$%NI<kb1y!`!<8x4n3i*o&pX3!hj3;?
z_P&&0(c&2Dla3e)m2*gIn&v=gKC3~#qr7!-9DoZ-rW%7|3~>!)T!I}W#-=UcCICJq
zTQJ0=ti~aeZHkL;7<tKU05fU#^vH$0%P&X5EAb91jW`;#AA+evFuR`?Y=C1Z-g5ca
z726Gh!;C{}w!3OU(X*^tXO^qHJd|<X*sv}T@}ixmU>T>d)lQ8``C6_r&AIYMflLbo
zUVD?7ni5*)Plqhy%<*hy2NA-sd_I#OPq8s^6PRBuq|1!#weUD<)bZ7^Rb1D0nnBVz
zi|91ItMs?zuP>vpZH-O4PC03b%q1y2^qarzrcK$Vh+QdYr|Cs%HJgf{D>r!~%#-sa
zN~;YjrKf9qfFDIj+aLW)j6bdq>^BtOc?X*6&sZ<bSC?i8nx}1#<#|$;dQPE*DD%_L
zJyy{wGo7w>RiZ($brIojFNqv%@Vs>l6DT^P9rtw}d#%h!Ee#^X8M!?m6;Bn0t6lXN
z2=6tb<^Ukflcv${T#tsR?l*`f=(TC#VVST1dE=CHn`_Pmhj&MM?)UwL7jc|Xm3WpP
zh~Q{GSRBF}GUsUqz+BF;u9_%e@P?KarWU<c_lP4YoTP4c=&yBb1BPCgW^!X4ixYp9
zY3d}^>^tf$y(_vdQsWrvwj3_2gS0Z(AdRL027Sw&P9H!F#}v^Y#UPIj57ld&Q-<mC
z6v}m8n^#GO-f-A;=GoE`a7OM|EvxHA&L54$R>W1n;wZY8X4bi(#1{sc5J=Qpx^)A(
zbd^&RhGjZztxe(iq&R_MM)%(L=5DotGFqb8ir;i>XS9_=AtxwXLZeUlE|=4j&JLK2
zQ*B1osbziC%{oxE9b19XYt@T%&7`U|1Q;7C`-L5IMvIhAp+~q|;z847q6{IA5<pzg
zfC(ycLA$L7KFo1@w_RA-Fi|R$;{?{^+$K~L0XNfum?~Vwt(ocP;f}zQa=4h;GY&Vv
zc|hj?$9Nx-Z3Mgl)8?B~c&OA#24ve)?7ELU@wh?(`NiVp8F&Fvydw`;)ky9(0BdBE
zu-ud5%#r{q?kKm@YOpXK^NA#%dD`E-M6{{Yo17lQCgQNue={&G2N8PaHn035Pr_7l
zi`!AobY2MoyTCIViHPqRBnuG+*9`%VXpN*vJz_pPY0)?B;VVw4OV->%b?q%8bj*e0
zDTb^BLEz%_j8?6OG#yH_>Jv2@*%pbAWI4scJWTUlY1I!E7#ajHm^YZ*KoUL0Qs>!V
zo|c?r&G8U24HpHLmvLto8A;fB)T>jhVZ=61dc+V9c%4n4MT<(VUMn)C7gH7!Yj1I3
zu)n=C((K7Fz~;8RehYo?;}-Blj3!5WE_;2?J6;zYKnD?+ZW?{<2_<NkrYy{^Q3;cU
zyQ&&WHOGBeG)TsZImdUyN+mRL(S%PVK~e8a04EV53=)VbvyN`>wh1Ur96jA!pp+`a
z^+d1XB|%7uFr?9WQA;_x+w#gGDrzobdgQl}q!LhxDN!e8WZQ7LxFZ}ZjMz%8{Eq(s
zNmHa^yAL8hVReSvu|-P^Vb3QKq09(K)%IOOo}KLh+RkQWx2Gitl%LG<v9L>eQ3Hfy
z&Tu>?4;N1nh^*3GPP5twbA$pK-~a>wWyVN{9m>(h*RM#jl{Y{wxPh@F&(6v_73~iZ
zY|CUjJ;IU`O*4wIQ3f=u6ty-@l4hLQ8&rle7Ol_cp3k40m5mx$TsA)+TD@G)Z7;iQ
zzM+$?LnYEg!y;k@!r<z)F!**;rwuOe7;A0#Z5r=7&Wl;Si|cioVb$q%f(ga|MtWVA
z+Y3LZP98<#^e@C?pu=Hs_4^9!%p7}!mw$lwwFA}CMYRS)Ql(m5tuq9Azy|*Si%)+e
z(NQ0k#Q3)t?8c$;z)V)&t%Vv|S<|)m4dK#kdTi=ArHgJG35^*kLvn`@E(|(O<jh!)
zn(Il`D&ZVl={cZt77?Utf9Sd<M-ovsEd2bnXB*Ab_Kzv%n5JQDRP$6SWXF$MVo}?t
zRC#I0kdA#*J<;3I+r&V196M;R`%hr)Nc^oL7!Trqdy@vL9jmAuuc}@~!N(}^Gjq@$
zg%<;D&usJc%u9=En?-dn)7lc)lB-smo6{itlPGe|jO(u0o@iNjQz7CZZt{p$;oKsV
zSw{~;zpG*i>gQ(;W?#fy*qtv3ZYLD-UiaI1kUWTkJfiAqx3(wti$IgY7&C4)ia%$#
ztj2hjZO1VS*W4}~=sLZhy2f17XnM6>cg_>1gl-12u^5aiD`wuC8{Hg*ETVsjrOqSS
z0?9g|tkBAJOJib9{{WkB0|07FtsjbN+kM$CCc~51wzpl7#De=(DiuZg^x1ZYn=9Lz
zVc>aMe%^O5;l=u2**NMB7<SWmf(Q+4iPSl;3up5Ht3T4HxCgn-^FKL*<W_;hn3}UU
zMcPc{wBRY?8I3vOdE)}KpVswk%JuO^Yigd}v$f5XZ!k$V+h`<_(O$`xl^|^w_<Q$T
z&W%~b?6))6D+MmgqT@tzQ%*~=^JT(8m}KMWyTf8ei#;zDA_GCtgKkFPN$495JG*66
z_B4Rp?hf1OFb^{6Jc8QT7iu`C3(r$rk)%mGGbea4H!&yt#h=Z#jpWp-e^LUg`9g9n
zvV|Hz5q`g*+vnABby}5NX49`<6RG)hTQXnD7acb|ElbU`Y&h@xM82k<!_|(NMp8*o
zv58hUyrJ={56v{(jFAbSo4K<5SbbiuFSQXDL;+;4{?izWOoz5nB_848)6`}h4rp6i
zQ)oLm+Fce>eyZrPP?RlZH79f@sb$ls+l}Nw_CC<ZC(>)k%rwxAM}hiWbW&@a$;x&2
z9}ugvb*i-rr2ZqMRK{{-+7g(k$}#6^4G&06GSAQ3`RP%;JWP<qYK^T)=DT&=#hgDo
zMdmbnWR1iwt#v22v#2-Ik>=8L{uA9i<x5h$p|FitOMt+OguG(ee0*OY?DC5gYPS@^
zSz<M8i5+cX)3b$-Lm@(?Cs87FIvD-I<~}gCB{v&ks&*Z%Vp7D2lxns)aiDd_cnQf2
zF-~lo_z6uf)-mPiG}iwB)QaSj_gGW^0Nt+u9)MG}`cCU&q|^H!PfxtB*VthH0OXmL
z+OSY++f*a|s$7cb^?6Kau`Rk7ZPNt2(F2ncA^!lfbpDJPU&%0Rs^zx4EHiD(%oO)!
zE|;{LG5dxXYl?}DkzU2SDVcQV=5ZNnby`h0BwtpX1+P0M3@0@`s^F@`gHi!F*JZP%
z`3H5b_FKadG|3RADrM1Ti%7!NCLEax?5mPsK6#_yF#3)vRHIP4xaja!nO&i<_R6_2
zv;P1lN~^MMRB^22r1Pa~vo!YGeVKCjU365a?lXj8bry=Tu`qvrMO>ub4~ceKq_e0J
zJ>R4*y4h#)cwx<^v)cAmw<<kec*5qX^EDu)&ilKc$PIqnWLNu5ONWx)FNLc@Et&(H
z?`hM=+_YQ49QKNs8ttp#4EcyB=)XeS5B^R{{{WN*$u=d2yrD2t@|g1gR%GP+Z0kZp
zT%d#6wjmB+?U^Y241E#du)o@v{ymJV_BcDU#bqupYuLco%U<>p152r&S)h{a?IS|9
zaX8!-3kgJV7YeRxNf}%mTTz~<aMU)S%1ns`r0V5B^3U>M*%k=eW;xsMl};&YUX3tz
zQ}FdX+wBvL^;l<bn?lc|E?kB<aBP1c;iaEsaJXK|sFyzNd8eA-I5d_6mJfN;d4VBz
zaN(Rzkm4o9%%;Y)mv=!UPzS8?Ow0fW;w$3j*ZiUA8%ug?1nfRdiZ%~w(9@^M=oHI6
zJzI^)`bQ4ogAWz>jYC?Y)RQrP)nH?B@2U+P3q>}LLM42w{1BY4)931^TOjSVIkx1S
zbLQrM8PezBUCYDh>AEj5hO=rKk1Y;+vVjn81<t2U4Y^L)rRJGnXat6dJLx)oZnyRd
z!IkVSCo??tL8l8(M@OWM&tzD8{Ta;~52BM>=}awJs~k_YJwEh<cz6YiH5<Q$(f5^N
z+iO!S4sreb&w#q7>JMx3fUDau!J~|JpGlUJe(70Ce8bMWZ%>?W{u|>9h*xG0iBhac
zW_CWubD6sA97eHwZfvyh6d#DD>Tq3@x5z>FfK;s>%9b#ls!L8Ts}r+qvncDDZ-Z^Q
zZ`$uzx51m)@Y@KN)Ff^t)(5BxzY!(a$^5PO53Z`&JG4ofK}|J3LUl+!Xoa$kfL3u5
zpqbi~g!-7x$dJof_Mw-KY9WqxiuEsu@PyMuKPa*wy-mBp5zgWyDRyu>$kSE2B)d(#
zRTfdD=q+)!Br_9r7eKI3E>BXPlbM=gWDZPjJDPInn2}<jx$hF<31_C8o28MqDdN=P
ziA-pP0?ssfxmTN+nt5ha_%z(I?~FMJS6N-DNji)%Iz}>9*yfM;LOjt5UQ=k;SSAIR
zY7UM=!cGO(!L*@Y7Jf>m!$i$M36i%>z%huD8Z#Y}5=;%#_t|2xP#G#~PPsKM%%rr&
zlMeZpNC%qRZ+|{ffq>#XGXcUWHx*bmGf25Uvy4Con$yJLF?gJF#Ts*^?HU2P+S?q!
zlX$ObSaM3AHp?`oY4vy6n4g*K%+L0DM;>$$OQ?)fCe!GQQ{tHH6SzHMUL5+SpL9ve
z+`=2DvLv|8+({cN7%Vm+xb{OKtPZ){k#c-Zg{;sQTlovm$xCbx%W|0ZVNso<PywIC
z_*}HAM0u^Y4-c*OVyD{&lH3uQ?>;x#zeU@CGHYT85oYq{e!Q%|R2wF!G8k89lv6O*
zYYQ+8>3(PjrG26*k>>Cn-XBOCNHO6hY2t%UCFQ*T07+V-X)=>FiF{2<k!4PyF*egI
zH+Rk%3qmo?^63rSrXT>~zc*f9^v5Xp-xXH{y-Iw}453ZSJa*pS<CP(^hx&0osM_Gg
z;=UZbLG2ypKI|%Z*0t2$0iNji<}2q$1?~8Q@Ym#dX54m14oy^^WTu*yxu7P*#Wgb2
zE-n!r>mHkm*JTG_E8-kL729(p4(spzc3GG@PJ3|-rX-qYmF^cqU|dy6+D;Oks<>)Z
z1_7x`gM>Efaak`r`g{)iZ7&pK;<5ZUf>jN5VUU?&7t21g4@>4@HiBAXD4k3;9}&bU
z7CEOd5Of6cI_t3<@4PCmHDYd`v)nh@B$Q_V00^g=d<((&yG8tCt_K;1#qqlJu+MA3
zqe=@;=6A|R!=mMip!!%QQ>Y+^x_!gv3s~U-X1R-~lxC#dCmom`5eVAE;v0v;KCwCl
zMur{L@l2)$rERUGPSTjJPRMa2wLOv6{<Ey2juNqb0Z@B?>=n#aER~Bt4!$drDmav;
z9(nf1sI@v{Wv(r3ESE8)%8F*VH|yzUeylP{xj<+dD2bP5A9Gc=GH1jdq%84bj>)9v
zDQ)pK>GZf?l{f>*w5fJ&0tieg_@;K0_!~1C!Y9Tv19+4(nc#w9KxJXg61OuL(vyv)
zo|NobQ$xVVoV!dvHx+<5M*>|lua4z*F^Xw{;o-~4Yg(kkGTP?*h=L^-n(fsWG|WPh
zz>E!QuF=Vw7L}$}czC0hCZMN8=G}I~Q*-NhexWkP^G^J^U6Wu{Ddjlv5%l)ETrbo<
zk55;iXz-wF@cJ3Z85#g+)C0NYn^_mj0%Ae!jv<)3hM1hhoX1J&&N}aMG*lPMip|lo
z+>K~!EgLNION_emiCd<bVF7-1^adP#lagxQH%D8}x}Ko^_37kG8C*%=usMQqKACmn
z+5j~3cY4<uL6TsAu3<yVi(GwXz9*a;_;cstB{QnFExzkceB*9dKI~Yp4xk@~AOb;x
zJEm=*BPqSe-%v<gsn2R#R1*OB+VR$6(`8!i9(HqAD}=;&2l_D|;Ss<%<gCZuDmqU&
zi}~uMiEKTNZ{Z&aO#4$#6{6xKXlSaXGW9*Te;EQXwjxW2fn@TwWTrC@(=#9uw4es2
z<uj9Dsm{5hG);rbO?&as85dB1G)S5umzVQSsk!g?A!a9J;004Ul*gPM>y#2wi=drI
zZ5<a?J>OOd<dj|tZSu2*g>fbDX1pMQvTB0F<D}sV`_LS&D7#Y$xffbnn&OC&8Yd%l
zUYa+vE5Z~QlA3EIS5zdm4#v^;u5h@t`1+y`%xOF{OoMMD%}KBx(s_@X?Hoh0g2ywy
zNVrme6tnUK)ds66Q=Mp?FOg|76*;HpW>(>_QMOdqgt>>bW6W^OQNi%+bE&rB?8~Lz
