torchcell.literature
The literature capture subsystem. It reads papers and supplementary files from Zotero (zotero), OCRs or extracts them into markdown (ocr, extract, scanned), fetches supplementary data (si_data), and records a sha256 provenance manifest for every stored artifact (manifest, provenance) in the on-disk mirror. server is the read-only HTTP endpoint (tc-lit-server) that serves the mirrored artifacts, and sync reconciles the mirror against a Zotero collection.
Classes
Summary of a mirror-wide backfill run. |
|
Provenance + integrity record for one paper's captured artifact. |
|
Runtime configuration for the literature endpoint. |
|
The shared |
|
Credentials for a single Zotero library. |
|
Synchronous handle to one Zotero library. |
Functions
Regularize one citation-key directory to a |
|
Backfill manifests for every citation-key directory under |
|
Capture a paper's full artifact from Zotero + external SI data. |
|
Generate a Better BibTeX-style citation key from paper metadata. |
|
Scan an artifact directory and build its manifest. |
|
Write |
|
OCR every PDF in an artifact directory: |
|
OCR one PDF to markdown next to it ( |
|
Build the FastAPI app bound to |
|
Factory for |
|
Resolve a DRYAD dataset DOI to |
|
Download SI data files into |
|
Construct a pyzotero client from a |
|
Run a zero-arg Zotero call with bounded exponential-backoff retry. |