Feat separate winnow predict filtered dataset outputs

[JemmaLDaniel]

Feature: Separate metadata and features from predictions and FDR metrics in CLI output

Summary

Implements a feature to separate outputs from winnow predict into two CSV files: one containing metadata and features, and another containing predictions and FDR metrics. This improves usability for downstream analysis by allowing users to work with different output categories independently.

Changes

Implementation (winnow/scripts/main.py)

• Changed parameter: Renamed --output-path to --output-folder to better reflect that multiple files are written
• Added output separation logic: The predict command now:
  1. Separates columns into two groups:
     • Metadata/features: All original metadata and feature columns
     • Predictions/FDR metrics: spectrum_id, confidence scores, prediction, psm_fdr, psm_q_value, psm_pep (if winnow method), and sequence (if available)
  2. Writes two separate CSV files:
     • metadata.csv: Contains all metadata and feature columns
     • preds_and_fdr_metrics.csv: Contains predictions and error metrics

Documentation Updates (docs/cli.md)

• Updated winnow predict output section to describe the two-file output structure
• Fixed parameter name from --output-path to --output-folder in all examples
• Added detailed descriptions of columns in each output file
• Updated workflow examples to use the correct parameter names

Files Changed

• winnow/scripts/main.py: Implemented output separation logic and changed parameter name
• docs/cli.md: Updated documentation to reflect the new two-file output structure

Breaking Changes

This is a breaking change for users of the CLI:

• The --output-path parameter has been renamed to --output-folder
• Instead of a single CSV file, two files are now created: metadata.csv and preds_and_fdr_metrics.csv
• Users will need to update scripts/workflows to read from the new file structure

[BioGeek]

Looks good to me!