z_B1qVZy-$V&1vFkcG$X2CfA;T87JboOKTfMe`(u5#kl>+9?egmnYv0xq-k{az*&=&
z{{V9=4Tt-<^f+ohg@mVtX`G_xnKyCsbf4(6YQMIs>b%XoG#>?L!n6{>b2Pk7wCeNJ
z`g)GDG<j)Yn5|WqFA6s9K`z^ay7`+^BH#yOg=*fK$Kh}+6(+^Rkxje$!tQW40h`Bo
zF05(8xwWf1EN!cZ-rEDbpzRM8(W@A91Xggy6Oxg`6pFl|X)xBL`?Do5B=2Z+_@<Xl
z6HP>041~1WY_(%I5l*%r5IrWNp5h&IW|oMIjTUX^Ff!t9ON142nWt9{^GE<KZNa{$
zdqf>I8p?{3rBb9ssu-SN+oOw)Hz_MFLK9UnY^>nBF=#hss8=V2v2?n1O<I-R+2G*w
zce2^P^E;=!fT-ndF*R+|blSg(rm+dhwCICWxiB{A^!ZK+X#g&(Q_@~CJBXL8$~se|
z%sWVIl5SjmXC7qyfUwXANzQZPcGLIGR=SIXqG~#ovTzpHaqi7&^Z0)cRhx4J&K9Ld
z4AtbCsnF_W{Up4=t2z6?AYQA36>%7BE-~V@R>Pa4!7X_3a5Lm(y%$zKr6%NQw<B@x
z*=fufD%_?V!IKAhIWs!s$;)b5(P(jr{CUMH6pKSK4DV)fC8XGf0U8Zvb&#!9cn@_4
zYzBhz@1~or^-im?F)Y(JE~{W7=r}xJUf9>8QI#2#NG;SG5##!a2V$tyW0-Rr$*~$A
z>&zRaoaoDs21w_-9i*;yO$QKgh@Mtm#Sc)FwM#c$2XD3P$`W@&a${4EeX!?ZglQ?J
z+CQ61;p&U*tbtycZKG?p5&;(-uKeU3LPkxH!fsLb`&`L>G2(nq)B62?wG{CfdIb7S
zH+C*<^9Td1jR-sQ>bugZRg+1>I6lexa_F}nC!calVp|23?Djc5J&h}_2N2(U{XRq-
z`(2%5%<xCl#(lNo_{zGrmKlb`4Ic6Y-`fQE^jjQ3gAew914)P0byq?DAkpv<<h7Km
zubY5-l0A7V1zZ_)bQ|ULqB8{vpKEQNyy9Sqlq?NNw`~zLiE3pEMxOI}_K2e0?!+8m
zBUN#Jq;hh$w$4Y>(;PDrY`OtDsOc)C#N^uy14^!4V`4ryi^u?%7*;8Rm<F${K1lvc
zxVm-fem~n5(31=7?W-ITb<Ikw{{V~4x8eA^9uMsY9>Y;<xOWJ{`ejE*Fh1?F2=6<8
z2J6g1^}H*E<A82E7XHKy0dzS0;B|x0FV{bjua<VeXgeat3}Clr-r3d`n3tzLTArTm
zF4U(?3CB3kAm9W=D+EO@@kfT?ynh}40H?L9VYrq6J=mZshEpU|bF&d}OPa>zvsl)G
zP04X_BF7QoIG?mws#WlqmeepoF_n#L19#QZU8{*E`U@>7{H>c|*bW||;3|eVz!a=T
z<w{-2hWlf}*JEU@CiL8EseaZ`YP;(Fp!;V0XW70WL3Jj&tC+^+X(V=3PG4En>5ln%
zc@8;cv6M%I(tZzw)jUast!+%LW3Mg%<|H^6+A`%mSLG{d)Ba4jj;S?Vz;*i70<!1i
zrk!qnQLl9faxh5uxC;LOYA_58#)^1=H*B%SIf)>XzKWhMmjcF1sJZls*uih=7c?C%
zw}<fS7<P$>rDtfZmlI~)*YRFir_)%`!f^@@z|p6Qrsn3k1+I&dc`VBKC$kXuQF}~j
z1Os3OiS81{n95eENSCg#Wld%KZZ0!6{ao(iwk3<i(QN|SM84Ls%-&jyx38=bg{1aY
zYu+Q|N7LlGBW!yNt2>8(3MdRI(X^SL0}(IABwp!r{oA{}k%Pnj(cs@-(W<zYNp6;F
z+js(I<4w-w0Sj9l!fLjcZ%9D1wYh1~^aDfHOYA3Q&fqPB?XPb$%nI2l@&z!1i(YBM
z2+QK!UY0ZM-wfci$uUjU97y#*sL{26f_a5+hv1!s;26w%hNLy3buf3_%rBnPT9oH$
z{{Sbj;#NiH9MIyuqP(Nq@T!1d-I#&}O1J)UC(bqZuQiLN=T;2@<*m1px3eLOnx`AM
znq<aI{OLCN@kPgDXua7fU%_#13acBQu!XOySn`o9hc3YC8gO16BOTZuQ@ragPje|8
zth%GSjoWj;NCZ7278=kjs=39+N{g7pvhJ~CE!&?^gmwUextNIwgEg0$C-k3Gym;bb
zG*HzcYZ=@R8r=tWP@5dYgQ7w?+^~p*COWd<Ht|A2K~AziN))GXp2*i-m3mm)dm?So
zFt^saBIqn2iLvXlZJfe6kDBdoDY~v2!_xaWcAs)%sq76+rOyt7!SZmkH9Qk8=3FLo
z88$v|atm*GmOQ@z6>%6c*V&Znz2C;;1+pjAU?tZVac<4N&k)N-xpd2%&_9K{`MAjF
zB<fJSlQDelM{3mO>oczWp4}p+s_>8gtak7EQETl3qk=uNtJwptY55N&&xJNPfN&Ar
zb>MCnT=@cG*4Jg8R?XLHhEu0Rq_HI3_cSxVjr9*#5UQa2N-;jyw7K{qb~St^8y^W;
zoCX>(8AssvOTJtAFV<$yLfOwJ{6gKx%>Epnl;CsCzy7ZWTP07hs~%?z_`%_vH)qaI
zn!j1Ue^?!z@_<w0Y2(GWfvAwwrQ8P_>I-!FM4IV+xv06Ih+mqI<vpf5gSLgV$_owI
zMQaMtCS}Om;Bs8*m17K)y;CfU+9ewtDf5b=(|D-i@T96@S_KD@_(%Rr<W9~iU8r&H
zji&J7xyv|j^*_cvHxqDUuvyG~`#0$CaO$&gV>4?<{j>84?#eMOR{&#5J}ssQnKbs4
z%MN_p^rM4olI83%`<Bw~hmWOCh~XcFs7R3@?$-}h!>B&-BX!dV=ixC7tidpE$c@LE
z`%Sg|meFx;!gh0yHfv96VwqWe+L_HvGEX>DK69x(Z9Ig%qaQQjTt)a#wYcmT6LDjy
zwW0xlM|2Vnq!J_<(mJnUjKU0O4Wo&|0O`H&@LcKSiLmdN_K}P6-E389*lLqCQk=S9
zNw6-FjsF0&99z-&bqo$1-w^h(x09@&f}_GP0od2#tN#Gg1o`-DYp!tGMKXF*0#tEb
zLA6=TNmZG4T)(2f{t>f-s>fqngr<MFYeDgk>a+1Q-RP$h;r;83cz(n#R*}LLikrX`
z$#o|@R`_Fa*qT(EH44cXdX*g%!n^+fquFX?w`r<Orm1@6z(3}S{+8;dDFxfnTS!ep
zsTFy7IdU+k3Y(t0+#o+xb!(jA_l7=elY<%R1*t0&lX``$WJUWCak+B$RcFYMZmR`Q
zGa|cmb(Vn&yg2Hu*iyanElOUXZK;SC&2;JvaIW6ck!tTmnr{z<zxJoInV!psvBd{6
z!S(hv*wgv8-5wWO`1j=**eqyqehRVgQZRfC!TNZsxXd7EW-^$T)E147;QZ6Qo8Wq$
z64`drcBL^Yi6}FYR5?;V%D(aUeGyFJIGkf@@mxO*X}N2jd;s5&Cf!u<_+HC~*T<0%
zLq~~PYNIiT0~2C)KL_SrrF^3`^K!2Kbhy!Nwk_eGgLVTCnyXq`#=*=dO%3M!y~>4V
zR~U+p3@!!6(LQ^HXpq3OZ#LClc3D12GazNTheBJmx<8X!&fvGl`yBXVQ;JZc{{TmQ
zK|7Ep2RH%=y@=CSG)2Z>YE&JrTXH+}*2A9Mibf`(O*u)#6Z5qO4tymDiKZq&(;G*C
z5gE(@2QC!5lvTxJFqN<lqlF~hq*#GxwbLdECJpXMU1?$J(!`3DdV_JuivjY{Y0OJm
zo0khFX>Ks#c#L?=V()E*s2}%%*{%h{L#9r;$OOqI^8)I=O3iTYL|EJsK?D!nX%|$Q
zm1<F@>Dq28xu#A30Nxk)#o8@naaG#Q@)71E>8h<zxZ&wJ!u{d#l>s+Y4NICo(eTBw
zqTr$3t4-c6Hugj?^%p{ia;H6c-0j2a-35k-P2Q<Ec9$Z73A@!JveR>l1SrFuxqT2J
zHF9OV_>L6W36jon^}%CwCo=Uhl_xF?^MFD`N>t5Rn~R5yeOM%Dm;pq)HV+pL?EEmb
zNLy6hW*f_izEA=)08?K`Ds;MTGV)g0U*O$hRz|Md_^%6MO8r`Iu~Vce#ZqE^U<mEG
z#UMFxv4M1rChPz#jyz8fal>(%?5Ydy$1#n&i@<d-V!#HR-RBU{Mb3+F6A!|I(-R<*
zU>ZO_9+zB~HH)wg&eOAOj#0_HP#-ysYtmv5wtlp%x#kjI)=X-<>h@^#orWhF-r__7
z{{U&E<^FeAzpF$u_?BRGIx*<?KZyST;x_TsXeyM`lXH`k)6*v8{KJIw#KgnI>#w=%
zkrxrKwk|dr16m?10I&m)0>CatTx68qsZ4R*o@1)k<dp_ORjay<-<-KVhVRpi-X&3j
zr3yNn)8GxS;%qz*C3ghF6)GM;c$uGo`Kt|6{d+$<QkHO>opaAQE%`wUE*m{@5boxm
zMl4~f3REdksxrs{r$(P1PCw>VsxZ0Eb8i}r$Co8s?OY;Fba-Ong~%^<&G8h`U7Hdf
zsDfeuPpOO>FJW3+Vx}^*kDDF?6MU@y0C~<SP|rz}+VL>Su(-}izdt2vscYWy40pHC
z<^KR>uqRGVcIk;%RUQ(zP&iGJ^?%?|pypA;v{h@OG;?(yz1nPgDvdQL62djq9S`rC
zitSF{r7(VT*7f2dtLZpX#lV{f4^;F&g!CkWbUP^)Hl(``er?=%a~~AsgV!vEJIFdN
zOTbNq&6Bx|0NPPyGQ~}n+YsjuTcH#uq0clYjm=rtlyL3cSmnXDH|92Hya{*m`H;Hy
zVUTeLFHdXq9%!LCrWb(u!5fAfXoBHVySqcozytVOa6E2$q+UQRr`?}<@j9P?T?4j>
zkCJ6J8-`__B4SqoH#z4Epr@-9DWm=o=C}mb`$5MsPu{q1_$kwG&1>;a=M3SrM%1VL
zPI>yRFx^gRw+8T4k5pk|Pepj4Pdhc{E%Rp_odBU8vn4q#WY&`V*FeWjkfFO&$6fhH
zOu{UURlf*I&d<(pXKZk*Q!7;3(n`0~sZ(b0c8QHSaN+l^iv7Dst94<NOOJZXWZ>+7
zhjv`8&e7xr%thgO&LV97J{YB1jea6BRWLaPrADqF0CQ4oo@iK)Z#dr=PIin-U2xhC
ziBUU2zjYEFE+2!!u(pFDbv8<_Hx-F?YR^6!FR8mk+Jw%{{DyWLvKAf|M-W!uoWw1&
z!>UHpa|R3aS+HZl;c!gipZbB0;}r&O!jnZg*ybz*Pgb~v>G2qW!~V|UaE!62d;b8#
zcaibGk{v6`Dr3sW$?qEMryWjRkEk=x*D&FcaTK$K#Vm{H4-9p81*7&$?F9}i!-iUQ
zS}D{U0xlimck3l8c=5-$U2H9XU57JzfiOIT0nA4w^0Kd1r^>A3RSfkSxJU?kW7yWr
z%pFS9_Gv!_=Njg^u%_V`_bZhWjYh6bwCuJaWF}oZnrZj!)%yG~QP_&OcGAPsVRld^
zE(8EC`fa#6D>Ug-u3b|&5_jJI_0g1%CyaYFD&1~kjXIRJ{O8Itl=7-cv_Bba%wmi9
zrw^v?#-&!0eTO?CK2Lz-)R^P~yxt#*fbh}>;sY@EwqxDmzWaGBc_!mY{2Bg>vAmr+
zr-&&Ge6LQY(rT?uyHlafHX%`^l64B%=Nk8$kC$=oBUk?b#D%;Cl|@RdFLgGsxaOrV
zr&Xb(wa$RxTGN@#n-F1oJUf|%;WW-U%?!+GhqxT0FwxQh4b5jrfgtJ@`0|YYOFUVJ
z{G2vFuq+o<mX^XYm6^KsAjDILr{?v1(Am^yrSwreMG>os`$6_&?O*=@i@w%i$T+M!
znMgIjjM;${<lVM!X&G|>;0!uIqemOU_|6ghCj|U~u3!y`Bmm;`BfIMypv;f~0O-CA
zZNc9rPbWF*mKE7llHwhtH|AcemlUkA{XU<mo4%ZT53}F?D*H=-W;=_2!qan};LXS{
zb>e5u#D5BoJ^T-YJF>yIwy6S1&|9C-X}Mo2wu@H9c&26AmLsU5Fntn?dA5c=hlHhv
z!>}@lnB3~To+7P$L(A(QA!VZG8??w<=#3VFbyO-5vW-H?1-^0EO0I5E6v2TK+?m9s
zLI~9|c6R<d(k2fyp#WyhQro$~BXpt{ie&RH=nJGkP80bWbMwop6U_#HoH&U$x*#UA
zBGkOahUQ?NA%Qevl5OF-T5V95quqo=>W>B8TXEYqFPEb1Cl=tElk~Bw;fJ1{Ial~9
