feat: add --feedback flag for search result quality feedback loop (#604) - #611
feat: add --feedback flag for search result quality feedback loop (#604)#611lb1192176991-lab wants to merge 1 commit into
Conversation
|
👋 Still actively working on this PR! Acknowledging the competition with #607 and #622 for the same feature (#604). I've reviewed all three implementations and here's where we stand: Current blockers on #611 (shared with competitors):
What makes #611 competitive:
Next steps I'm planning:
Is there any specific area you'd like me to focus on to make this the winning PR? Happy to iterate. cc @Ikalus1988 |
|
Thanks for the detailed update. To answer your questions:
Fix DCO: |
e515422 to
d7eb671
Compare
- Add --feedback flag to search_knowledge.py - After displaying results, prompt: 'Was this helpful? (y/n/c=comment)' - Log feedback to data/search-feedback.jsonl - No PII in logs (only query + result info + feedback) - No external dependencies - Existing search behavior unchanged when --feedback not used Closes Ikalus1988#604 Signed-off-by: lb1192176991-lab <lb1192176991-lab@users.noreply.github.com>
|
@Ikalus1988 DCO fixed! Rebased onto upstream/main, resolved conflict in search_knowledge.py (preserved both |
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
|
Closing — this work has been superseded by main branch implementation (v2.13.0/v2.14.0). The intake endpoint, feedback flag, and triage pipeline are already in place. Thank you for the contribution. |
What
Adds a
--feedbackflag tosearch_knowledge.pyso users/agents can report whether search results were useful.Was this helpful? (y/n/c=comment/Enter=skip)data/search-feedback.jsonlwith: query, found_results, feedback, timestampWhy
This feeds into #574 (feedback hub) and #591 (demand board) — enabling systematic tracking of search result quality and identification of knowledge gaps.
Testing
data/search-feedback.jsonlcontains valid JSONL entries--feedbacknot used