Add WebMD mirror + task verifiers (site by @alidezhihui, verifiers by reviewer)#68
Open
Django-Jiang wants to merge 6 commits into
Open
Add WebMD mirror + task verifiers (site by @alidezhihui, verifiers by reviewer)#68Django-Jiang wants to merge 6 commits into
Django-Jiang wants to merge 6 commits into
Conversation
Facts sourced from openFDA drug labeling and MedlinePlus/NIH public-domain health topics; all prose original; authors are fictional personas; images are generated placeholders (HF-managed, not committed). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…+ quality fixes - Breadth: every natural task query now returns >=6 results. Added 11 near-miss drugs (BPH alpha-blockers/5-ARIs, melatonin supplement, liothyronine, methimazole, allopurinol, colchicine, pravastatin, lovastatin), 6 conditions (Enlarged Prostate, Pseudogout, Labyrinthitis, Meniere's, Vestibular Neuritis, Vestibular Migraine), 10 articles (vaccine news, sleep supplements, thyroid labs, uric acid, dizziness triage, statin choice, BPH options). openFDA/MedlinePlus-sourced facts, original prose; all 20 task answers verified unchanged. - De-leak: removed author byline + published date from section listing rows and dates from author-page article lists; answer strings now hit zero of 127 listing surfaces (audit in scraped_data/harden_audit.md). - Consistency: gout/hyperthyroidism/high-cholesterol related-drug lists now match their treatments text; reviewed>=published and slug referential checks pass. - Quality: symptom checker batch-loads symptom links (selectinload); /drugs active_letters via distinct substr query; dead Drug.letter property removed; empty-note <p>-inside-<ul> fixed to <li>; needless |safe on save-button literals replaced with plain if/else. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
webmd.tar.gz (seed DB md5 c310c56b..., 233 placeholder images) uploaded to huggingface.co/datasets/ChilleD/WebHarbor as PR aiming-lab#42. Pinning to refs/pr/42 lets this branch build end-to-end before the assets PR merges. Bump to the merge sha (or back to main) once HF PR aiming-lab#42 lands. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
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
2 participants
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.
Reviewer PR for the WebMD mirror (site by @alidezhihui, PR #62). Adds the deterministic grading contract: sites/webmd/verify/ (verify_lib.py + verify_0..19.py) and backfills verifier_path + judge_rubric into tasks.jsonl (no answer key). Mechanical/visual/functional/task-quality review passed; grading contract validated (no-op FAIL, correct PASS, wrong FAIL, state-mismatch FAIL, eval_judge --verifier True path).