zvja2pswLhRsQsU#9#4hvt{qaQI|sxmioK0!1eZu@Xe|;5w%~tNxs1l+7%=tl?=0;p
zf<z5Lxwtx=Rfoyr8Dq>@Gyecfag)!#SQ>+Jfi=br%E~cok7Pcy%dL<5m!pRl!0_xF
zNx@Rd+2jr~<KC$5=<gGw8s}T_Hdky8Cly`}CHb>vfdE;8=EGdX0XDO@T^r<uSDD8?
zPJUOI+ZW{JdkfQW+mt!RZ`74;WG9Wzqw}cdm?xaO#o~;vKY_w<??V`D!!w>AE}OTI
zuou0yAza~joJKOGZqh6oL<q5sb=uxCzgZXxhfu;BHS%A>?6!$BNpg`zWZit87j>kZ
zm$`gd9n0d+qA}Rx==2C4%YGyDD$}9PsPnl$Kj^DFBHHE?+cbq+9^iVAy@y$@RjE3S
z-TdO8I`o^Ow+eXMRr_NM2Q~8X69=7*!Se)pA>BMY#;fFa@}y!aJv(T0A)Itqv*2c5
zaiRe&GAg}Lclf$SC5OYDRw|oXNSBackVFO@Mv@$)DhgCyYBymLBpqkmLivBPj6Wrh
zsuii~1~X-u#%RvG%z<3Yu{s9bq^bAs_u|JAVBwgGEOj@y7lyc#ocDnuC!=JE+y&DL
z#^Y(c)2Icc2!UcR9(ME9T^`tHm5vC0Y5H-`&0!ImB4%caI*mNngw;d7%(|gd;vZtM
zbeh-Yrt&V=xy>QM#6fh}%II)Bdk*<vFB4(_GACUM`w!$nOR8fmF+$j0JExO#Rp~dT
zDKm8g^R3qyKQ(u<WwJs0G2nRJDh%S!;=CF8$*9^vt&%dit$-X)<2IA4hl1IqR}q82
zhhkWM=w^5#`Ad#t9NGqN07mP1;=d$^wiD#(<u{0Fl=2FuF{#SaDXeMv>3$D7!hgOc
zPqP(U$1&VPh;W=|6JcX0I-^UTKn2u*bx8&cv=#%QX^?NCVZl|$RlxAqX3obv$L5B!
z#6-9oKyX}N<ptz1BSd2bP3?mcF-6&)9iU3dwo@)!H&>b_^(HL=_t!3;40;Y5_NNr;
z)$sf-gR9l8QO3i%#^}i>pq@*eOdk%%GMk@R22yNV_meV208CmWk+5GsvCTf2PNKff
z$gny|U}<ZI<?7H{b0h1DbocWsqt@XoSE*I5;?=Xg-}KpVIvwrgs8?edNRAy#BU|me
zam;PzZoYEkO64+EarYUh9Ng{PQx2aDdh9!1SR@AY-Fdq1Yg}OFKLw>~xSnokZE(>V
z^#`UjTE;f;s6FpGs+nRI+AluBFZZH(XG01h#A=^Q%;4&(>C`V_{{YXm(lN=$6-3KB
z-4xm|@3XtBbl`s{+)a-iZL9%Ok*S|oVP<u<=RT&~B){y&ekY6J@Io=obuY|Xa`})5
z1I7ZoM~2qK!NR$<IsyZ{V38~9{{WG%lf7SPl+0)3akQ!g#XlI-gXL@@Rfz`}DJttA
zL=E}OxMo_+bNVskxYrWH;IVjLwLBuLYJht;hB!Mr9zcs5xvz*Gw3yv&;b>NG46nnu
zjWW>4e39yrVrRg~9D#!{w%3#=kcQUin+Wn_*ft8;<lM$NsnjYs>c1w;u^X~NPlUX+
z>)}9bPpT;3KGAUV$GC?Y;#GSlifpGxKXuMvFU;WBo4v3PORn{>Z>hs6;4pwVQ>ou@
zBQAIZ-eAUH0NyMY%5Ao6cV*iR#aNZReT|x|W8$5dd*mEBI8RE`$eiJA$G&SmTY~JP
z(r~^f#JJ}T!~XzhVa*0Pjf$08GZ06t+G0ukvU}R-30c^@76XRXd$E-}S4r+NKph<8
zq4<r$;(GQv#2A9K0WE7Z-hh^sBNm0fQ$C_d^pvo4nMNKrnK7K6WauY-_R`A|wkoUN
z_i#E1%i+w0B-D(_J4s?l-g1~n_}q_ZEm=10w*35I?e6jVW0oSF8h8r9tqdWi2ml9f
z2a}>=;&lgQ$&Y)n47oQFCga1H{ley#=B5SBS$UFKYrT6&huaz;Q^aA9ZqPF}#0zOM
zG=fgV0{{!ktnAdPjD~WX-toxmZlM;ZHFXkEw_Q@QaqArMkQ;6P0D{D*Ili5lt;?E_
zn1Cg&zMUN)@*yg%BM$`5SD5f3-=XU&w>0nK;bG=O8WYV`RpXg|J46-<LPn(6@f!vD
zFint#b58zkQWV+(6H}PAIJ%&nkfGn3-`tSrVuc*D8jyq`U}ks+i<CJ7L?k&lIq^hw
zLTKdY3BZJeo3$;wC^nFVlbalmi-}=RcoZ+H)G16Y{{Yx`WM7IgH&(7~OA^R@k6n4v
zrB2lTNT$0iNUfN=B;(AYwk|Z_<gGL5cJCEL)o0oaPEuRhFcRj*VBXQe#DgKt2fSKF
zgv3JIaBJ{Qm^p}mOagLDTHxBoY<yP+&td!piKeFIFpeQQJ}GU|W!yK7AJG7KO44X1
ztVXjE-E8A>{{YsMdQn>$hXL5O3W7a)K$~kOoR<!A!r{CCvr8z|KXh`R!r}+{tUQPX
zYTQYLv7Q*LwTE#w8JLx>JM$8gZE2L<-6RXr@Rof|+@ADUr&|w*rQ6Y{90sHvH@_=h
z2nV9(=2fQt&FvuakKFjFY-gb==imgZs8V#we9W}L_+Oncz4QGf%k0KpAUdxMtHd5I
z_c6nx1Au7vOOH}j+D>aRtim+<p~J<*OOL!qlGjr8H%d~6INoSQPm?|p=^<zb_(p~t
zHm(V*aGgOLYE@dZPY~h`g<09CQ}HQ_H93Z*Ys>!L=LD?be@XeJ_s4hV2gL{8;XPS<
zd1~hPZsE|}Z7EADepB#2q4}r^q^(wyQGS<Ir#RI`t94m)c!uth_mH`Y@?Q*hYnfGo
zr86Mz7cvW6TI)iODTn_6HmL*@q%Dt45tzgkbGJ5{HYG=g)N51tq;sUx=?{US4IT||
z^j#iDWDhV;_=>na=g#SoUYkYUmqY9Ed<y97>Oh{+lEchz;xs=#tB`qu$3waw-IEfO
zK095#pAoh_9})f|a(x))UzNErZ2%iitv)lSfwxsK+{ZNo<)QLH(hGie=KW#+0G=p_
zZCo@O9rpem$MBuU<vvA7mR2NH?w)%S`G&`+@gW!`7&~nxTfJE9%9QI<&q&YjeaC^i
z=H|Yevb;y{Wbd=LE+S5zQ3Hq**<st-5v*uDPJ`yADLXD|ry%REkktGF5ziMxL+0QV
zO>Xur*}o{WN0zShw?uY=ov`CCUu*OpSL&QJ`5&>*dOPSLa#OO?V9PFEei(Cj2wR73
zbpn{$1v*X8w4^X@7Le1r>p#`QFkkqw<-W~H977W`+F~lZ%e8w)!J%#Ol0&f6Pq9jv
z{{Zc@{Z_PP&k;Eum=YEupQO)=4}oveE*Oc?F><;-nM5pEMt)s^&T~re_oA7t*;EC^
zbcq&PLCR~azT-0uTIe4XSTz9>)PqXu-JRhxq~Le)BaD3enoV8ejOx`*O8uBRD-DJ=
zR)xnniT?nV9g=9*QNFVGiY!v519LeBq0yyl$$8h!!)O{(zr!8LfS?>fm9uaKCHn3r
zqct~~G7gK6x$6MMTtz`g4C3AFtYN{cU9EqK@LhZ56|^6e_HVN0oqbeFw~3x-*0Hu}
zGN0Tiz~MN4%Zy_j29)sN_?KcHf7BMo5wF_eFyr%8_W?<x<X0+SZL3b;Y*$aK$jlN5
za5^`2wc;4u#bMcu!#b14UvkaE*M1ui+O0ZRsSU7I8lI`Sh-tZYX|oM`eQUa+h5rDv
z-GhGc(f%MU<)XP%wBm*)p7&K6eRfuP>z|cUn7ZN(P9pe!gj;@U4~CloVl)zW0(Rd0
zV5-e6c$c^L(RRm@@8R8R7gHqmBeZI*FBrIAi<GOA4Th)1pYp~B9IJrC;A&RmbQxKI
zUUHKj4j&`nWo>ZAVrya>ONZ5S3AXX)w0MYLxvwCqRyD);xGCFj#+3T*(Dkg7>=aoy
zXXrj>=N)yNjhnZpoZh~u$I8CX;&==H0QDOS=-4tE=N8iihNHVdn=MfHIlYr<w>E6K
z%e2c9=IQaaVw=DZX7@Hfyg7|FJV=H%oRIfVdoT?GCP`gw<&#IRVvU*NnILLuj5jMQ
zsWCFsauafI36-wuwK<Sl(C^OvS4P+Vp;Y6znv|*x4Q|5-X}`5?P$?IZT184tXsp_-
zIa^8L#gO^c(lUO{zYB-Qv^mT!s~`iMI6$^XHbKpBTa?KVU`bz#&nOBW!Yl4nSlMM)
z4Ako&ZUdi#wHGLb^))k_Z~p+hJfpAKY)9G}JU*NF!!Z=WE+e&It3CHjxv=K-4s7*a
zkBC#@^~v;6OuKN`hMg^Ihj~YW`1`bc4_K3^P4=$PCWc;aVXmKroH6wrCNh)hf+KA|
zs`0p-HC9w5(mY4$WzV%a8M$|ePr5XlQLYjed+C<y6<x;-l`bV*qo{NXt_7qh`I-B^
zsGN<8=a-s}?K+)3(B!C)77}Fb_`?QJ010C5Xg+h49YRnh6oWHFJtna!Nf6u<Q8V62
z3sl5JlxcS#GqskKtm67SQO>PuIA&+WJTx0$`YAG<TF%Q!vbtvwQeVc~8QE<=2b#>|
zY8ITV^&PTsoM0zS&OR?Uma}2fTeWzj&tb9HTqhmJI}-5Pb%=9cu`Or>2$F1Fu((>)
zu+?jD+$f|a)Wi{~5>CEwR?gTQ{Jk4*n4W=8a_LGY6q}_}&;HY(QkOmOb=%M2=*KQ5
z<6rFUJXwjYRh1pnHfSHp+^z8orVKWp6@3=}0MlV;{#>$r&SU1cN08nvHhh{qv*>dZ
zjkhN;hpFK0-A?JZI_$Wo`{GpiJsH1iTJKH4IlkMTeA?J2%mHqMtDFe1;qeoza{Gz>
zWq!{#g;2zl?T1KX3>St{dlAv3@a<|sp7ifvYHnq&%HXn2SVJ#&h^#ESt&cBE{{X{b
z+VdZGN=OqO#_b|+)GZPSH3$>lsRDKH9oH&(B~@li24jkRuFl)0;TAD2C;2vMiiESF
zQh4z_Z(MZQ8=11uTaT|>hoqyq%d1^72bXoRdGQgnav?wmWtRxo=6^}^=Dt_|04$|5
zv&~@aMJvNGQ^PZK)8i|g>{WeFqm{4a`cZp_ww)YJD&%_4bDVq&T0Q`HmCCkyl<HID
z8V8thKQ-}BV)chccWBJXO%gEH52yn~9_E)9+Ar-!?-F3igEa=V+E0lNw~&G4s^Hv1
zs)iVb5Im#js`@qL#kKvo)8weLI42d&4MtLaUWYSuz=$-<oKmK?^(W2``>>BA#_${o
z^&jcf+xL?3Voqp>0W%U|&@CjdLxII(`!LoV*+>K%h}uT>w~$|`-y|$qF}68b3|X{Y
zol;K(%DU{(;<>Tx9iGB(PSf3}R*l>@N6IiX_!TZOQw5HzP1KoH-Knct^czl?$aAu#
zHv;N|APi<&jxGXP1h`z{#;y|%KeOvpZAz^Hw3+vrr#5M0fy&u*Oo7T**}3{P2NOxk
z5H!Ks;7&5`fIFl`#i*BekS{-aKS5)7SK0pnwR|;#i)k2hYO&^%Di#`mIn@JPAU>rs
zMrR2bgP@SUZ%(c&1yRCW*#)EmKp<KP8i01t#4m=vS=$Ge!MK*Yw2H<Qn#7w(n0e`$
zd76Vi>bgT^b)YT$<id~HNci3=jyHhu+W!D&xQ;U+bF<WYTKa8ydyC<Zc$YTum_uwp
zm@DY~CmBnL;Mj2V*v8RlYgsv-NB|cY0iY5lF5*q&Rryr%ti^jt#Z_vHn8c?Av6wA6
zN9FrGw6D<t;RyPF1H*9KEf$O!VUcnUXF5!oJO24Bt|?m?TBC)0*4Wq&4qr9%cM(gx
z>+Y7iq<ZQ=2HknOj_<8oP?2zyub5k++ia^6O0Ky{GA=2F{TS*6lq7ONYLsGs&NKTk
z5h#_q_F+V<&th6Eh8~iA`oqtTgA&r@>!<J$iWtXHrG|U!kN_8w0P`hV!wRp&v6KNM
zn?RH9U#$=1^4iuS@?)I96ptzmy_Dqi^!&WN17_MRR_dhdh5*&(?qN^kR9uAf;g6AV
zU+ny68E+cl95B?Ugd(k1<!fow?B)4d*S9OAi&V->P1*yY!`HfoJ={3(+-+dj)frk`
z_Xji{c_q)Z_gpdwxwJ2*{{YHg<qxr(S+UxsB5M>^nZW-5E=}QVM_Iv>3=7g#1`CE+
z_T1xrvf{SK2Rle&*A(KIf3@6q4Oi^;9fkDT=Mh+C6+j)=02*@!gEXG%fHq$=<_gW>
