Usage: python tracery_markov_model.py [-h] [-n SIZE] [-l] path/to/corpus/file
Optional arguments:
-h,--help— Show help message-n SIZE,--ngram SIZE— Specify n-gram size (default: 3)-l,--lowercase— Convert all input to lowercase
The input corpus should be a text file containing statements separated by line breaks (preferably in English since it's the only language I've tested so far). The output JSON file will appear in a folder named output.