wassname/simpeg
1
0
Fork 0
mirror of https://github.com/wassname/simpeg.git synced 2026-06-27 16:16:53 +08:00
Code Issues Packages Projects Releases Wiki Activity

Updates to the documentation.

Browse Source
This commit is contained in:
Rowan Cockett
2016-01-10 17:21:01 -08:00
parent 8a61259cab
commit 2c5b19b7a0
23 changed files with 316 additions and 282 deletions
Download Patch File Download Diff File Expand all files Collapse all files
AUTHORS.rst
+7 -3
View File
@@ -1,9 +1,13 @@
- Luz Angelica Caudillo-Mata, (`@lacmajedrez <https://github.com/lacmajedrez/>`_)
- Rowan Cockett, (`@rowanc1 <https://github.com/rowanc1/>`_)
- Eldad Haber, (`@ehaber99 <https://github.com/ehaber99/>`_)
- Lindsey Heagy, (`@lheagy <https://github.com/lheagy/>`_)
- Seogi Kang, (`@sgkang <https://github.com/sgkang/>`_)
- Dave Marchant, (`@dwfmarchant <https://github.com/dwfmarchant/>`_)
- Brendan Smithyman, (`@bsmithyman <https://github.com/bsmithyman/>`_)
- Gudni Rosenkjaer, (`@grosenkj <https://github.com/grosenkj/>`_)
- Dom Fournier, (`@fourndo <https://github.com/fourndo/>`_)
- Dave Marchant, (`@dwfmarchant <https://github.com/dwfmarchant/>`_)
- Lars Ruthotto, (`@lruthotto <https://github.com/lruthotto/>`_)
- Mike Wathen, (`@mrwathen <https://github.com/mrwathen/>`_)
- Luz Angelica Caudillo-Mata, (`@lacmajedrez <https://github.com/lacmajedrez/>`_)
- Eldad Haber, (`@ehaber99 <https://github.com/ehaber99/>`_)
- Doug Oldenburg, (`@dougoldenburg <https://github.com/dougoldenburg/>`_)
- Adam Pidlisecky, (`@aPid1 <https://github.com/aPid1/>`_)
CITATION.rst
+2 -3
View File
@@ -1,14 +1,13 @@
Citing SimPEG
=============
-------------
There is a paper about SimPEG!
There is a `paper about SimPEG <http://dx.doi.org/10.1016/j.cageo.2015.09.015>`_, if you use this code, please help our scientific visibility by citing our work!
Cockett, R., Kang, S., Heagy, L. J., Pidlisecky, A., & Oldenburg, D. W. (2015). SimPEG: An open source framework for simulation and gradient based parameter estimation in geophysical applications. Computers & Geosciences.
BibTex:
-------
.. code::
LICENSE
+1 -1
View File
@@ -1,6 +1,6 @@
The MIT License (MIT)
Copyright (c) 2013-2015 SimPEG Developers
Copyright (c) 2013-2016 SimPEG Developers
Permission is hereby granted, free of charge, to any person obtaining a copy of
this software and associated documentation files (the "Software"), to deal in
PROJECTS.rst
-36
View File
@@ -1,36 +0,0 @@
- Electromagnetics (`simpegEM <http://simpegem.rtfd.org/>`_)
.. image:: https://travis-ci.org/simpeg/simpegem.svg?branch=master
:target: https://travis-ci.org/simpeg/simpegem
:alt: Master Branch
.. image:: https://coveralls.io/repos/simpeg/simpegem/badge.png?branch=master
:target: https://coveralls.io/r/simpeg/simpegem?branch=master
- Potential Fields (`simpegPF <http://simpegpf.rtfd.org/>`_)
.. image:: https://travis-ci.org/simpeg/simpegpf.svg?branch=master
:target: https://travis-ci.org/simpeg/simpegpf
:alt: Master Branch
.. image:: https://coveralls.io/repos/simpeg/simpegpf/badge.png?branch=master
:target: https://coveralls.io/r/simpeg/simpegpf?branch=master