zcq~r_p;E3qhUaSwY)K7%rLG6K?^U%srO;b=Uybd*c|YN*rBbOkvE8&&D$<RUjvXd{
zQjt@X_%{Clnm@ASdaeb;aX3sRA;$18qFmPf<6@&7XcFOefdB!PAZcQ_<FHimFLh2G
ziDgJ8SWZ_D2>>2u@^xHCO4;5dV-p=Zvq5o}+WFmUqt}d18nqtMi^Hc#UVf!2)SZ^A
z3t>uPYIKGTpOL9mrS&lz<k(ZGWMn@lwD%8ebH$l9zG=Sy0G)S&!Fz>6Qx$}*Nru^7
zCz%Z=D@n9S?>uShw9JtPWv7T^j<O@A;!n_Ts^r=EJ~4JdwMOl<uRoYE>di~<`%!ZX
zS2hZpF$QI=^97<UsXBr=-qIH;tRoQjQrXAK5A2U}iJ2DuPJO^}O2<Lc;pxB8_eUxp
zca5U<h|JeFS<?Pi+VSvHs&nDq(q;=!zTY(isjJ}NjZDyBl}jnrBh8QbiWX>0G!Ubv
zYK*i$IhjBSC?YJXvvp7?kk(Qol7l*j33GyPAt)(>lTGzput`ZUaqGB(2|A$-9&^PG
zP`s}hUG0T1ge5KgP~^qXim&dv?r?#D5QX3{&H{z>F0SxZf0HIzvI1ABJ8Y*t)5E7X
zh9(1_(rqKl;bxcHdf@wXs(+TNPt>hiknRk_$g||&xt_TH0QzLZ6ZqGj)u@t};z0bS
zBBwXNHAMBCWO7N5NZ9vtv9(YC0L!ZkDfJ$EpOxB{9*0&L*Sl*fZj4PcUB0t@Va#cm
zb2~fVx-QDG{wl$EPZUl}R5h>#J;r-227SKKAv0Dpxtm14o2cZfWiCw?S8?x-INUxT
z7m24ugDc-Tp4&UK0OpvC(qW)D5N~TKVd_<*MfFX?bO-SrB6N*y8zv+h9>wpyRHRKz
z(D7ZaymWeotyY&(a?<l{JU@)f5$D1)K<a<%103-~iPOMUp;!vwx7KVm;vmbKPE!$b
z!Vl$Bq3F-WQ>emNXG0`!q1gEQ){B#no|D5ketG#v5}Hp;(^;A}i$(XEeEiZENfm@E
zmp+)Y!|FBoT!86#-Ut_tQzCh;c(##<E{%V1lW@1`NV;IE&{e*Pis{88FNbk-k<4MH
z_+^e<#Am4Yc$y>O%e<R&fq$buD6dkLM!eU;9HiWX>bu90{KEX?S#)c`9i3r3j^p}t
zTVn$GGVNz;n1Cqz9-#cUN6|UAtFkV)DNm$J2hKe-<jJE{jXcu+*lnk%Zi<SRCu8um
zm`)`2WNOk5Y1=3sxH}PUaGzmRxOQVS=tu6}Yeq=PgXH|fGbEx;IDeA$t^NxfX<_SB
zV0#9#cIV^%+ba*M#@XH*{`}MMl9*=g8OH|iRqH>q6ZC9g^<5v$A0ngMVa~1hfPQJ?
ze<(CUV^DX8#8YL%3}N_({m=g6zED!_aKDHAqxDiQ{Gp)x;O0{f9o}|k-?XXxf3+M7
zaa0JSMxi8`${t<t%MJ0e>ojV&JK6_Q0Q%i7W^+s!?An9og{NI|Z-v*qd2sNddJBKj
zh&v68tyuQeJ2LHyjr_pBzEc%wQ>HsMW{!g1TX^V-{LI6HuEW`|KDSeGYM8}SccSLV
zMuN~r<-d1+zR=xrd7$CjWz07sZR6gz$sxJPp)RrgB98fZkF#07;Roc<xn{zD+dlaH
zREsln1?v9PDB^(Lvp>Lp`2=RcM0Rff01*8YlYOpd%!RysFy|dlzUqv%{X#Uq<3RLh
zH~`0A^9}z1m2U03H#Qq;+G|a5XjGb~xAbMC<)!}syBJ?)JG>ZFcemdE02`j~{Fa{@
z?{OzL@W}Tyt@<r<FC|xg(U*~Ne$0A2;$l`tQ?OQQW*fvaJ)zFYmX90Px&fvz7qmsu
zmD+{ceoa)W-Daoq9W(*VQ7LrviyEKLLBE`5^fQdyVC>u4I>5j^CqlZ@hj&)|@`dt1
z#M0Q8XL#2DQqE%p<EeJ0=;bfP=Zvg2ej;0<z8e_c%B4ES4LO<VznWwa@WFY}b>7=N
znku_3!Wd%=%u`5Zs2EhS9an?$ubzyve-gD;DdDReMORdxUC=xfs>?(WzDH9w6_4?(
z?9AmlnBg#PlamuRa0z=OZPlpJY&pw_mlZ2NvWYAp8Z+}E<lyx*&lldY$W^IV)v}#}
zu1MAXMLEZsBty2*21oeBDzMasld88smJH=V?P9Z0o4|Vtl;XIlHM)bJ`$cMF{{Z2n
zaN+E0c$X4*+NCD{00FfyO3#N5!}EDv@oqkef@kEY@9|Z4(W^iS&S&#$`Z20>RDmmH
z(kG_t6zN$f8dWeM9wF-S3}j<k8&aD^vpjqOD@P9FQke%}wgwu&(k;{_-^A0FNe>p2
zd`>z3SgiW^#%A<<ropmvLa@pd00$CCxjRqq<hz##c|cXL?Fw%VY>IKJ4JGMfVs8~2
zH%{B`FU}vwlzGfo3aN>!=^RC)5JklqJ)Gax&<_nvK?`Pw5ELo)7$3O?#-+1F?Q73H
zI<LE4Lf%j~56S-kX&zF0Ql-&B4uI`J`o<utJK8FBsf&xuqV!7!<V=Tp-_;%yjyLUY
zAB}LFX9&awJ+J$kZD{Q2w983?4t&jXgWYIsfEre<IvhKR@N88Iu5bbz_ZJy+h2(*5
z@3IZ0cAaDSQ1N!q@Z6>VsLspD*Ku6EVo}4=2N|SQWhOg2a{mC;%xRHv^FuF0W$_x{
z`h~=+aUL7RF~<`@uftWS%%O$sd^m<U32RxAx#@yNYv^w9d?W1j8n|p_78bj_87`1W
zpIRfme(4b*vP6=<DIQcCR>QT7ZCRDZ`zeuPx-Jl$^%7sOJSXoXd>O~oKFZ*0VK@a3
zL&d%A?-19V<JWPqJ@1c|mNyr~;opsWaTFTZi?%_~3x&oL*t|qE=(wH!%Af4ho^sgr
z?WAY#X^#<QhU%MBdGIQWJ&7H8NM;yC&%`m^tcWcui=GL(sLVdh((twVF=3_65LX8V
zxmTu|AS+yHP`&o!pmOHsg&r(Sn%cuex%#6M4v$%Lb70v^9vrF4Fb_>hnl!ydZ$ufv
zIaEtbN{UeioQZ?UMj<I*mu*B%@k3u&q1@<|^isgv=0pW7SEspjab<dSX^N5?GpQRG
z+^&wK=D#Svky}G$cxu0iDm8fu9v6Chv`hS$^t&>;pFgNKXJU(1iYndR!d8}1m&NfE
zs8FO_@FCf*LG?wX9Yyrhot4}L8nr4sF{RQYL%IHowy-V@sSYp3RQ9%P64c4jbFxX(
zD76A|k}}5IUp{}%Rn5dD-mPjkq&uU^SH)o+LYb!PgWhwV)0#mhTkOV`fCK^!irb}!
zuE)Kk6FLCWA`EIZ=i_m#ErtGtW@z=yZM1p@0jU-{oROLHy^>|n6DEg*Ld!>?5;?z$
zKDH8{52;<5SVRGr+0snjB~7>-mEVe|Nz&<?J9yZRi&E_S3+<z3ct>a$GX?A-vQ(@`
zCM?wZbpmc>(GhLH8!i;n=kIS3g~T|8J}=pvZFU~alH9N^L>Yo+-4w96N|+4o!_s)p
z05S>C$cqSHvHTaeJ(^bho#?cTS+!bQIP36QSZAc)sy5rM&30zbOpv&<1L4cuj5oTC
znjQ=cx{UyUul6{A;zcqqI&)m@<g>SJk3F&5o!WeZGp^!HRdaK5z%RB%i7F+(_x-q;
zep3}(0U@W9OF{WAVwFcQc5Nmt<|HqZU&^tJaaPpvhR-DSSBB>+7*<kBs?)PHcAlNk
z*LjZSgUwxLdH3HIuvm&zYMRw6G(j0!;vhxOj$?8>!IjFs9<gX+X&ZwiN$~d%_FN6*
z?XXN`QLR^4s_fpUCrX*7yH9oMyo}7VX4&<J6KTJ6cj8=d6G%F%9hBz*mohZmjrT?q
zhrH(dw>Cr)>safsmG$RfSSw;aPE@M3W}AfHt79I^t=Xms&5$NQh)U8PeyYcg;+T##
z_K}f`s3_BEIzVv;SiI;qg}aBPg5i=W0{1#RBrfRKR@rAZiM6Z&{+F>(&s{#j({N=f
zUdnwPks;%|z%7nj`K{eC<v6Dg!f?DZ@mL%O-LSO1&v|xXYYk02!ULScAa{p&Uh7Rk
zgsV=CQBCz2=*r&o!MQrP@|XblFRb`y%hv!-;|;9hdSv+y%&5)NsbFuu*vmIkdT+CM
zc!h2s5ppKu(^30h$1prvw;$nHcQ&3XpWNYIdyUE|4%p;xcm}b>pb}it7ziG(n5p)0
z#|4JLRC`TA=a;-*&1~h8bp(J!4)M7<ua+KPzEaym+3kI-eSq<F^EiI6^TRbL6GJ;8
zCs1qWm?oxk(?i@qt=re^io6nzAMH**6I&lmviBN=guce1CEc5gk90pV4LP!h0V`-y
zJQo2<o0tQsjG^JJ1h-a_Km?fXnI`S(7saimd2Z4%$iFAF&7R0i3!m*0ereCqb1?MB
zxWRp!sl%Kf!tpO;Y;Ha6gXF2#;`H$oOBX~tsQ#gFt~6s}n84nc!_x~dHTj1;+<JUF
zxW*6$;(&Kmhh4lg6E36OT<>4N8U*8#Q&-Ezy*yDPycT6lIc(%IwUd};qpD5zQ&RmX
zwkxtEl~~iMKB|_I!6G+JRo+pT4xw-Sd@;@Vmk|^~lPH1Ss9|jSPTH$D3ZF4V*~F4b
zYKEKD8Fv2w8gxFFR8gscrHrU6W6x<94KtK^Xsx##>?;i{5DcAMHTUcBT@U0*v^Qit
zp>USaaWvbGQ*hzTVu`|dh_zgnWR1SB=_4!c4iwe<J6DHc8Ud~aCd6AYgYcDigVpUg
zqibS`A(@=(@)p0E>ph2gR?XF`lDI!<wFPr$`%I?P8Lw+O{NAPV-0`tCp70u{)frgM
z3d&SYBZjvt@J?xGR(327Vf?ND$0I7$9wgeen@bmV*AAb`bGiuffJZ3g7d`nwc|KQZ
z(^#TzQfz{!M5qu<c6J=P&bFRUVGn$5@XTPTfy3peA?$67fV>MG)+REU!{cf(?}#Tr
zdF}eC{{SXT3sq|gV@!WknswJ{Qig2GJ0);5;J0v+A(f(A^AP3IR`)@SsO;jE8eyRN
z$2QtQAY4t0hkH>cRO!FD-+4MfgYcVzd1>IeA0Er>^}J6?{3!Wl+S$6@Msm)^(~dmj
zx7=4;e@ZUXsl?75&7)0?1VAK`*p0nW9Hs{)Sd1xtXSkm${{CC8YN+8?o~+fWtY&&M
z69VfxTnvynTv5{tR}#(}LZfMqRiGT#J3;7q0stT#QC*Y{CZH}iCPZnx_>Q9KwL=5A
z#H{wbkfxYk>0BZLB78m+yPQguYKEl(@JD_Iht>{}8H^I5_O~TQf)93m#B^GUuH>AT
z2sEYrnAj5Fs1O3z)3b99k$JVZ-j1LMx>8gFg6S>B8AH)V`lb%BM7I-}M5QGf<76of
zE2ak7<<$a6X=Yw<mgRv61Ry9}P$37+eejTy>x1or$b+hGZjav$2#EKGlqpUQtLWwl
zpOUP9Cq~^cqvqptZEXVs)moi?QSQMjYgI{Oz0n12l05hXx0R)Ad3uiYf3{7;2D#G8
zgDXa78EXxjzL(^+TESa{WXu}@w<ZOgMz>sLg|sk+BBEk?F4QHOugXk|(%&b@O-*t1
z+2;g6(79#FP@`6^CaxxvsXGTyf-jhc4<7DwZFHL|)acWtP(8SYNaZKU`l9-4b<ntv
z2W)Zba|4PoJpCgN$kd*fnUkVW*`Ab>sypXgXzrXtVMKQ}oM~uUCRs*p98Vn1>a^*x
z?syPi+DT(<(gRrp6C|4%Hvw9!O92h1nZPvO3vA=4B$Ee~h=_$RhBj^FXE-gpH|18;
zl=zjw^2W}w4ArIK-zN2tt5*9_QI2D%5(pw18ru=g>~Z%;l1yg{V}WTJ`%Wj|BSjJ9
zu}r4FOWGwmv*m+_@O-m%!(w9%<Jw(PsYb<(n?U8})V(?RL(=8vn-b$23M&Co^eMiE
z8P$jYhcM-HVnv(_Tqjr#i(D?13tL&ud%&J#L%fZ?yo%ZtPa?gA)G4@vu}Sjw*_)H{
za+8%Km7`+1)4j6ZJto+8f-g&@yT32t>E{+dw7BIM55vDuJIrW2wa?xp_%0@f4eo&L
zZ)XViZ60ewVU3RL=8b|d?%1(Kb=kZ+-FcdP{U&msKP5`G!FNH))&T*dJkd1T<MH3q
zcT>c8ty+%4(xxUXK<{$13yr#w6S7{9gAn(=I#1LH@LR%PY1oHrIM#=OFm+a%-T{-9
znPrWiVZ1LPC4{9ZBtYcZK9dtDZifJqug|{WqO=a{0i$tTw!IoAwkni3ne80vou*-+
