[docker] Mirror Functionality of CI scripts & Fix docs (#10349)

* first-pass

* add back build examples

* remove unnecessary test

* add gcc and more formatting

* doc fixing

* small fixes

docker/examples/Dockerfile

@@ -1,17 +1,36 @@
 # The examples Docker image adds dependencies needed to run the examples
 
-FROM ray-project/deploy
+FROM rayproject/ray:latest
 
-# This updates numpy to 1.14 and mutes errors from other libraries
-RUN conda install -y numpy
 # Needed to run Tune example with a 'plot' call - which does not actually render a plot, but throws an error.
-RUN apt-get install -y zlib1g-dev libgl1-mesa-dev libgtk2.0-dev
-RUN pip install -U pip
-RUN pip install gym[atari] opencv-python-headless==4.3.0.36 tensorflow lz4 pytest-timeout smart_open tensorflow_probability dm_tree
-RUN pip install -U h5py  # Mutes FutureWarnings
-RUN pip install --upgrade bayesian-optimization hyperopt
-RUN pip install ConfigSpace==0.4.10
-RUN pip install --upgrade sigopt nevergrad scikit-optimize hpbandster lightgbm xgboost torch torchvision tensorboardX dragonfly-opt zoopt
-RUN pip install -U tabulate mlflow
-RUN pip install -U pytest-remotedata>=0.3.1
-RUN pip install -U matplotlib jupyter pandas
+RUN apt-get update && install -y zlib1g-dev libgl1-mesa-dev libgtk2.0-dev && apt-get clean
+RUN pip install --no-cache-dir -U pip \
+    gym[atari] \
+    opencv-python-headless==4.3.0.36 \
+    tensorflow \
+    lz4 \
+    pytest-timeout \
+    smart_open \
+    tensorflow_probability \
+    dm_tree \
+    h5py   # Mutes FutureWarnings \
+    bayesian-optimization \
+    hyperopt \
+    ConfigSpace==0.4.10 \
+    sigopt \ 
+    nevergrad \
+    scikit-optimize \
+    hpbandster \
+    lightgbm \
+    xgboost \
+    torch \
+    torchvision \
+    tensorboardX \
+    dragonfly-opt \
+    zoopt \
+    tabulate \ 
+    mlflow \
+    pytest-remotedata>=0.3.1 \
+    matplotlib \
+    jupyter \
+    pandas