- Ground Water Flow (`simpegFLOW <http://simpegflow.rtfd.org/>`_)
.. image:: https://travis-ci.org/simpeg/simpegflow.svg?branch=master
:target: https://travis-ci.org/simpeg/simpegflow
:alt: Master Branch
.. image:: https://coveralls.io/repos/simpeg/simpegflow/badge.png?branch=master
:target: https://coveralls.io/r/simpeg/simpegflow?branch=master
- Direct Current Resistivity (`simpegDC <http://simpeg-dc.rtfd.org/>`_)
.. image:: https://travis-ci.org/simpeg/simpegdc.svg?branch=master
:target: https://travis-ci.org/simpeg/simpegdc
:alt: Master Branch
.. image:: https://coveralls.io/repos/simpeg/simpegdc/badge.png?branch=master
:target: https://coveralls.io/r/simpeg/simpegdc?branch=master
- Electromagnetics 1D (`simpegEM1D <http://simpegem1d.rtfd.org/>`_)
.. image:: https://travis-ci.org/simpeg/simpegEM1D.svg?branch=master
:target: https://travis-ci.org/simpeg/simpegEM1D
:alt: Master Branch
.. image:: https://coveralls.io/repos/simpeg/simpegEM1D/badge.png?branch=master
:target: https://coveralls.io/r/simpeg/simpegEM1D?branch=master
- Magnetotellurics (`simpegMT <http://simpegmt.rtfd.org/>`_)
.. image:: https://travis-ci.org/simpeg/simpegmt.svg?branch=master
:target: https://travis-ci.org/simpeg/simpegmt
:alt: Master Branch
.. image:: https://coveralls.io/repos/simpeg/simpegmt/badge.png?branch=master
:target: https://coveralls.io/r/simpeg/simpegmt?branch=master
SimPEG/EM/Analytics/FDEM.py
+2 -1
View File
@@ -58,8 +58,9 @@ def MagneticDipoleWholeSpace(XYZ, srcLoc, sig, f, moment=1., orientation='X', mu
from SimPEG import EM
import matplotlib.pyplot as plt
from scipy.constants import mu_0
freqs = np.logspace(-2,5,100)
Bx, By, Bz = EM.Analytics.FDEM.AnalyticMagDipoleWholeSpace([0,100,0], [0,0,0], 1e-2, freqs, m=1, orientation='Z')
Bx, By, Bz = EM.Analytics.FDEM.MagneticDipoleWholeSpace([0,100,0], [0,0,0], 1e-2, freqs, moment=1, orientation='Z')
plt.loglog(freqs, np.abs(Bz.real)/mu_0, 'b')
plt.loglog(freqs, np.abs(Bz.imag)/mu_0, 'r')
plt.legend(('real','imag'))
SimPEG/Regularization.py
+2 -81
View File
@@ -115,86 +115,7 @@ class BaseRegularization(object):
class Tikhonov(BaseRegularization):
"""**Tikhonov Regularization**
Here we will define regularization of a model, m, in general however, this should be thought of as (m-m_ref) but otherwise it is exactly the same:
.. math::
R(m) = \int_\Omega \\frac{\\alpha_x}{2}\left(\\frac{\partial m}{\partial x}\\right)^2 + \\frac{\\alpha_y}{2}\left(\\frac{\partial m}{\partial y}\\right)^2 \partial v
Our discrete gradient operator works on cell centers and gives the derivative on the cell faces, which is not where we want to be evaluating this integral. We need to average the values back to the cell-centers before we integrate. To avoid null spaces, we square first and then average. In 2D with ij notation it looks like this:
.. math::
R(m) \\approx \sum_{ij} \left[\\frac{\\alpha_x}{2}\left[\left(\\frac{m_{i+1,j} - m_{i,j}}{h}\\right)^2 + \left(\\frac{m_{i,j} - m_{i-1,j}}{h}\\right)^2\\right]
+ \\frac{\\alpha_y}{2}\left[\left(\\frac{m_{i,j+1} - m_{i,j}}{h}\\right)^2 + \left(\\frac{m_{i,j} - m_{i,j-1}}{h}\\right)^2\\right]
\\right]h^2