zY&413F`xr+1Qh<^&+Mo^N8tcRS(D3swuQmeDLbSZEp$!>{{XXp_oLI$4iuxH9|`)b
z>V#+a{+~rHGb!0w35mJZ+hf8;JzoUze&~&A^svGnZ^&w$I&Z7$`}S4e!YKxtKaw2>
zXD+ln=;|4JJR$aX`_WDH>(&Q5vJOMmJ||^iwAo1aMsBB1a%vTgS&?RTqR{&HMCn$e
zNuiYCyzG7_bqBHaJI|Buw}&NVr?wL;H8vAz=Lgh_V=Zb>=pox|e3t_>-`%t)!z2wB
zq_o-t{{TbHm-#<*bA1w7kax%7CgYY@Y(~?kH8#{4^B(EuH1irBuo1Ln$i=cPH%r1p
z9NSs+iEU$;4dcAg`Nre(w%qxtx|peYU<ufIr}$3C@}DEJnU<547X*(r^|)fyoHI1c
z>By?)K_J=|(!fO_H;S;uCAmRTnW@ULb(~*1F;g7w!-irpAA+0DxusdcH?+r5rgwXO
zVPC8GwpORK9imnym|Z!%O;&tJ+y4O8-2VWCWAG*lTq2zyXDeC$hW%w<;%+rD@2MKO
z{{ZVNXyJ;s5~}ujS&JL+iFC$|PSKs<RRAQy-*|UnI7b>eFY-l6mnYr)n-k5n@7!_2
z(<;)IyKZ2@ZewE0ul8BwDYXn4itwg2!q};4HK{4KXJlW*)LFS%LUan@9cDn8%jrdW
zyiW;Mlc`n!&5RMhLv({4!U)@*{XEy}lgW?BOAX>&D&&qPp-DNiS&FQd(ePWe>YG!;
zg)43k`o@-oA2`^|akQiFbjEHU4O@sXxVV7k#&HA@q!2D02jM!HU0s!gPqJ;`Z{VZ+
zsl1LbXZ<{P8d7Q%n$|1CVNj-FOff5Sj#1<#LS9Z;#|tLqfOw6<D|fjPt4_8Ct6NS1
z??Ack$cGcpMI6m{dGfz8URe8K*~ZRh>KGc4Ga|JKTble8!H;WG>3-w0JujcEjNKRT
zn0mEHtozTj?feHX70J%_S|T|wikQC};v7dd<=`#K7fiIj2nYG2VYJyyeyIzp({lAW
zxnmZs7X5b*h9KbF+|{r^Z-SoM6$-ZWmdRwP%%gzD3#aN?>pd|!G5z$4cn1Mf9Z}xa
zCa3vXN}v7<zAdjU!n6%2v*PEw{{SVfiwxCdUACh54-NKeT^Ebm;oQ2c))DI|%2eo5
z>Qwn{*0|fH)5YP8+(<PsR74XQhy%htrCp%rvYW{<qS$O|)rIMf6c8^N-5lxCIh^`}
zIs?kuu8w(gL28bfoU6<|*>tbCgl=rQu8U2lMW7h+TuoP#E9rJD#nn6~g)zPWka0U^
zE7~N%&wer>dLJqAe+cqJ2S**nnZ#<geIA#0#I*c+?7bJHI|;&04Tt;ZiSUu|T89#C
zQxLnh?WE|`P$|q+o|u+m^YbRu<im}cCxv(MeK>^-9uA%m98SHRHj`;Nsfh!g+gbsI
z1B$KbPVB5RA1tQF=9qaw(d9OdQdsU;=7s5oni-KY>MJ_p-vD6Ym<C`8klfjK)z01q
z=i<3>GwF}?mfB){+WD?Fo5ECAJku@$&E0Pc*2k!8$bwcAgcEfF<_mGY0CBzkvBbl~
z*+v|qW+=~6S68*RhM&?$K;ZopNFYd1@6SxDZEMeh+0q)F5ikHS0)(Dw)lS9INYFJp
z7Mgxhv$#qk>TV`lQ_M6e%=J!zQ)#xJogFgzPyqJc5K9QTRC$^O5%*mtE21s0fPjiR
zW~s?TD>gqh_4oNRX>zpfo$Xp<&AWall^tRc;kKSUKdwG~_R;0o{38iArtSX#I`zM1
z+R=r?81~!#mGsfKU2W(4JGPqohEd}Cx|^>|CqPqK1|C;mHeNy0A=J}u+xUCoNC!n2
zH@cQ+9VZ^ts5ql{tp;vNSx)76TLf|MKDQOTS3?dkqX7wIwNK~Q*5ikq%2CM)$vbZI
z%MUrwrUVeRKE*bFiLho4o5Q$-8yWuq!YcgvHV)n`w=-0jm}@_6gr8h>uU?f(!Sw$C
z0|%IB2g|POh8~;g(k*uGaq=!CE#oChm2RNf6CzINml=H_iq5F_2$)s?%5JQOQnlM!
zf2YSB!zSfOzSd2D2;7!&zQ?)CI^tr3Uw5eXjvlJoyBNeXG+Mo17R(x}vNDh85q@yS
zDDgT?txlc^gV!QQfnBh~cd9>^@Au@KRdHjo`!tuKk*8{_)sCw*aE+uwHimGi;hO#<
z+XmW`Lr2y=UTAelr6*TU@2Bi&wWi;(T#nPWRY<{=!e5kXeWj7?<L7%r4S_cDdu)#U
zI)iu@7C6?34LbrxS4@K%35e=BrCsdDVBLpzv?I(f&jvno<XP&ho8kwi_K~_yYK(nv
z1y>J<YeggL34#coaUd@Xh!rU~wfhyLrYkbczGmF8A^1(N;ToZWA!oVHl{+Jhn`OWH
z%QN*(5vI~~K&U{1RGH&;bjXT#hx1|lVm*mE1uoP9bj6HmtR&wL5q#Ta)AdBfCopDQ
zolv+!ES&WoaK9XHE4|wxknk5i>0L*&O`ov-6UCEsN`Nf&L5{>4Z*l3yb|uE9Dw7`H
zROYzdT%cb^4^hnage!!vlQz(2X!|IZ!@}G1lFwAzITL%cOqh1ycXj^&OT!%fB8abx
z&uq<Q^&X>v`FvL=W)!GS#oy}(>b~ZAKWwIBXt*_uO|*<K^B9i~{{W~ltfbtT>daxu
zhpm>Y<U__%j^w56;*5$-r%sTghjg|YK?X9I0D&IK)1KA@mERLn(pqY}P*0ftnuWi1
zq5Pb)dnEF4XJiz~6+A(XvBaGhRj)WlD*WEO!<qPyC7M;w`be_jh^1Y)+*fxfhVqyo
zK<adkYKeF>yDEG0_+Mr5t8Aai`-m$T&lPPGZWBX?sPlCv-NS`9{{UFjq?$We7aHpn
z(?PY)Y2&<ig~~X?g`&zBW|v#MHj*TkpAtzPsoDv=4Kr95I>Vj*JmX>)6?sK9v;Cdp
z9Ct>dQyOBU@}$u?&;V)y>X(E#VeqzMt6-@-5<@`(bkyi?9^<;}iYyi0-f(^3FPyld
zzc*K#nr3!$A?)r0WF9XN#rKBNIz&`!t~ZbmLpL+GNl|3F^D@#t7Wn)$(P`PLZlT}X
z6@JWYXsK!yer}sBovuus(Bz^9>Nx1~eOiX2^x(|x6krt|?IJ?r9xv00g9#A=LWX{o
zC|v|e5|t7qMY~3p=@-`cB@h&2w08i%ohRtQGL$AIM|_sv))z0*f=YJLVDR+Fuk(Oh
z;Yd=;2#Z8`fP+|87GMztIK#LurHU~&8^W~rnHrT+r7266IfJo~eqr!nE?am~;!Bs&
z5vftMXg<57<ptK(8VR>sSq*8Z@W$>y_n7hbCwPwoRB)arui;9Z<4IGd&76eLae23c
zsyN~CBO8WhCYuNjxYJEW{Jtx_Y6lBl)E?n$Z?5|;*V!H<qR4p|`C)AKsQLV_Q<1{e
z3Tv|R&N|4sC5VL1<@}$Tabp^@9)8r}tF^ilI%sJTZhC*ZwAlX3c%8rV3r~45?AAjS
z?GkdMGd85<29H$4QY{^#-A17yChYT4s-CF>MSNM}_^09U0*hdFg{^l+In85?+wlN8
z%%#KVw!SbOa>)~>U<-gFm5h#23sd3B+t_Bgh*_#QvsiAat0!BMGq|T&)w=Xw&Nh5a
zm!>b&rfX7TsX7Y)EjF=@{{Sh`!`!Nhf}KaUppfE2hWhK`H3!94s`xT)R=4skDrDx4
zX5W)ZhAwiV)1NeGA>Lk(P0>1a+Uh;ENM>JH6B$U<5^N)!+(8<yZ>HuyBRkWiZ^(^x
z{)JiOD0CWrCdH$)Pm~&!X~}Afaun9vZf2rZ=@?$`1iB9EipM?CjBE?)fa*9R+vWhe
zU;ya`1Q$rqTbWuIwz69MMCqW^m=>LYQ`m!RIF5ROH-fN>Aw=2p{{YTtv<F;o<v;%b
zFzJ3VKEKx;zN<kRRPeN!RAMWWRo{uwW20N@uyTv_RGnGGWT)I5ht%qPH0Y19Ty3>2
zmewk`lV)2(nxj*Rk(I8@>gv;{rHlQU<{i#nqKEmXns+}}z6?un)x&94rP>`@J+Bb$
zSUI|F)fD?ig628~^Ks!Dt=%=XPbZC}VZ2*_tC(JU2^~(TbwaLS*$H>(^@FrDnVOIb
zNtW9*Qf(wANI_j9?9UQs6`MmHhp608BDn$~#tV7makT9&5<rOnu7&h!156x%?le9i
zUN?z91<UkiX6JturAqnJB6C&#E+1OnfcUv|Utz}Lfo(C_gL&M8%ajdg$#HPQ(C<|5
zk!aW+zV$qGCV<cJgLZJfb>c7caTIoKJT?)tPgCt~eE$G`3I+9Rm~po)lW@Z9Ic37D
zZ%-U?c1yW*XerQOmKfuexju=<9PMkRraFSU27(~0y@re!)swox;j%l@uGd93u2m^G
zOWL3ZOV*1kf^s(5)tGv*7crF}ICve$f!Rno%st)p>*6#jL8n9m&mRZAT_0bHDmk_A
zKYK*xZsNmUT%FWd{vArN*HXmyDhN9re6OhSD{bsZ9)D|?6SO&Gn-w<ZI6QW7-TJvk
ze;2GVKWFiE0f=dCUBPv`+wWnbx5KG+JYNped(7$}UdjfaUzM3=P4~Z?KdCvz{UgLs
z@D8U!g<5X35$6p8-24Yc3|4xz81|Wwev(O(;;yvmx1AP8RDrW}Tl0UD4y&WJYEG_L
zYmy|018X1iSacR#THwjac+~S-Ijl%c)4olal)9+tH*?-M55hzAqni##6Is_~pl{T8
z{1Q2gWCG!R<Niz?pZQtgsj4>AEB;fpEC;Z*9l80bs=?b-{{Yq7mpi*z7ob}x%E!A6
zUbw^3XnV}u3p13R@6TN%1v+_~CqdVL($-(NZIx}$XBAu#gXex&DR8My{FO-gU?rsH
zu?)vXsDoTNx@fki3az~{p+cn!4fIDiW?{->E!+rcJ;`YqtO&6@^-(R%b2EGRd)FTS
z04QIUhm)Sm_73#^(J=(``zx!~=dgWNr-oW$b-4<QJ#udYok&?0?{+C;aTIaX+lRuq
zx(S#ucEyN`i^zf|Bxr_$B4fRzug31(X#PVRPJVK=v7M{a+te1`Zc=7aW`{5uog5Ba
z&VNMWG;#RsAq=d?p0Ey@?w?^Y;U5+AYZ+{Xx?J!2y1LQIYMUZ$%(m#<JzWJF4C1&)
zwjN7-T1V=EhO06H)z!=yQIlZaW^4}8xVTUJv2Knh4~Q{H%9ToQrr!FI+Z#0*98pB3
zZ?aDfVfj05@qC)We|ciJ3nBOy)<5+tVjb_S@IC-5IN;{unM_|TIz#^ew%PfuM)%4s
zaX86a>lA4BN~$w7N{I(ds_$yM=Npy}+CTV3a@N(x5Jrq27k@?2IJS@=>hGep$W`YW
zwb@#khp(o6MGy^jHur{MGt^Jab=7p4yG(qQ;|NqL)l5@|Bu=7c+tnFCa^n7&AwATa
zK-Cey$u?&-LrUokbB~Upw>nKZiIh>7#iT8f7e)16<cx|GgLqq>M9njx<in}c$!ON9
zU5sTp(ZFB$fYE7UtADMrjbWG3W+t7eJ?94U7Qx3jzxRdz0CyFQ7SzXQaGZjJ02*c@
znILn9gFKY8meVDqjK-t9yTEDHUZ}XG%C%Z@lu|m5NwClqw6g}`JZ`ilYM}a}pZAf@
zs~whu8vLzpR~axR)c|NSXn;?W=}qOq@+VL=IyUy#SX!31Ej2=@%)LJ*tv=VB5S(P#
zw7%SLwc`@{J%h7mC)FX+0~}+!Jr--)Xy-LVK->|j9Ml8Yr2cdTvM*R2i+z)uX97Hz
z0`P7KDh2#T!-d05ZGXKJ%2}x@1;QW&6yuu5dlCR5UJ;2{ZlXlo1s^ip)~%WfqZ0%H
zs##n%QgYNP)XfULi5X?IrqiY*rDjCm0ebnJZ97A#)6JF3$v8bD9Nk`%F%<B|W#PnH
z(8!rO1Ktn~-Euurn5Oeh)kcjV8{fS#ZdNnNLX5-UVRIyf?+kU2bW!TQwf_Jo<n?{M
zl<gj3kwJEHW?vJ9vdvCQ%aeMP_HHECD@&3`&VJe&LyO@%Q<iBP!8%ktzk07>_H?_@
zE0ziGjllLtHQ&zF_OQ*f4eU~O*9TRv`b$O|gQvwY9z=Q>0QUO+NB)a@HGQcc@iX;J
zXR&6}s8SV*#P-If#zs|#rk)|hv`&VG>%W|Ou)}b#aBFL89>e!B%WeJ=g6u;YUTGUi
z^_jPx>a(-Gv}1j!NIzG?n<Ae|V^d`*d$qby4?vdhJIHg1R-;1*;L4STlL4;q<+_bJ
