v2.0.0
Additions:
- Added setup.py for easy installation
- Added the CAPICE training scripts, so users are now able to create new CAPICE like models.
- Added config.cfg for additional setup.
General improvements:
- Changed code structure for be more object orientated
Command line interface:
- Changed input argument --input_path to -i
- Removed input argument --model_path
- Made output argument --prediction_savepath optional with -o
- Added flags for verbose (-v), force (-f) and train (--train)
Output:
- Changed output of CAPICE to no longer remove duplicate entries upon 'chrom-pos-ref-alt'
- Exposed transcript identifier
- Removed unused columns (prediction and combined_prediction)
Bugfixes:
- #16 Apply command-line argument conventions
- #17 Expose transcript identifier for predicted score
- #18 Add tests for Python code
- #20 Add training scripts to Github to enable reproducing the model
- #21 Log warnings for variants that can't be processed
- #24 'there shouldn't be any nulls' and 'feature from the model not in data' log messages
- #28 PIP install pandas==1.0.2 on Python 3.9+ fails to install