If we let D_1 be the derivative matrix in the x direction
.. math::
\mathbf{D}_1 = \mathbf{I}_2\otimes\mathbf{d}_1
.. math::
\mathbf{D}_2 = \mathbf{d}_2\otimes\mathbf{I}_1
Where d_1 is the one dimensional derivative:
.. math::
\mathbf{d}_1 = \\frac{1}{h} \left[ \\begin{array}{cccc}
-1 & 1 & & \\\\
& \ddots & \ddots&\\\\
& & -1 & 1\end{array} \\right]
.. math::
R(m) \\approx \mathbf{v}^\\top \left[\\frac{\\alpha_x}{2}\mathbf{A}_1 (\mathbf{D}_1 m) \odot (\mathbf{D}_1 m) + \\frac{\\alpha_y}{2}\mathbf{A}_2 (\mathbf{D}_2 m) \odot (\mathbf{D}_2 m) \\right]
Recall that this is really a just point wise multiplication, or a diagonal matrix times a vector. When we multiply by something in a diagonal we can interchange and it gives the same results (i.e. it is point wise)
.. math::
\mathbf{a\odot b} = \\text{diag}(\mathbf{a})\mathbf{b} = \\text{diag}(\mathbf{b})\mathbf{a} = \mathbf{b\odot a}
and the transpose also is true (but the sizes have to make sense...):
.. math::
\mathbf{a}^\\top\\text{diag}(\mathbf{b}) = \mathbf{b}^\\top\\text{diag}(\mathbf{a})
So R(m) can simplify to:
.. math::
R(m) \\approx \mathbf{m}^\\top \left[\\frac{\\alpha_x}{2}\mathbf{D}_1^\\top \\text{diag}(\mathbf{A}_1^\\top\mathbf{v}) \mathbf{D}_1 + \\frac{\\alpha_y}{2}\mathbf{D}_2^\\top \\text{diag}(\mathbf{A}_2^\\top \mathbf{v}) \mathbf{D}_2 \\right] \mathbf{m}
We will define W_x as:
.. math::
\mathbf{W}_x = \sqrt{\\alpha_x}\\text{diag}\left(\sqrt{\mathbf{A}_1^\\top\mathbf{v}}\\right) \mathbf{D}_1
And then W as a tall matrix of all of the different regularization terms:
.. math::
\mathbf{W} = \left[ \\begin{array}{c}
\mathbf{W}_s\\\\
\mathbf{W}_x\\\\
\mathbf{W}_y\end{array} \\right]
Then we can write
.. math::
R(m) \\approx \\frac{1}{2}\mathbf{m^\\top W^\\top W m}
"""
"""
smoothModel = True #: SMOOTH and SMOOTH_MOD_DIF options
alpha_s = Utils.dependentProperty('_alpha_s', 1e-6, ['_W', '_Ws'], "Smallness weight")
@@ -311,7 +232,7 @@ class Tikhonov(BaseRegularization):
if self.smoothModel == True:
mD1 = self.mapping.deriv(m)
mD2 = self.mapping.deriv(m - self.mref)
r1 = self.Wsmooth * ( self.mapping * (m))
r1 = self.Wsmooth * ( self.mapping * (m))
r2 = self.Ws * ( self.mapping * (m - self.mref) )
out1 = mD1.T * ( self.Wsmooth.T * r1 )
out2 = mD2.T * ( self.Ws.T * r2 )
docs/InversionWorkflow-PreSimPEG.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 49 KiB

docs/api_FiniteVolume.rst
+19
View File
@@ -0,0 +1,19 @@
.. _api_FiniteVolume:
Finite Volume
*************
Any numerical implementation requires the discretization of continuous functions into discrete approximations. These approximations are typically organized in a mesh, which defines boundaries, locations, and connectivity. Of specific interest to geophysical simulations, we require that averaging, interpolation and differential operators be defined for any mesh. In SimPEG, we have implemented a staggered mimetic finite volume approach (`Hyman and Shashkov, 1999 <http://math.lanl.gov/~mac/papers/numerics/HS99B.pdf>`_). This approach requires the definitions of variables at either cell-centers, nodes, faces, or edges as seen in the figure below.