zwQILCDV=$@^@#IV@4^kQ56$dSbe<p!TKpxU@fF=1Q#QvuxLnX{f06$E6y~@S*&p#U
z{{WJv;GM2w?V!;Z#27<lbr)$*4VvF9QQ9UP^%PEG?Dxd$Qm2Ef*y3tI<(0-o?g$XM
z)2i%EF;;^K9xeS<vHV}OrerD;^BX3qHq$MW&J&eNK>4vAFUv;){3$Rhf06$HqUbwe
z-Zb>;-yUf)*sE!jcAgo{md_5Ok4$!Ot5TfP)I0R}rc}5y%c>Kq;LHC2yyQJsY&FwT
zhBpw_#zXIK>ZHRK?LKO2wEqC1DvxE+^9~haiXwa*L&Z0AN|C`;p31L7-FW(ZlX=82
zjMR@yxAjjN#T!LC8=2WO^M1RBR-;-2>u}BZV!&+xB-92vJM`agj4nm>N&f8n+xo4`
zE3uC&DeYN_ZzQZ#Al;sE#%Th}!sA!A=$~z{hSf#Sy72?FmngJwwc12qOMvJS0l#*1
zBW?U9nBq7af?^35iM6aJ$K5Kp{xQaMN<-9^c5zTSrw>f5nxy@&_O$p$Q<}=yp47ih
zEPqM(EimV@hFqrq0K@vNPqb~hSLl;?ep0Ekijxw~elIR=Jxg#kEhDpnac#6ypD4S9
z!y1jK<~vK99W<RZGy1}<Ubv!J+BAu<(^#HsOXIDt;>un)@^-1rQ)VREoy2bhjKi)=
z6HcCh@p+d^{z^7@b`_QB<6AS4f<?gwPch@7p^3zn8Xz?HiRA{}3i+Y5xsR7+nW|F$
zZgx-{aiH~pKe@#oqYmcbs?n!Z#!}%u2mp1xmNqDMt#w*^i>_pweaTlWd7($I=_ub}
zYuha{vK=ANc1)eFeq@C{h{s%@mCTtoQN>hrzbuh307#ojS(T~-XQ_5@$dL*%RP3E~
zzLx-3)w&>9B2W)Cx~cf2<tdg*J!{QroxpzdQgeR&hw&3X;Us*ecg#By-cj=ly|x?W
zQM6nuMZrwXGbdD|{Hrv{$XDb7d6x(Aa`ADtyZj{?_3*T6Vyncr?%I2@G%p(|wxhcW
zLIH=nUb$E7U->ezFl}25)Nypy#p+ZV2(ofc>(#qfAFS$dSPjkQ^M7EnirhMd7T0J5
zk&sD|l*o*xM9Jt+SX{9V9pKTq@wVWPjE{Qy?(II9=*vli4U5w3+bv)Qq@WP#+&-tG
zH_ItF3U&}J%x!IB_7y2Jn^pN${#1Mkx5|9}EPu)6`b*e!wRNXy^<$aDt~vSF4rh+<
zP+HRw>a@z$A2mt*I+iTc#LI(CkE|R4o4^tIt$ijYZhh$i=YEzQM$^lZ{G4sS4`Nzo
zF~kpI8hgytrU|K|IjgUHoR@9G`fGCYjhqE)RPe-8xXj@KE#5F@zymORh}|mSx!}6p
z*4*!F_?_1(nYS}kmkstj?w8Wx{{YG?;u`l;s$P9jn%u<m>GfRc76&p0r!T)>-PW<2
zoO{IDbjF{pak{E(VL-o7a>9uL$wZz?DCeyI04zA>l#-e`IviS#k_<}QAxAkj-QPK!
z2u<8{3yuzmOy6)xJMMuc9O2u=a>JBaIc8&z&hH3`(L1A+jIOg?F%xwp<K+pQJpDTs
zQ=G}kH=Cbq{YRKRQU1(ByHFZ0yx8~~eN%@pimlrIVRVO<H3%iVtM<agkMlXI)WuDo
zRGN6r!Tp%c;SY`hL9X^jQ{vbQLy5L{wL(3pKJhDWc_dQsW-QrHm0Xr0!P1kpDt5~;
zMM{|5(-(d?#+%K&iUy7W52Oq7Vd1{)&M53g8;)TrGNEXFN+91r;z0t(UPrG;od(mb
zfPFVMU_Z=(A#2Yj+i3n$eoq@T3B*lmqtv>)z87a|k`oeIUtzZdek4(IA2^N+j>ou8
zX3IJ<n}dPrnZUc;qk=w<#PHa1=!XE~)!?+xgykI<JGp}8#-%$_#T9jzrQT0BEh5it
zsLcU!aJD}*oBEz0Po)|Z@kR7oPYle)1eYAF8(iGD^k_X{RjEL;5c5>U&4t`&U_VtW
zC6DFWm72lWUO=f7`IZh+PJgwTPn(2ySmT?m7kLshq>1lw$2ZV(TrqWstXpy*Wr4dI
zChXJ~5=fCc1lv+aqcmij1HAGi?Cs(s!c}H(2xIE8HBgn=<x$FX)fJElrf73+v`(<|
zKthBqT3m@ZVwM^BTt5=39Z{ecIwina(nkGxD;4SDD`4rCvBto9r(p+Z*;JdL)TG~-
z`EH#w$S~nH^uK0sCJ4v!c^LV>iSmtSR%8C}4(6HOMp5Pud%#asL{g~Pbw-gJL7#M3
zc&MtCa&CO3rOZ*fdpCAzU{yw$Me1Gq$ydayny-50mQ$xnoghnYPNeF5V@TL&6dva|
z#*v`de)_1<m?Z>3UU{M-cmDv4bU8qk5Ke(~mKT+mmckULm}wE&r{3)c`Q1A#{qfqL
z{IgFr%-*S;l(S4TXU+H`qsz6GZGaJVBofRHpo>vVml%?}Ku2rUh@B9E;0QT8Z@N&n
zQa-6#YY?^!)+fcG%g)^z^vN`uPbHrC6&wdY(J}BGOGJjCE6&mAw8|dQv7HK;EoCY~
zqcYT^SCQ!mkO=j4bMZ!Q4y9aEyG%jNm7tD<i<??{>buozX;X@Jl4V{tBJ;n8Y&WUV
zBRb%HF3xdRkFuZkTdKgE%oEIJ(DU<GcuU{?pxSe2ar1?$>4aY-;m5A?jZ4ef`bQDN
zdX;K^8jX&bpdOb|ssVEcHvFY+>23$)ZSd<C{n5saKs!cs&GJ<naA%_W%}c;|Hw|se
zXe?21;|<HwDuTr|y*)sb4#vxzkKS1GRO;8N*43!uNjYbw?5$K|7+ltle?_or`(4^x
zi#5z`D~zTG6;ncLQk^@09D_VO0r|hP7irU^%B*8s64O0yHC0?mX(fGh*~Zf`?4~2~
z6{9tZUs`1|12u`{+Hkcr>Z3)Rn;E6M=L=m_xbEK`b<TTVRJ6oAk8^}8WC$WO1LU-}
zC&U$eSAqWkl*0jHDZP9g^){D`j-2abqBWy!lR3ms$&nJTOgcWU8U26qBjiZ%P4F9S
zkI{Z(9eK%$6EiNCp6q(pd*WpqnOmPP??oU?O45`)q~oaxj7-#yC3UuCq#h;#n?PIh
zjrV<P`cVpur<7SuvK?i6Jb8+zsy&K!rjAOS?As*UvdtF&%UnJf;2Hx<nn=?^M*Rxx
zOoKgsOQd$)srYq^W(H-!ak*-Jns1m7^EqMUKI~QDRq+jNDgovT-)a8<+<LK&YfdJk
zh}pmBYS)wV4v%8u`N8Ai>W^B`3yCW%{Z^)(Gx=V+cdC5PPY;GU$!&V2rafPgO3=~N
zpi4gZHNx?JlyfhskPOG(q}>YHn>u;c@tzN{#pABigunQvqJ}iOn?u;*BpuHsZw>)*
z1TN&mtxL^I?Lum1QL~$88cXQBYymjW<9*BFj}42fh^3F9%2gcD=9vM`0QZQv1`Vy?
zy&fHupNFXDuo!UE1F&C8Ka<xEVTZhxcEgKiu%v^ymR@qRFwE@LDqZ(L!db^dtUNAn
z%07Aavc5InY%thwNT>m)LO`hYj`N_N>xIteUbYL;@LE`L!0J})+adJ?<p#hu1;&C+
zm(y$kFC|sk?#w1;sjLgNI$a88f3r$8FWMY^mWkh*6zY!+iXZW{skoDs?*d6S4{#SG
zn*g_|6O=`*wYhRviryN-qbw39L6<o`(BSDmEDsI8ZX-r%G&+$j(%_JG+z_Px@yhG?
z-{!tMe=9x!+5QajNvva>CA0cmmNIV^VdjRZAw`v9XHCak)C+(v$g)e5zCMfg6Nz!E
zzie5<>f>uUhLvNG)u>w`t%C)`NSkumd5^SMYzsIs#4>{ev$o-l2IlvT*T%mmw`DQw
z)wZ#10-rD}y{6sNNiZSvW8Mu%(zvcGh~NM?Oifn*0RGyCbIrwhoE01{C;RFwKgiMY
z3#_F!RkJtE{agW0U;hBdHU3!h$Buoa1V%cq_}%)gCh-0g;7hRV{{RSE69?=!WA#eg
zQnd}6$Q-#3K}m!S<KJnr&!<%X01;9gH-tIF96LSBelg^|<n|=<W3n1<ldbu|?3^xG
zR}OH0Z~dmsKjW%f^C<rSW^j!L9k-2_&tES+k!;sxQ=20H07rN}UO;Lxbr)xQ7?AdY
z2PLe}^q2Ib+x?N^JX4F(#vUhK)p8nPxweq$_rP9zjp1;Z7f={VVVSecZw0aLmD1nI
zmIMBdc8&i4_e)>?!I;bYY5xG!u)qG1{{Z<4)B89703l!a$NE>-tYiY*?F!A`>b5_$
z<am1rK<2$GG$m?%rWgMJcAG_W#xxp@zx*?iI#<L~{wDmwRV_jB$o!Uu@+$uT`E6e1
z{{UTo{{SNqcZpvV3H(3#6~hPp!@Pcrvn7BvrOMyR3}<Rgt72txf08hd`Xzt=0MwoT
z0K*v4;}QL!KZoj8cMJajlOwOtE|SEYG#dW^{{WSD`z0BS1GI05`Gt7Nrs?E<N)o77
zRbIQpWJG}^{`D$iRnnj$6j}|wEnn`zhxO|DeNwDUGr+2J)9&_Khacn}{={7xLgGDd
zHTtI66+T*_hc<M~_3kuuYG@ER`SkW#c=k9(HmRiAl4Lhhc{i%@IFeW@Ij}Bv9NfP-
z+he3ts8#uSdQ`TXa~jhxaa=~`vG@M~2=RabNF5imW1RXxy6cQKz{^D59x`n>Y9vGd
z0K=SXaL@^=KQI14XXEZEPvSr6E0jE{OxVMoMUS#-F8&pGf2nFWI6!N$oJ<kykM$Lq
zk0fERL~}ny^Ol(~{M`ykxu;bB01%n${;D=BHLT(kG6?&@{*_1<I}ZHEZ>U-?8-TGt
zP6l@YQ{_wk?lv%FnobAtKmA2!R{rzIe3cGi!xOYhFE>TSGB;fHqDlwWD2$mvgQBoV
znO@HqiW9w3kkxD{S%<Wy`2A1(P>k$bol2EDl*~iEj#epv?7;H+h4!2IG<I!)F{ITR
zp^GrXGjkF(DXEE0)H;C6niAYubjBYOt60g*MAKx&wR5LM+JFVHc6+`-eWut>+cv3-
zys@@rhjA7)ee5|;#_|;D8gBqe$;;~>0Ja0BJd=H~QHLSJ<>ybr?h|bks*Aa!k@*$$
zM;z^|YJcd0t6%&xeJfJ?Uyz_PG!_ekPI6RUIehKqsJX?**Tg#jDwmlLb@$xIgzci^
zl9+J-dp}S<OW<FYp3w2Mhn5C4!&3N_^`Ol5YJ-xlo~vN{@~t}Q4gUZ&ajLFCYDH|$
ztB0#q_cN>yqOnD-b6OpL_g_3V?q^r4&P&xxS+?Zltk>f&=I!^#Y#mCJi<?ZuxX6Lg
zEKNFjv&Migmst9ZOO?eE-I_!~*|u%7>BiLxm0yOboX>IJ0AW(718X!lXAhWMp;O_R
zct$p<UYjS7`6_iicle*0s5;@7QAbg#=Aqs);)x+Jeb+APhwP}GQ*-NtmD6JBhE$Vc
zaP#VffV?C`qXKDN0#<H_>P(?|f+BbW!3i*}^qFQRs95S!VY{V*ob#d{smB+R+LXxV
zyVC*G-W|zxzm&60zqOx~9vtTnXEbS|bJlm{$!<T^v6I7zZVBugR;m8CEgWcXr7-^h
z?6Cg;lpuK(Rw`6!Zzw5>Y|W1U0Hd1oJtABVb~cZz+$Bh<$6D9o@Xc`5%=kIbqr=Gz
zc^hZ{07zXa<Po#{7gWO72Ws^^HpAOS4YyIG+pE48%r?O`!`f-;rl$C-$1xbHc*?v!
zo5*0G<CIuU#rfLTT0AQcg{y@)nv4a)qi??Jwmg_?)p?ze)bPR25aH=9pi(J}$+6~I
za#X^xdq#ot=vD>Vw`Zi^pkvN)7Pq(UPQ8@P>_Ddx$|UI;o@;9e=Va`*&k@dC>7H`}
zXOQH$<7Jcc*oy@1@~d2&H|Lzga+9~b<<`IWLrmjD@b#BZaoZa10-b6Mr)YF}c4cR@
zmol@Rdmr^RSuiB!M9P;<DzwkZfzd15UC?glII<h980Uy>H&b(*(0ag~Ij+=8fei$X
zl4D*<Px(e0ZGgianY3)1r&SppA2+}+vu{ivXZzGhb^F8pGLAD1BQ31&m|rMx)U=e;
z#ItjB#vB(N<5%SxaaS$L)H$cs9;xnoDU!z#17b(XdHQW*iOkS^Kr5&)Gw$F?@0Tz^
z{GD8NqLOlUl8GrWOd6t_cQuh@N{OET018G<DbK`E?C}mS8o?xq8XK$h187w3XaLg@
KAOK^;FaOy<@Ip2K

