Use env.txt to replicate the Anaconda environment using:

$ conda create --name <env_name> --file env.txt

Once you activate the environment, you may want to (re)generate some of the synthetic datasets used. Navigate to the datasets/ directory and use the bash scripts within to do this. In each dataset directory there are two bash scripts:

• gen.sh - generates new data based on random variables
• regen.sh - regenerates data in the exact same state as was used for training

Look at the bash scripts and semgen/semgen.py --help for documentation of the commands used.

All scripts use Click to enable command line functionality. Use the --help parameter for usage instructions.

Entry points:

$ python train.py
$ python predict.py

GNU GPL v.3