.. image:: images/finitevolrealestate.png
:width: 400 px
:alt: FiniteVolume
:align: center
.. toctree::
:maxdepth: 2
api_Mesh
api_DiffOps
api_InnerProducts
docs/api_ForwardProblem.rst
+1 -6
View File
@@ -61,11 +61,6 @@ If the forward problem is invertible, then we can rearrange for \\(\\frac{\\part
This can often be computed given a vector (i.e. \\(J(v)\\)) rather than stored, as \\(J\\) is a large dense matrix.
.. math::
u(m)
The API
=======
@@ -78,7 +73,7 @@ Problem
Survey
------
.. automodule:: SimPEG.Survey
:members:
:undoc-members:
docs/api_Inverse.rst → docs/api_Inversion.rst
+6 -17
View File
@@ -1,21 +1,19 @@
.. _api_Inverse:
Regularization
**************
InvProblem
**********
.. automodule:: SimPEG.Regularization
.. automodule:: SimPEG.InvProblem
:show-inheritance:
:members:
:undoc-members:
Optimize
********
Inversion
*********
.. automodule:: SimPEG.Optimization
.. automodule:: SimPEG.Inversion
:show-inheritance:
:private-members:
:members:
:undoc-members:
@@ -27,12 +25,3 @@ Directives
:members:
:undoc-members:
Inversion
*********
.. automodule:: SimPEG.Inversion
:show-inheritance:
:members:
:undoc-members:
docs/api_InversionComponents.rst
+11
View File
@@ -0,0 +1,11 @@
Inversion Components
********************
.. toctree::
:maxdepth: 3
api_DataMisfit
api_Regularization
api_Optimization
api_Inversion
docs/api_Mesh.rst
+1 -2
View File
@@ -24,8 +24,7 @@ the implementations.
.. plot::
from SimPEG import Examples
Examples.Mesh_ThreeMeshes.run()
Examples.Mesh_Basic_Types.run()
Variable Locations and Terminology
docs/api_MeshCode.rst
+10 -10
View File
@@ -9,6 +9,15 @@ Tensor Mesh
:undoc-members:
Cylindrical Mesh
================
.. automodule:: SimPEG.Mesh.CylMesh
:show-inheritance:
:members:
:undoc-members:
Tree Mesh
=========
@@ -21,16 +30,7 @@ Tree Mesh
Curvilinear Mesh
================
.. automodule:: SimPEG.Mesh.Curvilinear
:show-inheritance:
:members:
:undoc-members:
Cylindrical Mesh
================
.. automodule:: SimPEG.Mesh.CylMesh
.. automodule:: SimPEG.Mesh.CurvilinearMesh
:show-inheritance:
:members:
:undoc-members:
docs/api_Optimization.rst
+9
View File
@@ -0,0 +1,9 @@
Optimize
********
.. automodule:: SimPEG.Optimization
:show-inheritance:
:private-members:
:members:
:undoc-members:
docs/api_Regularization.rst
+100
View File
@@ -0,0 +1,100 @@
Regularization
**************
If there is one model that has a misfit that equals the desired tolerance, then there are infinitely many other models which can fit to the same degree. The challenge is to find that model which has the desired characteristics and is compatible with a priori information. A single model can be selected from an infinite ensemble by measuring the length, or norm, of each model. Then a smallest, or sometimes largest, member can be isolated. Our goal is to design a norm that embodies our prior knowledge and, when minimized, yields a realistic candidate for the solution of our problem. The norm can penalize variation from a reference model, spatial derivatives of the model, or some combination of these.