literal 0
HcmV?d00001


From 06220903a15e72b7603d9eeda329e57dc1368542 Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Thu, 11 Jun 2015 00:20:09 +0530
Subject: [PATCH 17/26] Formatting changes

---
 doc/examples/plot_seam_carving.py   |  5 +++--
 skimage/transform/_seam_carving.pyx | 13 ++++++++-----
 2 files changed, 11 insertions(+), 7 deletions(-)

diff --git a/doc/examples/plot_seam_carving.py b/doc/examples/plot_seam_carving.py
index 9ed7181c..c5250d32 100644
--- a/doc/examples/plot_seam_carving.py
+++ b/doc/examples/plot_seam_carving.py
@@ -66,12 +66,13 @@ importance of those pixels
 
 masked_img = img.copy()
 
-poly = [(404, 281), (404, 360), (359, 364), (338, 337), (145, 337), (120, 322), (145, 304), (340, 306), (362, 284)]
+poly = [(404, 281), (404, 360), (359, 364), (338, 337), (145, 337), (120, 322),
+        (145, 304), (340, 306), (362, 284)]
 pr = np.array([p[0] for p in poly])
 pc = np.array([p[1] for p in poly])
 rr, cc = draw.polygon(pr, pc)
 
-masked_img[rr,cc,:] = masked_img[rr,cc,:]*0.5 + hl_color*.5
+masked_img[rr, cc, :] = masked_img[rr, cc, :]*0.5 + hl_color*.5
 plt.figure()
 plt.title('Object Marked')
 
diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index c2ad8bfa..41ea1d75 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -41,14 +41,14 @@ cdef _preprocess_image(cnp.double_t[:, :, ::1] energy_img,
             for offset in range(-1, 2):
 
                 c_idx = c + offset
-                if (c_idx > cols - 1) or (c_idx < 0) :
+                if (c_idx > cols - 1) or (c_idx < 0):
                     continue
 
                 if cumulative_img[r-1, c_idx] < min_cost:
                     min_cost = cumulative_img[r-1, c_idx]
                     track_img[r, c] = offset
 
-            cumulative_img[r,c] = min_cost + energy_img[r, c, 0]
+            cumulative_img[r, c] = min_cost + energy_img[r, c, 0]
 
 cdef cnp.uint8_t _mark_seam(cnp.int8_t[:, ::1] track_img,
                             Py_ssize_t start_index,
@@ -98,7 +98,7 @@ cdef cnp.uint8_t _mark_seam(cnp.int8_t[:, ::1] track_img,
     return 1
 
 cdef _remove_seam(cnp.double_t[:, :, ::1] img,
-                cnp.uint8_t[:, ::1] seam_map, Py_ssize_t cols):
+                  cnp.uint8_t[:, ::1] seam_map, Py_ssize_t cols):
     """ Removes marked seams from an image.
 
     Parameters
@@ -121,6 +121,7 @@ cdef _remove_seam(cnp.double_t[:, :, ::1] img,
             for ch in range(channels):
                 img[r, c, ch] = img[r, c + shift, ch]
 
+
 def _seam_carve_v(img, energy_map, iters, border):
     """ Carve vertical seams off an image.
 
@@ -160,7 +161,8 @@ def _seam_carve_v(img, energy_map, iters, border):
 
     cdef cnp.double_t[::1] last_row = last_row_obj
     cdef Py_ssize_t[::1] sorted_indices
-    cdef cnp.uint8_t[:, ::1] seam_map = np.zeros(img.shape[0:2], dtype=np.uint8)
+    cdef cnp.uint8_t[:, ::1] seam_map = np.zeros(img.shape[0:2],
+                                                 dtype=np.uint8)
     cdef Py_ssize_t cols = img.shape[1]
     cdef Py_ssize_t rows = img.shape[0]
     cdef Py_ssize_t seams_left = iters
@@ -169,7 +171,8 @@ def _seam_carve_v(img, energy_map, iters, border):
 
     cdef cnp.double_t[:, :, ::1] image = img
     cdef cnp.int8_t[:, ::1] track_img = np.zeros(img.shape[0:2], dtype=np.int8)
-    cdef cnp.double_t[:, ::1] cumulative_img = np.zeros(img.shape[0:2], dtype=np.float)
+    cdef cnp.double_t[:, ::1] cumulative_img = np.zeros(img.shape[0:2],
+                                                        dtype=np.float)
     cdef cnp.double_t[:, :, ::1] energy_img
 
     energy_map[:, 0:border] = DBL_MAX

From 17dc21347710f8b2a28eac924bc3dd61e8c8703a Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Thu, 11 Jun 2015 00:26:57 +0530
Subject: [PATCH 18/26] pep8 changes

---
 skimage/data/__init__.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/skimage/data/__init__.py b/skimage/data/__init__.py
index 98fb4e5f..a2ea2f33 100644
--- a/skimage/data/__init__.py
+++ b/skimage/data/__init__.py
@@ -243,6 +243,7 @@ def hubble_deep_field():
     """
     return load("hubble_deep_field.jpg")
 
+
 def rocket():
     """Launch photo of DSCOVR on Falcon 9 by SpaceX.
 

From 3f9c800514daf1e3fe1a1b8e0560b54cae3cd6f9 Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Thu, 11 Jun 2015 00:47:13 +0530
Subject: [PATCH 19/26] Modfied test to use new API and added code to remove
 remaining seams if any

---
 skimage/transform/_seam_carving.pyx          |  2 ++
 skimage/transform/tests/test_seam_carving.py | 22 ++++----------------
 2 files changed, 6 insertions(+), 18 deletions(-)

diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index 41ea1d75..7159e38a 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -202,4 +202,6 @@ def _seam_carve_v(img, energy_map, iters, border):
             last_row[:cols] = cumulative_img[-1, :cols]
             sorted_indices = np.argsort(last_row_obj)
 
+    _remove_seam(image, seam_map, cols)
+
     return img[:, 0:cols]
diff --git a/skimage/transform/tests/test_seam_carving.py b/skimage/transform/tests/test_seam_carving.py
index 57df75e0..1ff0b32e 100644
--- a/skimage/transform/tests/test_seam_carving.py
+++ b/skimage/transform/tests/test_seam_carving.py
@@ -3,34 +3,20 @@ import numpy as np
 from numpy import testing
 
 
-def energy(img):
-    if(img.ndim == 3):
-        img = np.ascontiguousarray(img[:, :, 0])
-    return (1 - img)
-
-
 def test_seam_carving():
     img = np.array([[0, 0, 1, 0, 0],
                     [0, 0, 1, 0, 0],
                     [0, 0, 1, 0, 0],
                     [0, 1, 0, 0, 0],
                     [1, 0, 0, 0, 0]], dtype=np.float)
+    energy = 1 - img
 
-    out = transform.seam_carve(img, 'horizontal', 1, energy, border=0)
+    out = transform.seam_carve(img, energy, 'vertical', 1, border=0)
+    print out
     testing.assert_allclose(out, 0)
 
     img = img.T
-    out = transform.seam_carve(img, 'vertical', 1, energy, border=0)
-    testing.assert_allclose(out, 0)
-
-    img = img.T
-
-    img3 = np.dstack([img, img, img])
-
-    out = transform.seam_carve(img3, 'horizontal', 1, energy, border=0)
-    testing.assert_allclose(out, 0)
-
-    out = transform.seam_carve(img3, 'vertical', 1, energy, border=0)
+    out = transform.seam_carve(img, energy, 'horizontal', 1, border=0)
     testing.assert_allclose(out, 0)
 
 

From 5a869dc70e24d41b0dce5854def0a10fd33002b3 Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Thu, 11 Jun 2015 00:51:52 +0530
Subject: [PATCH 20/26] Added code comments

---
 skimage/transform/_seam_carving.pyx | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index 7159e38a..fb98d2b9 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -157,6 +157,7 @@ def _seam_carve_v(img, energy_map, iters, border):
            "Seam Carving for Content-Aware Image Resizing"
            http://www.cs.jhu.edu/~misha/ReadingSeminar/Papers/Avidan07.pdf
     """
+    # This reference has been kept to be used for the `np.argsort` call
     last_row_obj = np.zeros(img.shape[1], dtype=np.float)
 
     cdef cnp.double_t[::1] last_row = last_row_obj
@@ -177,6 +178,12 @@ def _seam_carve_v(img, energy_map, iters, border):
 
     energy_map[:, 0:border] = DBL_MAX
     energy_map[:, cols-border:cols] = DBL_MAX
+
+    # Filters often let the boundary be `0`. If the last row is of all same
+    # values, the least value in then penultimate row will give 3 minimum
+    # enries in the last row. Hence, two successive removals will always
+    # intersect as the 3 least seams will share the same pixels except
+    # they will differ in the last row.
     energy_map[rows-border:rows, :] = energy_map[rows-2*border:rows-border, :]
 
     energy_map = np.ascontiguousarray(energy_map[:, :, np.newaxis])

From 557354668ef2cd4a6544aabd3a316b01decdfeab Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Thu, 11 Jun 2015 01:04:46 +0530
Subject: [PATCH 21/26] removed debug print in test

---
 skimage/transform/tests/test_seam_carving.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/skimage/transform/tests/test_seam_carving.py b/skimage/transform/tests/test_seam_carving.py
index 1ff0b32e..b4eb8977 100644
--- a/skimage/transform/tests/test_seam_carving.py
+++ b/skimage/transform/tests/test_seam_carving.py
@@ -12,7 +12,6 @@ def test_seam_carving():
     energy = 1 - img
 
     out = transform.seam_carve(img, energy, 'vertical', 1, border=0)
-    print out
     testing.assert_allclose(out, 0)
 
     img = img.T

From 55fc07edb966b307c27318d66cd669933c16e725 Mon Sep 17 00:00:00 2001
From: Dhruv Jawali <dhruv13.j@gmail.com>
Date: Sat, 13 Jun 2015 23:25:36 +0530
Subject: [PATCH 22/26] removed reduntant continue, added Cython flags and
 fixed negative indexing

---
 skimage/transform/_seam_carving.pyx | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index fb98d2b9..d8dc3945 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -1,3 +1,7 @@
+# cython: cdivision=True
+# cython: boundscheck=False
+# cython: nonecheck=False
+# cython: wraparound=False
 import numpy as np
 cimport numpy as cnp
 
@@ -190,7 +194,7 @@ def _seam_carve_v(img, energy_map, iters, border):
     energy_img = energy_map
 
     _preprocess_image(energy_img, cumulative_img, track_img, cols)
-    last_row[...] = cumulative_img[-1, :]
+    last_row[...] = cumulative_img[rows - 1, :]
     sorted_indices = np.argsort(last_row_obj)
     seam_idx = 0
 
@@ -199,14 +203,13 @@ def _seam_carve_v(img, energy_map, iters, border):
             seams_left -= 1
             cols -= 1
             seam_idx += 1
-            continue
         else:
             seam_idx = 0
             _remove_seam(image, seam_map, cols)
             _remove_seam(energy_img, seam_map, cols)
             seam_map[...] = 0
             _preprocess_image(energy_img, cumulative_img, track_img, cols)
-            last_row[:cols] = cumulative_img[-1, :cols]
+            last_row[:cols] = cumulative_img[rows - 1, :cols]
             sorted_indices = np.argsort(last_row_obj)
 
     _remove_seam(image, seam_map, cols)

From e247e3bacc2f01b96facfc51580f0bad46ae15a9 Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Mon, 15 Jun 2015 20:19:52 +0530
Subject: [PATCH 23/26] Precomputation optimizations and docstring changes

---
 skimage/data/__init__.py            |  2 +-
 skimage/transform/_seam_carving.pyx | 41 ++++++++++++++++-------------
 skimage/transform/seam_carving.py   |  2 +-
 3 files changed, 24 insertions(+), 21 deletions(-)

diff --git a/skimage/data/__init__.py b/skimage/data/__init__.py
index a2ea2f33..14a507d3 100644
--- a/skimage/data/__init__.py
+++ b/skimage/data/__init__.py
@@ -248,7 +248,7 @@ def rocket():
     """Launch photo of DSCOVR on Falcon 9 by SpaceX.
 
