Feature: Add semantic duplicate issue detection for GitHub issues #551
Open
vaibhavbaghel wants to merge 4 commits intoAOSSIE-Org:mainfrom
Open
Feature: Add semantic duplicate issue detection for GitHub issues #551vaibhavbaghel wants to merge 4 commits intoAOSSIE-Org:mainfrom
vaibhavbaghel wants to merge 4 commits intoAOSSIE-Org:mainfrom
Conversation
| No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (1)
🚧 Files skipped from review as they are similar to previous changes (1)
📝 WalkthroughWalkthroughThis PR adds an automated duplicate-issue detection pipeline: a new Python checker script, a pinned requirements file, and a GitHub Actions workflow that runs the checker on issue opened/edited events. Also includes a small UI padding tweak in a React component. Changes
Sequence DiagramsequenceDiagram actor GitHub participant Workflow as rgba(0,128,255,0.5) GitHub Actions participant Script as rgba(0,200,100,0.5) check-duplicates.py participant GitHubAPI as rgba(255,165,0,0.5) GitHub API participant Model as rgba(128,0,128,0.5) SentenceTransformer GitHub->>Workflow: issue opened/edited event Workflow->>Workflow: checkout, setup Python, install deps Workflow->>Script: run with ISSUE_NUMBER, TITLE, BODY, REPO, GITHUB_TOKEN Script->>GitHubAPI: GET /issues (paginated, exclude closed optional, max 500) GitHubAPI-->>Script: issues list Script->>Model: prepare texts (title + 500 chars), compute embeddings Model-->>Script: embeddings Script->>Script: compute cosine similarities, filter & sort by threshold Script->>Script: format markdown comment Script->>GitHubAPI: POST comment to current issue (if matches) GitHubAPI-->>Script: comment response Script-->>Workflow: exit with status Estimated code review effort🎯 3 (Moderate) | ⏱️ ~28 minutes Poem
🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ 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 |
vaibhavbaghel changed the title ![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 4
🧹 Nitpick comments (2)
.github/workflows/duplicate-detection.yml (1)
24-36: Cache the sentence-transformer model between runs.The workflow caches pip dependencies but not the Hugging Face model that
check-duplicates.pyloads at line 103 (SentenceTransformer("all-MiniLM-L6-v2")). On cold runner starts, this model will be redownloaded for every issue event, adding unnecessary latency and flakiness to triage. Add a caching step for the Hugging Face model directory usingactions/cacheandHF_HOMEenvironment variable to persist the model across workflow runs.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/duplicate-detection.yml around lines 24 - 36, The workflow currently installs pip deps but doesn't persist the Hugging Face model used in check-duplicates.py (SentenceTransformer("all-MiniLM-L6-v2")), so add an actions/cache step that caches the model directory and set HF_HOME to that cache path in the job environment; specifically, define HF_HOME (e.g., an env like HF_HOME: ${{ runner.temp }}/hf or similar), add an actions/cache entry that keys and restores the HF_HOME directory (or the Hugging Face cache path) before running pip and the python script, and ensure check-duplicates.py continues to load the model from SentenceTransformer using the HF_HOME-backed cache..github/scripts/requirements.txt (1)
1-5: Use pinned dependency versions for the duplicate-detection workflow.The workflow installs
.github/scripts/requirements.txtwith floating>=specs on every issue event, making duplicate detection behavior non-reproducible across runs and over time. The rootrequirements.txtalready demonstrates pinned versions are maintained for this project (e.g.,torch==2.5.1,scikit-learn==1.5.2). Update.github/scripts/requirements.txtto use pinned versions (==) from the root file or generate a proper lock file.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/scripts/requirements.txt around lines 1 - 5, Update the .github/scripts/requirements.txt to use pinned versions instead of floating >= specs so duplicate-detection runs are reproducible: replace each dependency (sentence-transformers, scikit-learn, requests, numpy, torch) with exact versions (==) matching the project's root requirements.txt or freeze them into a lock file and reference that; ensure the file uses the same pinned versions as root (e.g., the root's torch and scikit-learn pins) or generate a pinned requirements.lock and update the workflow to install that lock file.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed. Inline comments: In @.github/scripts/check-duplicates.py: - Around line 141-145: The try/except around calculate_similarity is swallowing errors and converting failures into an empty duplicate list; instead propagate the failure so the job fails or returns an explicit error state—remove the blanket except or re-raise the caught exception (or return a distinct error value) from where calculate_similarity is called, and update main() to treat that propagated exception/value as a fatal error (exit non‑zero or surface an error message) rather than treating it as “no duplicates found.” Ensure you reference calculate_similarity and main() when making the change so callers handle real failures instead of returning []. - Around line 189-205: The post_comment function currently always posts a new comment; change it to upsert a single detector comment by first listing existing comments for the issue (GET to f"{GITHUB_API_URL}/repos/{REPO}/issues/{ISSUE_NUMBER}/comments" with HEADERS), search for a prior bot comment using a unique marker string you add to the comment body (e.g., "<!