Tikhonov Regularization
=======================
Here we will define regularization of a model, m, in general however, this should be thought of as (m-m_ref) but otherwise it is exactly the same:
.. math::
R(m) = \int_\Omega \frac{\alpha_x}{2}\left(\frac{\partial m}{\partial x}\right)^2 + \frac{\alpha_y}{2}\left(\frac{\partial m}{\partial y}\right)^2 \partial v
Our discrete gradient operator works on cell centers and gives the derivative on the cell faces, which is not where we want to be evaluating this integral. We need to average the values back to the cell-centers before we integrate. To avoid null spaces, we square first and then average. In 2D with ij notation it looks like this:
.. math::
R(m) \approx \sum_{ij} \left[\frac{\alpha_x}{2}\left[\left(\frac{m_{i+1,j} - m_{i,j}}{h}\right)^2 + \left(\frac{m_{i,j} - m_{i-1,j}}{h}\right)^2\right] \\
+ \frac{\alpha_y}{2}\left[\left(\frac{m_{i,j+1} - m_{i,j}}{h}\right)^2 + \left(\frac{m_{i,j} - m_{i,j-1}}{h}\right)^2\right]
\right]h^2
If we let D_1 be the derivative matrix in the x direction
.. math::
\mathbf{D}_1 = \mathbf{I}_2\otimes\mathbf{d}_1
.. math::
\mathbf{D}_2 = \mathbf{d}_2\otimes\mathbf{I}_1
Where d_1 is the one dimensional derivative:
.. math::
\mathbf{d}_1 = \frac{1}{h} \left[ \begin{array}{cccc}
-1 & 1 & & \\
& \ddots & \ddots&\\
& & -1 & 1\end{array} \right]
.. math::
R(m) \approx \mathbf{v}^\top \left[\frac{\alpha_x}{2}\mathbf{A}_1 (\mathbf{D}_1 m) \odot (\mathbf{D}_1 m) + \frac{\alpha_y}{2}\mathbf{A}_2 (\mathbf{D}_2 m) \odot (\mathbf{D}_2 m) \right]
Recall that this is really a just point wise multiplication, or a diagonal matrix times a vector. When we multiply by something in a diagonal we can interchange and it gives the same results (i.e. it is point wise)
.. math::
\mathbf{a\odot b} = \text{diag}(\mathbf{a})\mathbf{b} = \text{diag}(\mathbf{b})\mathbf{a} = \mathbf{b\odot a}
and the transpose also is true (but the sizes have to make sense...):
.. math::
\mathbf{a}^\top\text{diag}(\mathbf{b}) = \mathbf{b}^\top\text{diag}(\mathbf{a})
So R(m) can simplify to:
.. math::
R(m) \approx \mathbf{m}^\top \left[\frac{\alpha_x}{2}\mathbf{D}_1^\top \text{diag}(\mathbf{A}_1^\top\mathbf{v}) \mathbf{D}_1 + \frac{\alpha_y}{2}\mathbf{D}_2^\top \text{diag}(\mathbf{A}_2^\top \mathbf{v}) \mathbf{D}_2 \right] \mathbf{m}
We will define W_x as:
.. math::
\mathbf{W}_x = \sqrt{\alpha_x}\text{diag}\left(\sqrt{\mathbf{A}_1^\top\mathbf{v}}\right) \mathbf{D}_1
And then W as a tall matrix of all of the different regularization terms:
.. math::
\mathbf{W} = \left[ \begin{array}{c}
\mathbf{W}_s\\
\mathbf{W}_x\\
\mathbf{W}_y\end{array} \right]
Then we can write
.. math::
R(m) \approx \frac{1}{2}\mathbf{m^\top W^\top W m}
The API
-------
.. autoclass:: SimPEG.Regularization.BaseRegularization
:members:
:undoc-members:
.. autoclass:: SimPEG.Regularization.Tikhonov
:show-inheritance:
:members:
docs/api_Utilities.rst
+10
View File
@@ -0,0 +1,10 @@
Utilities
*********
.. toctree::
:maxdepth: 2
api_Solver
api_Maps
api_Utils
api_Tests
docs/api_Utils.rst
+3 -6
View File
@@ -1,8 +1,5 @@
.. _api_Utils:
Utilities
*********
Utils
*****
.. automodule:: SimPEG.Utils
:members:
@@ -52,7 +49,7 @@ Interpolation Utilities
:undoc-members:
Counter Utilities
=======================
=================
::
class MyClass(object):
docs/api_bigPicture.rst
+58 -6
View File
@@ -1,17 +1,69 @@
.. _api_license:
Why SimPEG?
***********
===========
Our essential functions as researchers are the pursuit and dissemination of knowledge through research and education. As scientists we
seek to find models that reproduce the observations that we make in the world. In geophysics, we use inverse theory to mathematically
create models of the earth from measured data. It is a difficult problem with many moving pieces: physics, discretization, simulation,
regularization, optimization, computer science, linear algebra, geology. Exploring each of these disciplines can take a career, if you
are so inclined, but as geophysicists we care about the combination: how to pull these disciplines together to answer our questions.
