Repository navigation
Conversation
|
Important Draft PR not reviewedDraft PRs are not automatically reviewed by default.
To automatically review draft PRs, update your CodeRabbit configuration: reviews:
auto_review:
drafts: true✨ Finishing Touches🧪 Generate unit tests (beta)
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. Comment |
Summary of ChangesHello, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request establishes the core configuration infrastructure for a machine learning pipeline designed to automate the histological grading of ulcerative colitis. It introduces a structured Hydra configuration system to manage diverse experimental setups, foundation model integrations, and dataset processing pipelines. Additionally, it provides detailed documentation to facilitate reproducibility and understanding of the project's methodology. Highlights
Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize the Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counterproductive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here. Footnotes
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request sets up the machine learning pipeline and Hydra configurations for automated grading of ulcerative colitis from H&E whole slide images, including a detailed README. Feedback on these changes highlights an invalid Ruff rule code (TC002 instead of TCH002) in .ruff.toml that would cause configuration parsing errors. Additionally, typos were found in several knl_patos embedding configuration files referencing a non-existent default path, and the README table is missing the configs/ prefix for the listed configuration paths.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
| "D107", # missing docstring in __init__ | ||
| "N812", # lowercase imported as non lowercase | ||
| "TCH002", # move third-party into a type-checking block | ||
| "TC002", # move third-party into a type-checking block |
There was a problem hiding this comment.
The rule code TC002 is invalid in Ruff. The correct rule code for ignoring third-party imports in type-checking blocks is TCH002 (as part of the TCH / flake8-type-checking ruleset). Changing this to TC002 will cause Ruff to fail with a configuration parsing error.
| "TC002", # move third-party into a type-checking block | |
| "TCH002", # move third-party into a type-checking block |
| @@ -0,0 +1,8 @@ | |||
| defaults: | |||
| - /dataset/tiled/knl_patos/0_224@_here_ | |||
There was a problem hiding this comment.
The default config path /dataset/tiled/knl_patos/0_224 does not exist. The repository contains 1_224.yaml under configs/dataset/tiled/knl_patos/. This typo will cause Hydra to fail when composing the configuration.
- /dataset/tiled/knl_patos/1_224@_here_References
- Hydra Configs (configs/): Reproducibility is paramount. Ensure configurations are correct and complete. (link)
| @@ -0,0 +1,8 @@ | |||
| defaults: | |||
| - /dataset/tiled/knl_patos/0_224@_here_ | |||
There was a problem hiding this comment.
The default config path /dataset/tiled/knl_patos/0_224 does not exist. The repository contains 1_224.yaml under configs/dataset/tiled/knl_patos/. This typo will cause Hydra to fail when composing the configuration.
- /dataset/tiled/knl_patos/1_224@_here_References
- Hydra Configs (configs/): Reproducibility is paramount. Ensure configurations are correct and complete. (link)
| @@ -0,0 +1,8 @@ | |||
| defaults: | |||
| - /dataset/tiled/knl_patos/0_224@_here_ | |||
There was a problem hiding this comment.
The default config path /dataset/tiled/knl_patos/0_224 does not exist. The repository contains 1_224.yaml under configs/dataset/tiled/knl_patos/. This typo will cause Hydra to fail when composing the configuration.
- /dataset/tiled/knl_patos/1_224@_here_References
- Hydra Configs (configs/): Reproducibility is paramount. Ensure configurations are correct and complete. (link)
| | I | Baseline Classification | `experiment/ml/experiment_i_baseline_classification` | VGG16 tile-level classifier trained on raw pixel tiles | | ||
| | II | Ordered Regression | `experiment/ml/experiment_ii_ordered_regression` | Cumulative link loss for NHI ordinal structure | | ||
| | III | Pathology Foundation Models | `experiment/ml/experiment_iii_pathology_foundation_model` | Linear probe on frozen prov-gigapath embeddings | | ||
| | IV | MIL & Hierarchical Modeling | `experiment/ml/experiment_iv_mil_and_hierarchical_modeling` | Attention MIL; three-task hierarchical decomposition | | ||
| | V | More Foundation Models | `experiment/ml/experiment_v_more_foundation_models` | MIL sweep over all foundation models | | ||
| | VI | Neutrophil Detection | `experiment/ml/experiment_vi_neutrophil_detection` | Binary tile-level neutrophil classifier | | ||
| | VII | Ensembling | `experiment/postprocessing/experiment_vii_ensembling` | Soft majority vote vs. hierarchical routing | | ||
| | VIII | Markov Chain Confidence | `experiment/postprocessing/experiment_viii_markov_chain_model_aggregation` | Absorption distribution confidence (entropy / herfindahl / std) | | ||
| | — | Final Model | `experiment/ml/final` | Attention MIL (virchow2) trained on all three institutions | |
There was a problem hiding this comment.
