pmc-rct

# JerryLHT/pmc-rct ## Overview This dataset contains cleaned PMC randomized controlled trial articles stored as raw JSON files. - Time coverage: `2023-01` to `2026-04` - Total files: `67501` - Classification labels: `primary_rct, rct_related` - Source types: `html, xml` ## Directory Layout ```text raw/ 2023/ 01/ primary_rct/ json/ PMCxxxx.json ``` The original directory structure is preserved under `raw/`. ## Manifest Files - `manifest/files.jsonl`: one record per JSON file - `manifest/stats.json`: aggregate dataset statistics Each record in `manifest/files.jsonl` includes: - `rel_path` - `year` - `month` - `classification` - `pmc_id` - `pmid` - `doi` - `title` - `trial_registration` - `metadata_month` - `source_type` - `section_count` - `sha256` ## Example Files - `2023/01/primary_rct/json/PMC10003283.json` | `PMC10003283` | `primary_rct` | Evaluation of remineralizing effect of zinc-carbonate hydroxyapatite on the reduction of postrestorative sensitivity: A randomized controlled clinical trial - `2023/01/primary_rct/json/PMC10003293.json` | `PMC10003293` | `primary_rct` | Influence of varying dry air temperatures on postoperative sensitivity, penetration depth, and push-out bond strength of an ethanol/water-based adhesive: An in vivo double-blind clinical trial and in vitro analysis - `2023/01/primary_rct/json/PMC10003580.json` | `PMC10003580` | `primary_rct` | Assessment of risk of bias in randomized controlled trials published in Indian journals pertaining to pharmacology ## JSON Schema Summary Each article JSON file follows this high-level structure: ```json { "metadata": { "pmc_id": "...", "pmid": "...", "doi": "...", "title": "...", "trial_registration": "...", "month": "YYYY_MM", "classification": "primary_rct", "source_type": "xml" }, "sections": [ { "section_key": "abstract", "section_title_raw": "...", "section_title_normalized": "...", "section_number": "...", "blocks": [ { "type": "text", "text_md": "..." } ] } ] } ``` ## Notes - This repo is intended as a raw-file dataset rather than a single tabular split. - Licensing may vary across PMC articles. Document article-level usage constraints before public release. - If you only want a subset, filter via `manifest/files.jsonl` and download the corresponding files.