This is the first problem we hope to help solve: to create a toolbox for the geophysicist that allows you to work at a high level and
keep your geophysical question in focus. However, a toolbox is not enough. The research questions that we are interested in surround
the integration of information to make better decisions.
We believe that the feedback loops in the geosciences could use some serious work. For example, collect multiple data-sets from the
same field area (geology, seismic, electromagnetics, hydrogeology), process the data separately, and then reconvene with your
multidisciplinary team. You may be rather surprised (or not) that the everyone has a (completely!?) different model. Dissonant at best,
but often conflicting in the details. Therein lies the second problem: how do we integrate these geoscience fields? Not by force or
even by default, but at least to have the option of quantitative communication and built in feedback loops. What we require is an
implementation that is inherently and unequivocally modular, with all pieces available to manipulation. Black-box software, where the
implementations are hidden, obfuscated, or difficult to manipulate, do not promote experimentation and investigation. We are working on
a framework that exposes the details of the implementation to the geophysicist in a manner that promotes productivity and question
based interrogation. This framework can be easily extended to encompass many geophysical problems and is built with the inverse problem
as the fundamental goal.
The future we see is a mix of tools that span our disciplines, and a framework that allows us to integrate many different types of
geophysical data so that we can communicate effectively and experiment efficiently. A toolbox combined with a framework that allows you
to solve your own problems, and creates opportunities for us to work together to better image and understand the subsurface. What we
are building is called SimPEG, simulation and parameter estimation in geophysics. We are building it in the open. We are testing it.
Breaking it. Building it. Fixing it. Using it. If you believe, like we do, that geophysics can be more innovative and informative in
the open and that these tools are necessary and invaluable in education as well as research, then you should get in touch. There is a
lot of work to do!
The Big Picture
===============
---------------
Defining a well-posed inverse problem and solving it is a complex task that requires many components that must interact. It is helpful
to view this task as a workflow in which various elements are explicitly identified and integrated. The figure below outlines the inversion components that consists of inputs, implementation, and evaluation. The inputs are composed of the geophysical data, the equations which are a mathematical description of the governing physics, and prior knowledge or assumptions about the setting. The implementation consists of two broad categories: the forward simulation and the inversion. The **forward simulation** is the means by which we solve the governing equations given a model and the **inversion components** evaluate and update this model. We are considering a gradient based approach, which updates the model through an optimization routine. The output of this implementation is a model, which, prior to interpretation, must be evaluated. This requires considering, and often re-assessing, the choices and assumptions made in both the input and implementation stages.
.. image:: InversionWorkflow-PreSimPEG.png
:width: 400 px
:alt: Components
:align: center
A Comprehensive Framework
-------------------------
There are an overwhelming amount of choices to be made as one works through the forward modeling and inversion process (see figure above). As a result, software implementations of this workflow often become complex and highly interdependent, making it difficult to interact with and to ask other scientists to pick up and change. Our approach to handling this complexity is to propose a framework, (see below), that compartmentalizes the implementation of inversions into various units. We present it in this specific modular style, as each unit contains a targeted subset of choices crucial to the inversion process.