The configuration paths listed in the table are missing the configs/ prefix. According to the repository structure, the experiment configurations are located under configs/experiment/.... Updating these paths will make it easier for users to locate the files in the repository.
| | I | Baseline Classification | `experiment/ml/experiment_i_baseline_classification` | VGG16 tile-level classifier trained on raw pixel tiles | | |
| | II | Ordered Regression | `experiment/ml/experiment_ii_ordered_regression` | Cumulative link loss for NHI ordinal structure | | |
| | III | Pathology Foundation Models | `experiment/ml/experiment_iii_pathology_foundation_model` | Linear probe on frozen prov-gigapath embeddings | | |
| | IV | MIL & Hierarchical Modeling | `experiment/ml/experiment_iv_mil_and_hierarchical_modeling` | Attention MIL; three-task hierarchical decomposition | | |
| | V | More Foundation Models | `experiment/ml/experiment_v_more_foundation_models` | MIL sweep over all foundation models | | |
| | VI | Neutrophil Detection | `experiment/ml/experiment_vi_neutrophil_detection` | Binary tile-level neutrophil classifier | | |
| | VII | Ensembling | `experiment/postprocessing/experiment_vii_ensembling` | Soft majority vote vs. hierarchical routing | | |
| | VIII | Markov Chain Confidence | `experiment/postprocessing/experiment_viii_markov_chain_model_aggregation` | Absorption distribution confidence (entropy / herfindahl / std) | | |
| | — | Final Model | `experiment/ml/final` | Attention MIL (virchow2) trained on all three institutions | | |
| | I | Baseline Classification | `configs/experiment/ml/experiment_i_baseline_classification` | VGG16 tile-level classifier trained on raw pixel tiles | | |
| | II | Ordered Regression | `configs/experiment/ml/experiment_ii_ordered_regression` | Cumulative link loss for NHI ordinal structure | | |
| | III | Pathology Foundation Models | `configs/experiment/ml/experiment_iii_pathology_foundation_model` | Linear probe on frozen prov-gigapath embeddings | | |
| | IV | MIL & Hierarchical Modeling | `configs/experiment/ml/experiment_iv_mil_and_hierarchical_modeling` | Attention MIL; three-task hierarchical decomposition | | |
| | V | More Foundation Models | `configs/experiment/ml/experiment_v_more_foundation_models` | MIL sweep over all foundation models | | |
| | VI | Neutrophil Detection | `configs/experiment/ml/experiment_vi_neutrophil_detection` | Binary tile-level neutrophil classifier | | |
| | VII | Ensembling | `configs/experiment/postprocessing/experiment_vii_ensembling` | Soft majority vote vs. hierarchical routing | | |
| | VIII | Markov Chain Confidence | `configs/experiment/postprocessing/experiment_viii_markov_chain_model_aggregation` | Absorption distribution confidence (entropy / herfindahl / std) | | |
| | — | Final Model | `configs/experiment/ml/final` | Attention MIL (virchow2) trained on all three institutions | |
* feat: add ikem validation configuration * fix: update institution name in ikem validation configuration * feat: add drop_missing parameter to create_dataset function * fix: correct typo in dataset configuration parameter * fix: update condition for institution check in create_dataset function * feat: add ikem validation configuration file * feat: add ikem validation configuration and update storage parameters in quality control script * feat: implement filtering of dataset by QC errors using qc_errors_uri * fix: improve line handling in filter_dataset_by_qc_errors function * fix: update qc_main to handle qc_errors.log more efficiently * feat: enhance QC handling in tiling process with NaN filling and mask availability checks * fix: update qc_mask URI in ikem validation configuration * feat: add ikem validation configuration for dataset processing
…odel requirements
…rkov chain confidence
1 participant
No description provided.