(cf. https://github.com/yoeo/guesslang/releases/tag/v2.0.0a1)
- Supports 30 languages (20 languages supported by previous versions). Including Powershell and Batch
- Robust and scalable training workflow, using Tensorflow dataset API https://www.tensorflow.org/api_docs/python/tf/data/Dataset
- Way simpler yet as performant feature engineering
- More concise documentation
- Simple check on detection probabilities to avoid classifying plain text as source code
- Exposes the detection probabilities with guess.probabilities(source_code) method.
- Updated Tensorflow to the latest version 2.2.0
- Use up to date Tensorflow canned classifier
- guess.language_name(source_code) now identifies empty content
- Guesslang dataset creation moved to a new dedicated repository https://github.com/yoeo/guesslangtools