.. image:: InversionWorkflow.png
:width: 400 px
:alt: Framework
:align: center
The process of obtaining an acceptable model from an inversion generally requires the geophysicist to perform several iterations of the inversion workflow, rethinking and redesigning each piece of the framework to ensure it is appropriate in the current context. Inversions are experimental and empirical by nature and our software package is designed to facilitate this iterative process. To accomplish this, we have divided the inversion methodology into eight major components (See figure above). The (:class:`SimPEG.Mesh.BaseMesh`) class handles the discretization of the earth and also provides numerical operators. The forward simulation is split into two classes, the (:class:`SimPEG.Survey.BaseSurvey`) and the (:class:`SimPEG.Problem.BaseProblem`). The (:class:`SimPEG.Survey.BaseSurvey`) class handles the geometry of a geophysical problem as well as sources. The (:class:`SimPEG.Problem.BaseProblem`) class handles the simulation of the physics for the geophysical problem of interest. Although created independently, these two classes must be paired to form all of the components necessary for a geophysical forward simulation and calculation of the sensitivity. The (:class:`SimPEG.Problem.BaseProblem`) creates geophysical fields given a source from the (:class:`SimPEG.Survey.BaseSurvey`). The (:class:`SimPEG.Survey.BaseSurvey`) interpolates these fields to the receiver locations and converts them to the appropriate data type, for example, by selecting only the measured components of the field. Each of these operations may have associated derivatives with respect to the model and the computed field; these are included in the calculation of the sensitivity. For the inversion, a (:class:`SimPEG.DataMisfit.BaseDataMisfit`) is chosen to capture the goodness of fit of the predicted data and a (:class:`SimPEG.Regularization.BaseRegularization`) is chosen to handle the non-uniqueness. These inversion elements and an Optimization routine are combined into an inverse problem class (:class:`SimPEG.InvProblem.BaseInvProblem`). (:class:`SimPEG.InvProblem.BaseInvProblem`) is the mathematical statement that will be numerically solved by running an Inversion. The (:class:`SimPEG.Inversion.BaseInversion`) class handles organization and dispatch of directives between all of the various pieces of the framework.
Explaining The Big Picture
==========================
The arrows in the figure above indicate what each class takes as a primary argument. For example, both the (:class:`SimPEG.Problem.BaseProblem`) and (:class:`SimPEG.Regularization.BaseRegularization`) classes take a (:class:`SimPEG.Mesh.BaseMesh`) class as an argument. The diagram does not show class inheritance, as each of the base classes outlined have many subtypes that can be interchanged. The (:class:`SimPEG.Mesh.BaseMesh`) class, for example, could be a regular Cartesian mesh (:class:`SimPEG.Mesh.TensorMesh`) or a cylindrical coordinate mesh (:class:`SimPEG.Mesh.CylMesh`), which have many properties in common. These common features, such as both meshes being created from tensor products, can be exploited through inheritance of base classes, and differences can be expressed through subtype polymorphism. Please look at the documentation here for more in-depth information.
.. include:: ../CITATION.rst
Authors
-------
.. include:: ../AUTHORS.rst
License
-------
.. include:: ../LICENSE
docs/api_installing.rst
+2 -2
View File
@@ -1,7 +1,7 @@
.. _api_installing:
Installation
************
Getting Started with SimPEG
***************************
Dependencies
============
docs/api_license.rst
-17
View File
@@ -1,17 +0,0 @@
.. _api_license:
License
*******
.. include:: ../LICENSE
Authors
*******
.. include:: ../AUTHORS.rst
Projects Using SimPEG
*********************
.. include:: ../PROJECTS.rst
docs/em/api_Utils.rst
+3 -4
View File
@@ -7,7 +7,7 @@ sources, and analytic functions.
Analytic Functions - Time
=========================
.. automodule:: SimPEG.EM.Utils.Ana.TEM
.. automodule:: SimPEG.EM.Analytics.TDEM
:show-inheritance:
:members:
:undoc-members:
@@ -17,7 +17,7 @@ Analytic Functions - Time
Analytic Functions - Frequency
==============================
.. automodule:: SimPEG.EM.Utils.Ana.FEM
.. automodule:: SimPEG.EM.Analytics.FDEM
:show-inheritance:
:members:
:undoc-members:
@@ -27,8 +27,7 @@ Analytic Functions - Frequency
Sources
=======
.. automodule:: SimPEG.EM.Utils.Sources.magneticDipole
.. autoclass:: SimPEG.EM.FDEM.SrcFDEM.MagDipole
:show-inheritance:
:members:
:undoc-members:
:inherited-members:
docs/em/index.rst
+2 -3
View File
@@ -9,7 +9,7 @@ Time Domian Electromagnetics
----------------------------
.. toctree::
:maxdepth: 2
:maxdepth: 2
api_TDEM_derivation
@@ -18,7 +18,7 @@ Code for Time Domian Electromagnetics
-------------------------------------
.. toctree::
:maxdepth: 2
:maxdepth: 2
api_TDEM
@@ -28,7 +28,6 @@ Frequency Domian Electromagnetics
.. toctree::
:maxdepth: 2
api_ForwardProblem
api_FDEM
docs/index.rst
+67 -84
View File
@@ -1,25 +1,38 @@
.. image:: https://raw.github.com/simpeg/simpeg/master/docs/simpeg-logo.png
:alt: SimPEG Logo
SimPEG Documentation
********************
.. image:: simpeg-logo.png
:width: 300 px
:alt: SimPEG
:align: center
.. image:: https://img.shields.io/pypi/v/SimPEG.svg
:target: https://crate.io/packages/SimPEG/
:alt: Latest PyPI version
SimPEG: Simulation and Parameter Estimation in Geophysics
.. image:: https://img.shields.io/pypi/dm/SimPEG.svg
:target: https://crate.io/packages/SimPEG/
:alt: Number of PyPI downloads
SimPEG is a framework and a collection of tools that aid in the development of
large-scale geophysical inversion codes.
