943baa3c0237657445c0defa61bfc3240b58caff
New 'mosaic ingest' CLI command bypasses HTTP collection to import raw documents (PDF, HTML, text) or pre-extracted JSONL data directly into the pipeline. Modes: - Single/multiple files: -f file1.pdf -f file2.html -s source_name - Directory bulk import: -d /path/to/dump/ -s source_name -r - JSONL extracted data: -f data.jsonl -s name -c tools - Real-time watch: -d /incoming/ -s name -w (polls every 2s) - Auto-parse: --parse flag runs parser on newly ingested docs Files are hashed, deduplicated, and stored in the cache using the same hash-sharded scheme as HTTP-collected documents. DevTrack: mosaic #498
Description
Intelligence extraction platform for open-source and declassified document corpora.
Languages
Python
100%