docker: production image + Gitea Actions for monthly refresh

Dockerfile: self-contained image with corpus + Chroma + BM25 baked
in. Drawbar's compose pulls + runs without volume mounts. Built from
sources.json (labels schema), PRODUCT_NAME=crop_chem by default,
HYBRID_SEARCH=true (always-on for production quality). RERANK_URL +
OLLAMA_URL get set at compose time.

.gitea/workflows/refresh.yml: monthly cron (1st @ 06:00 UTC) does
full scrape → reindex → image push. Scrapes Bayer (~30 min) +
EPA PPLS row-crop filtered (~7h). Skips reindex+push if no corpus
diff. Tags pushed: :latest, :<sha12>, :corpus-<YYYY.MM.DD>.

.gitea/workflows/image-only.yml: on-demand or auto on code-only
pushes to main (paths: docs_mcp/, rag/, scrape/, requirements.txt,
Dockerfile, sources.json). Reindexes from committed corpus, builds
image, pushes. ~10 min vs ~9h full refresh.

.gitignore: corpus/ now COMMITTED (4,159 labels, 265 MB of .md +
sidecars). Lets image-only.yml rebuild indexes without re-scraping.
chroma/ + bm25/ still gitignored (regenerable binary indexes).

.dockerignore: drops venv, eval results, PLAN/README/CLAUDE.md,
deploy/, .git/ — keeps the image lean. corpus + chroma + bm25
explicitly NOT in dockerignore (those go INTO the image).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

corpus/epa_ppls/19713-79.md

@@ -0,0 +1,48 @@
+# ATRAZINE PLUS LINURON
+
+- EPA Reg No: **19713-79**
+- Registrant: DREXEL CHEMICAL COMPANY
+- Signal word: Caution
+- Active ingredients: Linuron (30.8%); Atrazine (30.8%)
+- Label accepted: 1997-04-21
+- Source PDF: https://www3.epa.gov/pesticides/chem_search/ppls/019713-00079-19970421.pdf
+
+---
+Drexel Chemical Company 
+;97/3- '71 
+.-
+SUPPLEMENTAL LABELING 
+Drexel Atrazine Plus Linuron 
+EPA Reg. No. 19713-79 
+All applicable directions, restrictions and precautions on the EPA 
+registered label are to be followed. This label must be ·in the possession of 
+the actual pr()duct user at the time of application. 
+TILE-TERRACED FIELDS CONTAINING STANDPIPES 
+To ensure protection of surface water from runoff through standpipes and tile 
+outlets in terraced fields, one of the following options may be used: 
+1) Do not apply this product within 66 feet of standpipes in tile-outieUed terraced 
+fields. 
+2) Apply this product to the entire tile-outletted terraced field and immediately 
+incorporate it to a depth of 2 to 3 inches in the entire tile-outletted terraced field. 
+3) Apply this product to the entire tile-outletted terraced field under a no-till 
+practice only when high crop residue management practices are used. High crop 
+residue management practice is described as a crop management practice 
+where little or no crop residue is removed from the field during or after crop 
+harvest. 
+-- - - -
+EPA Reg. No. 19713-79 
+EPA Est. No. 19713-MS-1 
+labeling Date: March 14, 1997 
+Manufactured by: 
+Drexel Chemical Company 
+. P. O. Box 13327 
+Memphis, TN 38113-0327 
+1700 Channel Avenue· Post Office Box 13327· Memphis, Tennessee 38113"()327 
+Phone: (901) 774-4370 • Fax: (901) 774-4555 • Cable Address: Drexel Chern Mis. 
+4i if'i'fU' 
+'" -\.> 
+"-0 ____ _ 
+-U 
+C) 
+>-' 
+Ir;;f{