Feature/breast data

[ejdam87]

Summary by CodeRabbit

• New Features
  • Added support for B20-24 and MammaPrint breast cancer datasets across exploration and preprocessing workflows.
  • Added configurations for train, validation, test, tile masking, filtering, and embedding generation at multiple resolutions.
  • Added dataset metadata, artifact locations, stain statistics, and validation split handling.
• Improvements
  • Preserved carcinoma and epithelium measurements during tile filtering.
  • Increased embedding processing capacity per output file.
  • Added explicit zero-dropout settings across supported models.
• Documentation
  • Updated project naming and description to reflect broader carcinoma detection capabilities.

[coderabbitai]

📝 Walkthrough

Walkthrough

This change adds breast MMCI dataset configurations and exploration adapters for B20-24 and MammaPrint data. It wires preprocessing experiments, updates tile filtering and embedding settings, sets model dropout values, and renames the project for carcinoma classification.

Changes

Breast data pipeline

Layer / File(s)	Summary
Breast dataset contracts configs/data/breast/...	Defines shared breast data settings, stain references, filtering thresholds, and dataset statistics.
Dataset artifact configurations configs/data/breast/mmci/...	Adds MMCI B20-24 full, training, validation, and MammaPrint test artifact configurations.
Exploration adapters configs/exploration/breast/..., exploration/b20_24.py, exploration/mammaprint_test.py	Adds Hydra configurations and MLflow adapters for metadata normalization, label correction, tiling repair, split creation, and artifact logging.
Tile preprocessing configuration and filtering configs/experiment/preprocessing/filter_tiles/..., configs/experiment/preprocessing/tile_masks/..., preprocessing/tiling_v2/filter_tiles.py, configs/preprocessing/tile_embeddings_v2.yaml	Adds tile preprocessing configurations, excludes carcinoma and epithelium columns from threshold filtering, and raises rows_per_file to 150000.
Embedding preprocessing experiments configs/experiment/preprocessing/tile_embeddings_v2/...	Adds PGP and Virchow2 configurations for B20-24 and MammaPrint datasets.
Model and project metadata configs/model/*.yaml, pyproject.toml	Sets model dropout to 0.0 and updates project metadata for carcinoma classification.

Estimated code review effort: 3 (Moderate) | ~25 minutes

Sequence Diagram(s)

sequenceDiagram
  participant Hydra
  participant ExplorationAdapter
  participant MLflow
  participant MetadataAndTiling
  Hydra->>ExplorationAdapter: Load dataset configuration
  ExplorationAdapter->>MLflow: Download metadata and tiling artifacts
  ExplorationAdapter->>MetadataAndTiling: Normalize metadata and repair tiling datasets
  MetadataAndTiling->>MLflow: Log repaired datasets and metadata artifacts

Loading

Possibly related PRs

• RationAI/carcinoma-binary-classification-methods#23: Modifies the same tile-filtering logic for carcinoma and epithelium percentage columns.
• RationAI/carcinoma-binary-classification-methods#35: Uses related tile-filtering and dataset artifact configuration patterns.

Suggested reviewers: adames4, vejtek

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly identifies the main change: adding breast cancer data configurations, exploration workflows, and preprocessing support.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches 💡 1

📝 Generate docstrings 💡

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feature/breast-data

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[gemini-code-assist]

Caution

The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased.

[ejdam87]

@coderabbitai review

[coderabbitai]

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@exploration/mammaprint_test.py`:
- Around line 37-48: Normalize the MammaPrint metadata before writing
mammaprint_test_metadata.csv so it exposes slide_path as required by the
pipeline, preprocessors, and v2 tiling. Update the metadata setup associated
with config.original_metadata, or apply the existing B20-24 adapter’s
rename-and-drop contract to df before repair_and_log and CSV logging; preserve
all other metadata behavior.

🪄 Autofix

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 37ee6e1b-6ca1-4362-a81e-c5fc17cdffd7

📥 Commits

Reviewing files that changed from the base of the PR and between 1d81141 and 358dec3.

⛔ Files ignored due to path filters (1)

• uv.lock is excluded by !**/*.lock

📒 Files selected for processing (32)

• configs/data/breast/base.yaml
• configs/data/breast/mmci/b20_24_base.yaml
• configs/data/breast/mmci/b20_24_full.yaml
• configs/data/breast/mmci/b20_24_train.yaml
• configs/data/breast/mmci/b20_24_val.yaml
• configs/data/breast/mmci/base.yaml
• configs/data/breast/mmci/mammaprint_test.yaml
• configs/data/breast/mmci/stats.yaml
• configs/experiment/preprocessing/filter_tiles/b20_24_train.yaml
• configs/experiment/preprocessing/filter_tiles/b20_24_val.yaml
• configs/experiment/preprocessing/filter_tiles/mammaprint_test.yaml
• configs/experiment/preprocessing/tile_embeddings_v2/pgp/b20_24_train.yaml
• configs/experiment/preprocessing/tile_embeddings_v2/pgp/b20_24_val.yaml
• configs/experiment/preprocessing/tile_embeddings_v2/pgp/mammaprint_test.yaml
• configs/experiment/preprocessing/tile_embeddings_v2/virchow2/b20_24_train.yaml
• configs/experiment/preprocessing/tile_embeddings_v2/virchow2/b20_24_val.yaml
• configs/experiment/preprocessing/tile_embeddings_v2/virchow2/mammaprint_test.yaml
• configs/experiment/preprocessing/tile_masks/b20_24_val_224.yaml
• configs/experiment/preprocessing/tile_masks/b20_24_val_512.yaml
• configs/experiment/preprocessing/tile_masks/mammaprint_test_512.yaml
• configs/exploration/breast/b20_24.yaml
• configs/exploration/breast/base.yaml
• configs/exploration/breast/mammaprint_test.yaml
• configs/model/pgp_full_model.yaml
• configs/model/pgp_model.yaml
• configs/model/virchow2_full_model.yaml
• configs/model/virchow2_model.yaml
• configs/preprocessing/tile_embeddings_v2.yaml
• exploration/b20_24.py
• exploration/mammaprint_test.py
• preprocessing/tiling_v2/filter_tiles.py
• pyproject.toml