reprocess

Overview ¶

reprocess — One-shot script to reprocess existing ready pieces through docpipe + BufferWriter. Run once after deploying the new sas_ingester with MarkdownConverter + BufferWriter support.

Usage:

reprocess -db /path/to/sas_ingester.db -chunks /path/to/chunks -buffer /path/to/buffer/pending

What it does:

1. Lists all pieces with state=ready that have no markdown yet
2. For each: assemble chunks → docpipe.Extract → store markdown + write .md to buffer
3. Skips pieces that already have markdown (idempotent)