smrutil / forestdb.org

A repository for generative models

Forest

Forest is a collaborative repository for generative models.

Adding models

To add a new model to Forest, simply create a markdown file in the models directory. You can do this directly on github by following this link: add model.

For examples of the model file format, see:

• example.md is a model that is not shown on the front page.
• arithmetic.md is a model that is shown on the front page.

For models that are shown on the front page, use additional header entries:

model-status: code
model-category: Concept Learning
model-tags: concepts, program induction

Currently, the following status codes are available:

code
link
code-fail
stub

The following categories are available:

Concept Learning
Reasoning about Reasoning
Machine Learning
Nonparametric Models
Bayesian Data Analysis    
Undirected Constraints
Inverse Dynamics
PPAML Challenge Problems
Miscellaneous

Tags can be arbitrary words or phrases separated by commas.

Updating webchurch

Forest uses webchurch and webppl to run probabilistic programs. To update webchurch, follow these steps:

1. Clone and build the most recent version of the webchurch repository, following the instructions here:

    git clone https://github.com/probmods/webchurch.git
    cd webchurch
    git init (only if you are on Windows)
    git submodule update --init --recursive
    npm install
    ./compile.sh
   

2. Clone the most recent version of the Forest repository:

    git clone https://github.com/forestdb/forestdb.org.git
   

3. Copy the files in webchurch/online/ to forestdb.org/assets/webchurch/.

4. Run a local webserver in the Forest directory and go to 0.0.0.0:4000 to check that the changes did not break functionality:

    jekyll serve --watch
   

5. If code boxes run as expected, great! Commit and push. If not, you may need to adjust assets/js/custom.js in the Forest directory.