-- duplicate-detector -->") or by matching the bot's user, and if found use PATCH to update that comment (PATCH to f"{GITHUB_API_URL}/repos/{REPO}/issues/comments/{comment_id}") otherwise POST as before; keep using HEADERS, timeout and response.raise_for_status() and return True/False as currently implemented in post_comment to preserve behavior. - Around line 21-27: The dynamic runtime pip-install fallback around the SentenceTransformer import must be removed so missing dependencies fail fast; delete the try/except block that catches ImportError and runs subprocess.check_call, and replace it with a single top-level import from sentence_transformers import SentenceTransformer so an ImportError surfaces immediately (this touches the import of SentenceTransformer in the script). In @.github/workflows/duplicate-detection.yml: - Around line 18-22: Update the GitHub Actions step that currently uses actions/setup-python@v4 to a supported major version (actions/setup-python@v5 or `@v6`) to avoid Node16 runner removal; specifically modify the step referencing "actions/setup-python@v4" to use at least "actions/setup-python@v5" (or "@v6"), keep the existing inputs (python-version: '3.11' and cache: 'pip') unchanged, then run the workflow to verify the runner no longer errors due to the deprecated Node runtime. --- Nitpick comments: In @.github/scripts/requirements.txt: - Around line 1-5: Update the .github/scripts/requirements.txt to use pinned versions instead of floating >= specs so duplicate-detection runs are reproducible: replace each dependency (sentence-transformers, scikit-learn, requests, numpy, torch) with exact versions (==) matching the project's root requirements.txt or freeze them into a lock file and reference that; ensure the file uses the same pinned versions as root (e.g., the root's torch and scikit-learn pins) or generate a pinned requirements.lock and update the workflow to install that lock file. In @.github/workflows/duplicate-detection.yml: - Around line 24-36: The workflow currently installs pip deps but doesn't persist the Hugging Face model used in check-duplicates.py (SentenceTransformer("all-MiniLM-L6-v2")), so add an actions/cache step that caches the model directory and set HF_HOME to that cache path in the job environment; specifically, define HF_HOME (e.g., an env like HF_HOME: ${{ runner.temp }}/hf or similar), add an actions/cache entry that keys and restores the HF_HOME directory (or the Hugging Face cache path) before running pip and the python script, and ensure check-duplicates.py continues to load the model from SentenceTransformer using the HF_HOME-backed cache. ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 376fea32-ba12-447c-bb31-636e7976cfd8
📒 Files selected for processing (4)
.github/scripts/check-duplicates.py.github/scripts/requirements.txt.github/workflows/duplicate-detection.ymleduaid_web/src/pages/Text_Input.jsx
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed. Inline comments: In @.github/scripts/check-duplicates.py: - Around line 62-80: The current pagination loop in check-duplicates.py swallows RequestException and breaks, returning a partial or empty issues list; change the except requests.exceptions.RequestException block to propagate the failure instead of breaking (e.g., re-raise the caught exception or call sys.exit(1)) so that callers of the function (the workflow) fail fast on API/auth/network errors; update the handler around the requests.get/response.raise_for_status section (the try/except surrounding page_issues, issues.extend and pagination logic) to surface the original exception and include the error context in the raised error/logging. - Around line 240-242: The current main() ignores post_comment()'s return value so comment-post failures don't fail the job; update the block that calls post_comment(comment) to check its boolean result and exit non‑zero on failure (e.g., if not post_comment(comment): log an error and sys.exit(1) or raise SystemExit) so permission/API failures cause the script to fail; reference the post_comment function and the main() flow where the comment is printed and posted. ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 3eb61906-b169-4909-b104-8d3abe5eb2c7
📒 Files selected for processing (1)
.github/scripts/check-duplicates.py
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
1 participant
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.
Summary
This PR introduces a semantic duplicate issue detection tool for GitHub issues.
The script analyzes newly created issues and compares them against existing issues using sentence embeddings to identify potential duplicates.
Related Issue: #396
Key Features
sentence-transformersto generate semantic embeddingsConfiguration
The behavior can be configured using environment variables:
Workflow
Benefits
Future Improvements
Summary by CodeRabbit
New Features
Chores
Style