Repository navigation
fix(pathway): resolve remaining PR611 review findings - #1624
Open
parulsarma-a wants to merge 16 commits into
Open
parulsarma-a wants to merge 16 commits into
parulsarma-a wants to merge 16 commits into
Conversation
A lightweight environment for pathway inference on synthetic omics data.
An agent inspects a dataset, runs differential expression and pathway
enrichment analyses, then submits a hypothesis about which signaling
pathway is activated.
Designed as a starter biology-shaped environment that validates OpenEnv
mechanics (reset/step/state, client/server, Docker packaging) without
requiring heavy genomics dependencies. Future versions will swap
deterministic toy outputs for real DE/enrichment tools (PyDESeq2, GSEApy).
Actions: inspect_dataset, run_differential_expression,
run_pathway_enrichment, submit_answer
Made-with: Cursor
…s and tests - PyDESeq2 + Fisher ORA, compare pathways, expert budget, HTML traces - Gradio Visualization tab with episode state, tabbed results, default /web UI - Case JSON fixtures (toy, strict no-default, legacy, detailed stress) - Scripts for curriculum, benchmark, fake RNA-seq case builder - Tests for pathway env; app enables ENABLE_WEB_INTERFACE when unset Made-with: Cursor
…gent workflow docs - New action: summarize groups/samples, validate DGE contrast without DESeq2 - State: validated_reference/alternate, design_understood; DE uses same resolution order - Gradio Pathway lab button; README agent workflow (groups → DGE → pathways) - Tests for design summary and DE parity with validated contrast Made-with: Cursor
Adds reproducible GEO study runners/cases and benchmark documentation, plus advisor-ready HTML/PPTX decks and supporting scripts. Made-with: Cursor
Separate author-table loading errors from counts errors without fallback. Validate gzip downloads before atomic publication and request curl failure handling and retries. Accept plain GTF annotations for Salmon using the same extension convention as STAR. Regression tests reproduce all three Bugbot findings on 31d296c (18 failed, 51 passed). Focused Pathway suite: 102 passed, 5 skipped. Scientific mapping and State/oracle merge holds remain unchanged. Refs openenv#611.
Co-authored-by: Cursor <cursoragent@cursor.com>
Author
|
@burtenshaw thanks again for the guidance on #611. This follow-up PR is now ready for review and includes the remaining fixes you called out:
If useful, I can split this further, but I kept it scoped to the unresolved findings. |
1 similar comment
Author
|
@burtenshaw thanks again for the guidance on #611. This follow-up PR is now ready for review and includes the remaining fixes you called out:
If useful, I can split this further, but I kept it scoped to the unresolved findings. |
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Follow-up to closed PR #611, addressing remaining unresolved review findings.
What changed
PathwayStateno longer includestrue_pathwaygene_id -> gene_namemapping inrun_true_counts_pipeline.pytrue_pathwayin state payloadValidation
PYTHONPATH=src:envs uv run pytest tests/envs/test_pathway_counts_pipeline.py -q(19 passed)PYTHONPATH=src:envs uv run pytest tests/envs/test_pathway_analysis_env.py -k "state_does_not_expose_true_pathway or legacy_fixture" -q(2 passed)Branch
parulsarma-a:pr611-followup-fixesNote
Low Risk
Changes are scoped to a new environment package and benchmark scripts; main risks are optional external GEO/SRA downloads and heavy Docker bench runs, not core platform auth or data paths.
Overview
Adds a new
pathway_analysis_envOpenEnv for RNA-seq-style workflows: agents inspect design, run PyDESeq2 DE and Fisher ORA / Enrichr enrichment, then submit a pathway hypothesis, with synthetic and GEO-linked case JSON, an HTTP client/server (openenv.yaml), default Gradio UI, and published docs/site entries.Ships a
bench/toolchain (optional Docker image with STAR/Salmon/SRA tools, hg38/GENCODE downloaders, true-counts pipeline, per-study GEO runners, and a keyword-based agent eval harness) plus benchmark write-ups and failure-code docs.Evaluation hardening (PR #611 follow-up): serialized
PathwayStateno longer includestrue_pathway(grading stays server-side on submit), STAR/featureCounts exports map Ensemblgene_id→ gene symbols with duplicate aggregation for ORA-friendly matrices, and new regressions lock in those behaviors alongside FASTQ/layout pipeline edge cases.Also adds a root pointer to canonical
GEO_BENCHMARK_STUDIESdocumentation under the env package.Reviewed by Cursor Bugbot for commit 8e6af36. Bugbot is set up for automated code reviews on this repo. Configure here.