/code_notcode

Ever wanted to parse your developer logbook but wanted to separate code from your other remarks?

Primary LanguagePython

Ever wanted to parse your developer logbook but wanted to separate code from your other remarks?

Bayesian classifier. Pretrained on Python code and SQL.

docker run --rm -v $(pwd):/code -it nltk bash docker run --rm -v $(pwd)/pickled_models:/code/pickled_models -v $(pwd)/landing:/landing -v $(pwd)/processed:/processed nltk

prototype TODO: create docker environment with nltk installed ✓ paste some python code in ./code ✓ paste some HQL in ./code use some nltk demo dataset for ./notcode ✓ write training code ✓ pickle the model ✓ write classifier ✓ make it interface with a landing zone and output zone