{"id":51714092,"url":"https://github.com/rlarson94170/tyler_med","last_synced_at":"2026-07-17T01:01:07.172Z","repository":{"id":370378155,"uuid":"1290290288","full_name":"rlarson94170/tyler_med","owner":"rlarson94170","description":"Claude Code skill: turn clinical-journal PDFs into a token-efficient Markdown wiki + evidence table (study design, DOI, sample size, data source) for literature review.","archived":false,"fork":false,"pushed_at":"2026-07-09T01:24:28.000Z","size":70,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-07-09T03:12:07.699Z","etag":null,"topics":["claude-code","claude-skill","evidence-synthesis","literature-review","medical-literature","pdf-to-markdown","pymupdf","systematic-review"],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/rlarson94170.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-07-05T19:44:16.000Z","updated_at":"2026-07-09T01:24:32.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/rlarson94170/tyler_med","commit_stats":null,"previous_names":["rlarson94170/tyler_med"],"tags_count":1,"template":false,"template_full_name":null,"purl":"pkg:github/rlarson94170/tyler_med","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rlarson94170%2Ftyler_med","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rlarson94170%2Ftyler_med/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rlarson94170%2Ftyler_med/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rlarson94170%2Ftyler_med/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/rlarson94170","download_url":"https://codeload.github.com/rlarson94170/tyler_med/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/rlarson94170%2Ftyler_med/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":35563056,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-07-16T02:00:06.687Z","response_time":83,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["claude-code","claude-skill","evidence-synthesis","literature-review","medical-literature","pdf-to-markdown","pymupdf","systematic-review"],"created_at":"2026-07-17T01:00:34.768Z","updated_at":"2026-07-17T01:01:07.143Z","avatar_url":"https://github.com/rlarson94170.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# tyler_med — Medical Literature Wiki Skill\n\nA [Claude Code](https://claude.com/claude-code) skill that converts a folder of\n**medical / clinical journal PDFs** (JVS, JAMA, NEJM, Ann Surg, EJVES, Cochrane, …)\ninto a token-efficient, two-tier Markdown wiki for literature review — with\nstudy-design classification, DOI extraction, an evidence-table export, and\nduplicate detection.\n\nIt is a medical-tuned fork of the general-purpose **`tyler`** skill from the\n[`econtools`](https://github.com/johanfourieza/econtools) project (see\n[Acknowledgements](#acknowledgements)).\n\n## What it produces\n\n```\nWIKI_DIR/\n├── index.md            # papers grouped by study design (evidence hierarchy),\n│                       #   with DOI links, sample size (N), and data source\n├── index.csv           # evidence-table skeleton (one row per paper)\n├── index.json          # same, machine-readable\n├── papers/*.md         # cleaned full text + rich YAML frontmatter\n├── references/*.md     # trimmed reference lists (kept, not discarded)\n└── .wiki_state.json    # incremental cache\n```\n\nRead `index.md` (or the CSV) to navigate 100+ papers cheaply; open an individual\n`papers/*.md` only when you need the full text. The folder is also a ready-to-use\n**Obsidian vault** (frontmatter → note properties, `[[wikilinks]]`, nested tags).\n\n## Requirements\n\n- Python 3.9+\n- [`pymupdf4llm`](https://pypi.org/project/pymupdf4llm/) (bundles PyMuPDF / `fitz`,\n  used for embedded-metadata and DOI extraction)\n\n```bash\npip install pymupdf4llm\n```\n\n## Usage\n\n**As a Claude Code skill:** copy this folder to `~/.claude/skills/tyler_med/` and\ninvoke `/tyler_med`.\n\n**Directly:**\n\n```bash\npython3 convert.py \"PDF_DIR\" \"WIKI_DIR\" [OPTIONS]\n```\n\n| Flag | Effect |\n|------|--------|\n| `-r`, `--recursive` | Scan `PDF_DIR` subdirectories |\n| `--prefer-pdf-title` | Trust the PDF's embedded metadata title over the filename (good for older, badly-named files) |\n| `--keep-references` | Keep references inline in each paper file |\n| `--drop-references` | Discard references entirely (default: save to `references/`) |\n| `--force` | Re-convert everything, ignoring the incremental cache |\n\n## Medical-tuning highlights\n\n- **Reliable titles** from embedded PDF metadata + DOI, not just the filename.\n- **Mojibake / ligature repair** (e.g. `n ¼ 4,894` → `n = 4,894`, dropped `fi`/`fl`).\n- **Study-design classification** (RCT, systematic review, meta-analysis, cohort,\n  case-control, guideline/consensus, scoping/narrative review, QI, protocol,\n  editorial). A named registry in the title (NSQIP, VQI, Medicare, …) implies a\n  cohort; an explicit `EDITORIAL` tag in a title is honoured.\n- **Structured metadata:** DOI, sample size (`n=`), data source, trial/PROSPERO\n  registration IDs, and clean controlled tags.\n- **Duplicate detection** (shared DOI or year+title) with `⚠️DUP-n` flags.\n\n\u003e **Note:** study design, sample size, and data source are heuristic (best-effort\n\u003e from the abstract). Verify anything load-bearing against the full text.\n\n## Using it as an Obsidian vault\n\nOpen `WIKI_DIR` in Obsidian (or drop it into an existing vault). Each `papers/*.md`\ncarries YAML frontmatter that becomes note **properties**, and the index links every\npaper with `[[wikilinks]]`.\n\n**Tags are deliberately minimal and controlled** — only three nested namespaces, so\nthe tag pane and graph stay clean and queryable:\n\n- `design/…` — study design (e.g. `design/cohort-study`, `design/randomized-controlled-trial`)\n- `source/…` — data source (e.g. `source/nsqip`, `source/vqi`, `source/medicare/cms`)\n- `year/…` — publication year (e.g. `year/2023`)\n\nPer-keyword hashtags are **off by default** (they flood the graph); the human-readable\n`keywords:` property is kept regardless. Pass `--keyword-tags` if you want them.\n\n**Dataview** turns the frontmatter into a live evidence table. Examples:\n\n````md\n```dataview\nTABLE study_type AS Design, year AS Year, sample_size AS N, journal\nFROM \"papers\"\nWHERE contains(data_source, \"VQI\")\nSORT year DESC\n```\n````\n\n````md\n```dataview\nTABLE WITHOUT ID link(file.link, title) AS Paper, doi\nFROM #design/randomized-controlled-trial\nSORT year DESC\n```\n````\n\nFor a **by-topic** view (what to cite for a given point), ask Claude to build\n`index_by_theme.md` (see Step 4 in `SKILL.md`) — a curated thematic grouping with a\ndesign badge and one-line contribution note per paper.\n\n## Acknowledgements\n\nThis project is a derivative work of the **`tyler`** skill in the\n[`econtools`](https://github.com/johanfourieza/econtools) project by\n[@johanfourieza](https://github.com/johanfourieza), used and continued under the\nMIT License. The original `tyler` was built for economics / social-science papers;\n`tyler_med` re-tools it for clinical literature (study-design classification,\nDOI/registry metadata, evidence-table export, mojibake repair, duplicate\ndetection). With thanks to the original author.\n\n## License\n\nReleased under the **MIT License** — see [`LICENSE`](LICENSE). Copyright is held by\n`johanfourieza` for the upstream `tyler` skill and by Robert A. Larson, MD for\nthe medical adaptation.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Frlarson94170%2Ftyler_med","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Frlarson94170%2Ftyler_med","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Frlarson94170%2Ftyler_med/lists"}