The vision is to create a modular and extensible package for
finite volume simulation and parameter estimation with
applications to geophysical imaging and subsurface flow. To enable
these goals, this package has the following features:
.. image:: https://img.shields.io/badge/license-MIT-blue.svg
:target: https://github.com/simpeg/simpeg/blob/master/LICENSE
:alt: BSD 3 clause license.
- is modular with respect to ... everything!
- is built with the (large-scale) inverse problem in mind
- provides a framework for geophysical and hydrogeologic problems
- supports 1D, 2D and 3D problems
- provides a set of commonly used visualization utilities
.. image:: https://img.shields.io/travis/simpeg/simpeg.svg
:target: https://travis-ci.org/simpeg/simpeg
:alt: Travis CI build status
.. image:: https://img.shields.io/coveralls/simpeg/simpeg.svg
:target: https://coveralls.io/r/simpeg/simpeg?branch=master
:alt: Coverage status
Simulation and Parameter Estimation in Geophysics - A python package for simulation and gradient based parameter estimation in the context of geophysical applications.
Our vision is to create a package for finite volume simulation with applications to geophysical imaging and subsurface flow. To enable the understanding of the many different components, this package has the following features:
* modular with respect to the spacial discretization, optimization routine, and geophysical problem
* built with the inverse problem in mind
* provides a framework for geophysical and hydrogeologic problems
* supports 1D, 2D and 3D problems
* designed for large-scale inversions
About SimPEG
@@ -29,56 +42,8 @@ About SimPEG
:maxdepth: 2
api_bigPicture
api_license
Getting Started with SimPEG
***************************
.. toctree::
:maxdepth: 2
api_installing
Discretization
**************
.. toctree::
:maxdepth: 3
api_Mesh
api_DiffOps
api_InnerProducts
Forward Problems
****************
.. toctree::
:maxdepth: 2
api_ForwardProblem
Inversion
*********
.. toctree::
:maxdepth: 3
api_DataMisfit
api_Inverse
Utility Codes
*************
.. toctree::
:maxdepth: 2
api_Solver
api_Maps
api_Utils
api_Tests
Packages
********
@@ -88,20 +53,46 @@ Packages
em/index
flow/index
Developer's Documentation
*************************
Examples
********
* Travis-CI Testing
.. image:: https://travis-ci.org/simpeg/simpeg.svg?branch=master
:target: https://travis-ci.org/simpeg/simpeg
:alt: Master Branch
:align: center
.. toctree::
:maxdepth: 2
* Coveralls Testing
.. image:: https://coveralls.io/repos/simpeg/simpeg/badge.png?branch=master
:target: https://coveralls.io/r/simpeg/simpeg?branch=master
:alt: Coveralls
:align: center
api_Examples
Finite Volume
*************
.. toctree::
:maxdepth: 3
api_FiniteVolume
Forward Problems
****************
.. toctree::
:maxdepth: 3
api_ForwardProblem
Inversion Components
********************
.. toctree::
:maxdepth: 3
api_InversionComponents
Utility Codes
*************
.. toctree::
:maxdepth: 3
api_Utilities
Project Index & Search
@@ -110,11 +101,3 @@ Project Index & Search
* :ref:`genindex`
* :ref:`modindex`
* :ref:`search`
Examples
********
.. toctree::
:maxdepth: 2
api_Examples