{"id":49767054,"url":"https://github.com/retospect/acatome-extract","last_synced_at":"2026-05-11T10:57:35.051Z","repository":{"id":343541180,"uuid":"1178150903","full_name":"retospect/acatome-extract","owner":"retospect","description":"PDF extraction pipeline for acatome — Marker/fitz, metadata, block chunking","archived":false,"fork":false,"pushed_at":"2026-05-07T17:56:44.000Z","size":254,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-05-11T10:56:56.760Z","etag":null,"topics":["document-parsing","extraction","pdf","python","scientific-papers"],"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/retospect.png","metadata":{"files":{"readme":"README.md","changelog":"CHANGELOG.md","contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":"SECURITY.md","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":"2026-03-10T18:33:06.000Z","updated_at":"2026-05-07T17:56:48.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/retospect/acatome-extract","commit_stats":null,"previous_names":["retospect/acatome-extract"],"tags_count":7,"template":false,"template_full_name":null,"purl":"pkg:github/retospect/acatome-extract","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/retospect%2Facatome-extract","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/retospect%2Facatome-extract/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/retospect%2Facatome-extract/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/retospect%2Facatome-extract/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/retospect","download_url":"https://codeload.github.com/retospect/acatome-extract/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/retospect%2Facatome-extract/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":32891966,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-10T13:40:02.631Z","status":"online","status_checked_at":"2026-05-11T02:00:05.975Z","response_time":120,"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":["document-parsing","extraction","pdf","python","scientific-papers"],"created_at":"2026-05-11T10:57:34.334Z","updated_at":"2026-05-11T10:57:35.036Z","avatar_url":"https://github.com/retospect.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# acatome-extract\n\nPDF extraction and enrichment pipeline for scientific papers. Converts PDFs into structured, searchable bundles with block-level summaries and embeddings.\n\n## Features\n\n- **Marker PDF extraction** — structured block extraction with headings, tables, figures\n- **Fitz fallback** — recursive character chunking when Marker is unavailable\n- **LLM enrichment** — block and paper summaries via Ollama or litellm\n- **Embeddings** — sentence-transformer embeddings for semantic search\n- **File watcher** — `acatome-extract watch` monitors an inbox folder\n- **Bundle format** — `.acatome` companion files for sharing pre-built extractions\n- **CLI** — `acatome-extract` command for extract, enrich, and watch workflows\n\n## Installation\n\n```bash\nuv pip install -e .\n```\n\nOn **macOS/Linux** this includes Marker for structured PDF extraction.\nOn **Windows** it installs with the lighter pymupdf (fitz) backend by default.\nTo add Marker on Windows (requires C build tools):\n\n```bash\nuv pip install -e \".[marker]\"\n```\n\nWith GPU acceleration (embeddings + torch):\n\n```bash\nuv pip install -e \".[gpu]\"\n```\n\nEverything at once:\n\n```bash\nuv pip install -e \".[full]\"\n```\n\n## Usage\n\n```python\nfrom acatome_extract.pipeline import extract\n\nbundle = extract(\"/path/to/paper.pdf\")\n```\n\n## CLI\n\n```bash\n# Extract (RAKE summaries included automatically, no LLM needed)\nacatome-extract extract paper.pdf\nacatome-extract extract --type datasheet TI_LM317.pdf   # non-article types\n\n# Enrich — embeddings only by default; add --summarize for LLM summaries\nacatome-extract enrich /path/to/bundle\nacatome-extract enrich --summarize /path/to/bundle       # enable LLM summaries\nacatome-extract enrich --summarize --skip-existing dir/   # incremental LLM pass\n\n# Watch — extract + embed + ingest; LLM summaries off by default\nacatome-extract watch ~/papers/inbox\nacatome-extract watch ~/papers/inbox --summarize          # enable LLM summaries\n\n# Migrate old bundles to new summaries dict format + add RAKE\nacatome-extract migrate ~/.acatome/papers\nacatome-extract migrate ~/.acatome/papers --dry-run       # preview changes\n\n# Supplements\nacatome-extract attach parent-slug supplement.pdf --name s1\n```\n\n### Summaries\n\nExtraction always generates **RAKE** (extractive keyword) summaries — instant, no LLM required. LLM-based summaries are opt-in via `--summarize` and require an Ollama or litellm-compatible model.\n\nRAKE summaries are used as the default for search and display. To add LLM summaries later:\n\n```bash\nacatome-extract enrich --summarize --skip-existing ~/.acatome/papers\n```\n\n### Sidecar metadata\n\nPlace a `\u003cstem\u003e.meta.json` alongside any PDF to override metadata:\n\n```json\n{\"type\": \"datasheet\", \"title\": \"LM317 Regulator\", \"author\": \"Texas Instruments\", \"year\": 2022}\n```\n\nSupported fields: `type`, `title`, `author` (string or list), `year`, `doi`,\n`abstract`, `journal`, `s2_id`, `arxiv_id`, `verified`.\n\nUse **explicit `null`** to clear a field that the upstream Crossref/S2 lookup\ngot wrong — empty strings are ignored for backward compatibility:\n\n```json\n{\"doi\": \"10.1002/9781118519301.ch5\", \"s2_id\": null, \"verified\": true}\n```\n\nWhen `\"verified\": true` is set, the fuzzy-title verification gate is\nbypassed — useful for PDFs where the real title doesn't appear on page 1\n(Elsevier header strips, ACS internal production PDFs, abstract collections).\n\n## Dependencies\n\n- **acatome-meta** — metadata lookup and verification\n- **marker-pdf** — structured PDF extraction\n- **litellm** / **Ollama** — LLM-based enrichment\n\n## Testing\n\n```bash\nuv run python -m pytest tests/ -v\n```\n\n## License\n\nGPL-3.0-or-later — see [LICENSE](LICENSE).\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fretospect%2Facatome-extract","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fretospect%2Facatome-extract","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fretospect%2Facatome-extract/lists"}