Fix to ensure Factor states are aligned when adding/multiplying/etc.

Dockerfile

@@ -9,25 +9,28 @@ ENV USER ${NB_USER}
 ENV NB_UID ${NB_UID}
 ENV HOME /home/${NB_USER}
 
+USER root
+
 RUN adduser --disabled-password \
     --gecos "Default user" \
     --uid ${NB_UID} \
     ${NB_USER}
 
-# Copy package
+# Install widget first, so the (editable) install of 'core' overwrites
+# the PyPI version of thomas-core.
+RUN pip install thomas-jupyter-widget
+RUN jupyter labextension install --minimize=False thomas-jupyter-widget
+
+# Copy & install package
 COPY . /usr/local/python/thomas-core/
+WORKDIR /usr/local/python/
+RUN pip install ./thomas-core
 
-# Make sure the contents of our repo are also in ${HOME}
-# COPY . ${HOME}
 
-USER root
 RUN chown -R ${NB_UID} ${HOME}
 RUN chown -R ${NB_UID} /usr/local/python/thomas-core/
 USER ${NB_USER}
 
-WORKDIR /usr/local/python/
-RUN pip install ./thomas-core
-
 # JupyterLab runs on port 8888
 EXPOSE 8888
 

README.md

@@ -5,7 +5,7 @@
 # Thomas
 Very simple (almost naive ;-) bayesian network implementation.
 
-Example (module `thomas.core.examples`) contains examples from the book "Probabilistic Graphical Models: Principles and Techniques" from Koller and Friedman ([PGM Stanford](http://pgm.stanford.edu)) and from the lecture by [Adnan Darwiche](http://web.cs.ucla.edu/~darwiche/) on YouTube:
+Contains examples (`thomas.core.examples`) from the book "Probabilistic Graphical Models: Principles and Techniques" from Koller and Friedman ([PGM Stanford](http://pgm.stanford.edu)) and from the lecture by [Adnan Darwiche](http://web.cs.ucla.edu/~darwiche/) on YouTube:
 * [6a. Inference by Variable Elimination I (Chapter 6)](https://www.youtube.com/watch?v=7oRReD_ayWo).
 * [6b. Inference by Variable Elimination II (Chapter 6)](https://www.youtube.com/watch?v=QSSmx1ndUvg).