     This is the launch photo of Falcon 9 carrying DSCOVR lifted off from
-    SpaceX's Launch Complex 40 at Cape Canaveral Air Force Station, Fla..
+    SpaceX's Launch Complex 40 at Cape Canaveral Air Force Station, FL.
 
     Notes
     -----
diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index d8dc3945..f70c1e9e 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -20,10 +20,11 @@ cdef _preprocess_image(cnp.double_t[:, :, ::1] energy_img,
     Parameters
     ----------
     energy_img : (M, N, 1) ndarray
-       The array of cost of removal of each pixel. Seam carving tries to avoid
-       pixels with high costs.
+        Cost array representing the expense to remove each pixel. Seam carving
+        tries to avoid pixels with high costs.
     cumulative_img : (M, N) ndarray
-       The array to be updated with the total cost of lowest energy seams.
+        The array to be updated inplace with the total cost of lowest energy
+        seams.
     track_img : (M, N) ndarray
         For each pixel, `track_img` stores the relative column offset in
         the previous row which has the lowest value in `cumulative_img`. This
@@ -35,21 +36,24 @@ cdef _preprocess_image(cnp.double_t[:, :, ::1] energy_img,
     cdef Py_ssize_t r, c, offset, c_idx
     cdef Py_ssize_t rows = energy_img.shape[0]
     cdef cnp.double_t min_cost = DBL_MAX
+    cdef Py_ssize_t colsm1 = cols - 1
+    cdef Py_ssize_t rm1
 
     for c in range(cols):
         cumulative_img[0, c] = energy_img[0, c, 0]
 
     for r in range(1, rows):
+        rm1 = r - 1
         for c in range(cols):
             min_cost = DBL_MAX
             for offset in range(-1, 2):
 
                 c_idx = c + offset
-                if (c_idx > cols - 1) or (c_idx < 0):
+                if (c_idx > colsm1) or (c_idx < 0):
                     continue
 
-                if cumulative_img[r-1, c_idx] < min_cost:
-                    min_cost = cumulative_img[r-1, c_idx]
+                if cumulative_img[rm1, c_idx] < min_cost:
+                    min_cost = cumulative_img[rm1, c_idx]
                     track_img[r, c] = offset
 
             cumulative_img[r, c] = min_cost + energy_img[r, c, 0]
@@ -87,7 +91,7 @@ cdef cnp.uint8_t _mark_seam(cnp.int8_t[:, ::1] track_img,
 
     current_seam_indices[rows - 1] = start_index
     for row in range(rows - 2, -1, -1):
-        col = current_seam_indices[row+1]
+        col = current_seam_indices[row + 1]
         offset = track_img[row, col]
         col = col + offset
         current_seam_indices[row] = col
@@ -103,7 +107,7 @@ cdef cnp.uint8_t _mark_seam(cnp.int8_t[:, ::1] track_img,
 
 cdef _remove_seam(cnp.double_t[:, :, ::1] img,
                   cnp.uint8_t[:, ::1] seam_map, Py_ssize_t cols):
-    """ Removes marked seams from an image.
+    """ Remove marked seams from an image.
 
     Parameters
     ----------
@@ -112,38 +116,37 @@ cdef _remove_seam(cnp.double_t[:, :, ::1] img,
     seam_map : (M, N) ndarray
         Array with seams to be removed marked by non-zero entries.
     cols : int
-        The number of colums to process.
+        The number of columns to process.
     """
     cdef Py_ssize_t rows = img.shape[0]
     cdef Py_ssize_t channels = img.shape[2]
     cdef Py_ssize_t r, c, ch, shift
+    cdef Py_ssize_t c_shift
 
     for r in range(rows):
         shift = 0
         for c in range(cols):
             shift += seam_map[r, c]
+            c_shift = c + shift
             for ch in range(channels):
-                img[r, c, ch] = img[r, c + shift, ch]
+                img[r, c, ch] = img[r, c_shift, ch]
 
 
 def _seam_carve_v(img, energy_map, iters, border):
     """ Carve vertical seams off an image.
 
-    Carves out vertical seams off an image while using the given energy
-    map to decide the importance of each pixel.[1]
+    Carves out vertical seams from an image while using the given energy map to
+    decide the importance of each pixel.[1]_
 
     Parameters
     ----------
     img : (M, N) or (M, N, 3) ndarray
         Input image whose vertical seams are to be removed.
-    iters : int
-        Number of vertical seams are to be removed.
     energy_map : (M, N) ndarray
-        The array to decide the importance of each pixel. The higher
-        the value corresponding to a pixel, the more the algorithm will try
-        to keep it in the image.
-    num : int
-        Number of seams are to be removed.
+        Cost array denoting importance of each pixel. The algorithm will try to
+        retain high valued pixels.
+    iters : int
+        Number of vertical seams to be removed.
     border : int, optional
         The number of pixels in the right, left and bottom end of the image
         to be excluded from being considered for a seam. This is important as
diff --git a/skimage/transform/seam_carving.py b/skimage/transform/seam_carving.py
index e037b6e1..95e975ce 100644
--- a/skimage/transform/seam_carving.py
+++ b/skimage/transform/seam_carving.py
@@ -7,7 +7,7 @@ import numpy as np
 def seam_carve(img, energy_map, mode, num, border=1, force_copy=True):
     """ Carve vertical or horizontal seams off an image.
 
-    Carves out vertical/horizontal seams off an image while using the given
+    Carves out vertical/horizontal seams from an image while using the given
     energy map to decide the importance of each pixel.
 
     Parameters

From 958f410a5c55e124f9ae10a9724104eb03e0691b Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Mon, 15 Jun 2015 20:47:24 +0530
Subject: [PATCH 24/26] Changes example explanation text

---
 doc/examples/plot_seam_carving.py | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/doc/examples/plot_seam_carving.py b/doc/examples/plot_seam_carving.py
index c5250d32..f1dcb948 100644
--- a/doc/examples/plot_seam_carving.py
+++ b/doc/examples/plot_seam_carving.py
@@ -4,9 +4,10 @@ Seam Carving
 ============
 
 This example demonstrates how images can be resized using seam carving [1]_.
-Resizing often distorts contents in the image. Seam carving tries to resize
-images while trying to keep important content intact. In this example we are
-using the Sobel filter to signify the importance of each pixel.
+Resizing to a new aspect ratio distorts image contents. Seam carving attempts
+to resize *without* distortion, by removing regions of an image which are less
+important. In this example we are using the Sobel filter to signify the
+importance of each pixel.
 
 .. [1] Shai Avidan and Ariel Shamir
        "Seam Carving for Content-Aware Image Resizing"

From fee667b35cbac9165cbdbbd797a5909d8959927a Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Mon, 15 Jun 2015 20:49:26 +0530
Subject: [PATCH 25/26] preallocated seam_buffer and added noil

---
 skimage/transform/_seam_carving.pyx | 24 +++++++++++++++---------
 1 file changed, 15 insertions(+), 9 deletions(-)

diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index f70c1e9e..9fe0789a 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -8,10 +8,10 @@ cimport numpy as cnp
 
 cdef cnp.double_t DBL_MAX = np.finfo(np.double).max
 
-cdef _preprocess_image(cnp.double_t[:, :, ::1] energy_img,
-                       cnp.double_t[:, ::1] cumulative_img,
-                       cnp.int8_t[:, ::1] track_img,
-                       Py_ssize_t cols):
+cdef void _preprocess_image(cnp.double_t[:, :, ::1] energy_img,
+                            cnp.double_t[:, ::1] cumulative_img,
+                            cnp.int8_t[:, ::1] track_img,
+                            Py_ssize_t cols) nogil:
     """ For each row, compute the lowest seam value for all its columns.
 
     This function updates `cumulative_img` such that `cumulative_img[r, c]`
@@ -60,7 +60,8 @@ cdef _preprocess_image(cnp.double_t[:, :, ::1] energy_img,
 
 cdef cnp.uint8_t _mark_seam(cnp.int8_t[:, ::1] track_img,
                             Py_ssize_t start_index,
-                            cnp.uint8_t[:, ::1] seam_map):
+                            cnp.uint8_t[:, ::1] seam_map,
+                            Py_ssize_t[::1] seam_buffer) nogil:
 
     """ Re-trace the optimal seam from a given column in the last row.
 
@@ -77,6 +78,9 @@ cdef cnp.uint8_t _mark_seam(cnp.int8_t[:, ::1] track_img,
     seam_map : (M, N) ndarray
         The array used to mark seams. If a pixel is marked as as seam it is set
         to `1`, else `0`.
+    seam_buffer : (M,) ndarray
+        Buffer used to store the column indices of the seam currently being
+        checked. This is preallocated to save time.
 
     Returns
     -------
@@ -84,7 +88,7 @@ cdef cnp.uint8_t _mark_seam(cnp.int8_t[:, ::1] track_img,
         `1` if seam was marked, `0` is seam inersects and was not marked.
     """
     cdef Py_ssize_t rows = track_img.shape[0]
-    cdef Py_ssize_t[::1] current_seam_indices = np.zeros(rows, dtype=np.int)
+    cdef Py_ssize_t[::1] current_seam_indices = seam_buffer
     cdef Py_ssize_t row, col
     cdef cnp.int8_t offset
     cdef Py_ssize_t seams
@@ -105,8 +109,8 @@ cdef cnp.uint8_t _mark_seam(cnp.int8_t[:, ::1] track_img,
 
     return 1
 
-cdef _remove_seam(cnp.double_t[:, :, ::1] img,
-                  cnp.uint8_t[:, ::1] seam_map, Py_ssize_t cols):
+cdef void _remove_seam(cnp.double_t[:, :, ::1] img,
+                       cnp.uint8_t[:, ::1] seam_map, Py_ssize_t cols) nogil:
     """ Remove marked seams from an image.
 
     Parameters
@@ -176,6 +180,7 @@ def _seam_carve_v(img, energy_map, iters, border):
     cdef Py_ssize_t seams_left = iters
     cdef Py_ssize_t seams_removed
     cdef Py_ssize_t seam_idx
+    cdef Py_ssize_t[::1] seam_buffer = np.zeros(rows, dtype=np.int)
 
     cdef cnp.double_t[:, :, ::1] image = img
     cdef cnp.int8_t[:, ::1] track_img = np.zeros(img.shape[0:2], dtype=np.int8)
@@ -202,7 +207,8 @@ def _seam_carve_v(img, energy_map, iters, border):
     seam_idx = 0
 
     while seams_left > 0:
-        if _mark_seam(track_img, sorted_indices[seam_idx], seam_map):
+        if _mark_seam(track_img, sorted_indices[seam_idx], seam_map,
+                      seam_buffer):
             seams_left -= 1
             cols -= 1
             seam_idx += 1

From 85946c8eed61417a5da8555136fe2438e4c68fc8 Mon Sep 17 00:00:00 2001
From: Vighnesh Birodkar <vighneshbirodkar@gmail.com>
Date: Tue, 16 Jun 2015 19:48:14 +0530
Subject: [PATCH 26/26] Changed return type to bint and improved code comments

---
 skimage/transform/_seam_carving.pyx | 22 +++++++++++-----------
 1 file changed, 11 insertions(+), 11 deletions(-)

diff --git a/skimage/transform/_seam_carving.pyx b/skimage/transform/_seam_carving.pyx
index 9fe0789a..3499514c 100644
--- a/skimage/transform/_seam_carving.pyx
+++ b/skimage/transform/_seam_carving.pyx
@@ -58,10 +58,10 @@ cdef void _preprocess_image(cnp.double_t[:, :, ::1] energy_img,
 
             cumulative_img[r, c] = min_cost + energy_img[r, c, 0]
 
-cdef cnp.uint8_t _mark_seam(cnp.int8_t[:, ::1] track_img,
-                            Py_ssize_t start_index,
-                            cnp.uint8_t[:, ::1] seam_map,
-                            Py_ssize_t[::1] seam_buffer) nogil:
+cdef bint _mark_seam(cnp.int8_t[:, ::1] track_img,
+                     Py_ssize_t start_index,
+                     cnp.uint8_t[:, ::1] seam_map,
+                     Py_ssize_t[::1] seam_buffer) nogil:
 
     """ Re-trace the optimal seam from a given column in the last row.
 
@@ -84,8 +84,8 @@ cdef cnp.uint8_t _mark_seam(cnp.int8_t[:, ::1] track_img,
 
     Returns
     -------
-    out : int
-        `1` if seam was marked, `0` is seam inersects and was not marked.
+    success : int
+        `1` if seam was marked, `0` is seam intersects and was not marked.
     """
     cdef Py_ssize_t rows = track_img.shape[0]
     cdef Py_ssize_t[::1] current_seam_indices = seam_buffer
@@ -191,11 +191,11 @@ def _seam_carve_v(img, energy_map, iters, border):
     energy_map[:, 0:border] = DBL_MAX
     energy_map[:, cols-border:cols] = DBL_MAX
 
-    # Filters often let the boundary be `0`. If the last row is of all same
-    # values, the least value in then penultimate row will give 3 minimum
-    # enries in the last row. Hence, two successive removals will always
-    # intersect as the 3 least seams will share the same pixels except
-    # they will differ in the last row.
+    # Filters often let the boundary be `0`. If all the entries in the last
+    # row of `energy_img` are equal, the minimum value in the penultimate row
+    # of `cumulative_img` will result in 3 minimum values in its last row.
+    # Hence, two successive removals will always intersect as the 3 least seams
+    # will share the same pixels except they will differ in the last row.
     energy_map[rows-border:rows, :] = energy_map[rows-2*border:rows-border, :]
 
     energy_map = np.ascontiguousarray(energy_map[:, :, np.newaxis])