gh_plot_reports corpus (4,299 plots) + concurrency + 4-GPU pool

CORPUS — 4,299 GH plot reports added (3,797 written + 502 from the
earlier slow run + 319 sitemap-listed URLs that 404'd as
discontinued). Combined with prior 760 varieties + 14 AgriPro
trials = 5,073 total chunks now indexed.

scrape/sources/gh_plot_reports.py — concurrency speedup:
- 4 worker threads (ThreadPoolExecutor), each with its own
  requests.Session for connection-pool efficiency.
- Shared class-level rate limiter (0.25 sec between ANY two
  requests across all threads). Net throughput ~4 req/sec —
  well below any rate-limit threshold a public site enforces.
- Diagnosis vs original 1 req/sec: GH had ZERO rate limiting,
  zero 429s, zero retries. The 1 sec self-throttle was just too
  conservative. Bench:
    1 worker  / 1.0 sec throttle:  ~0.4 plots/sec (190 min ETA)
    4 workers / 0.25 sec throttle: ~3 plots/sec  (~25 min actual)

rag/chunk.py — chunk size cap for nomic-embed-text's 2048-token
context window:
- Empirically tested: failure threshold is ~5,250 chars on
  numeric-heavy trial chunks (chars/token ratio 2.4 vs 3.5 for
  prose). Cap at 4,500 chars to be safely under at worst-case
  2.2 chars/token.
- Applied to BOTH variety and trial chunks. Marked truncated
  chunks with metadata.embed_truncated = True; FULL text stays
  in the on-disk .md for get_page to return verbatim.

.gitea/workflows/{refresh,image-only}.yml — OLLAMA_URL pool
restructured for the 4 GPU-pinned endpoints. Bench (50-chunk
batches on nomic-embed-text):

    .0.125:11434  (RTX 40-series)  242 embeds/sec  ← weight ×4
    .0.2:11436    (GPU-pinned)     108 embeds/sec  ← weight ×2
    .0.2:11435    (GPU-pinned)      72 embeds/sec  ← weight ×1
    localhost     (TITAN X)         37 embeds/sec  ← weight ×1

Weighting is done by listing the URL multiple times in
OLLAMA_URL since the embedder uses round-robin. .0.2:11434 is
explicitly EXCLUDED — it isn't pinned to a specific GPU.

Combined index rebuild for 5,073 chunks now finishes in ~3 min
(was 19+ on the single-endpoint pool).

Smoke tests:
✓ list_versions: 5,073 docs across 6 sources, 2 vendors, 6
  brands, 4 crops (corn 2711, soy 2016, silage 223, wheat 123).
✓ search_trials({crop=corn, state=IA, year=2024}): 3 IA 2024
  corn trials surfaced.
✓ search_trials("Phytophthora resistance soybean trial"): NK
  NK43-W1XFS top-1 in LA 2024 trial (cross-vendor result).
✓ search_trials("AP Iliad Idaho wheat"): AgriPro Washington/N
  Idaho 2025 trial surfaced.
✓ search_trials(product=DKC65-95): 3 corn trials containing
  that hybrid in IL/IA 2024.
✓ search_trials(product=NK1701): 3 corn trials in AR/MS 2024.
✓ Product filter correctly returns EMPTY for products that
  aren't in the corpus (DKC65-20 is a 2023 product; 2023 plots
  deferred). Anti-hallucination contract preserved.

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

rag/chunk.py

@@ -296,7 +296,13 @@ def chunks_from_variety(
     """
     sidecar = json.loads(Path(sidecar_path).read_text(encoding="utf-8"))
     text = _render_variety_chunk(sidecar)
+    # Same 2,048-token cap as trial chunks. Varieties are usually
+    # under 3 KB so this rarely fires, but Bayer hybrids with long
+    # characteristics_groups can run wide — defensive cap.
+    text, truncated = _truncate_for_embed(text)
     meta = _flat_metadata(sidecar)
+    if truncated:
+        meta["embed_truncated"] = True
     chunk_id = f"{meta['source']}::{meta['source_key']}::0"
     yield {
         "id": chunk_id,
@@ -525,6 +531,34 @@ def _flat_trial_metadata(sidecar: dict) -> dict:
     return md
 
 
+# nomic-embed-text caps at 2,048 tokens (Ollama returns HTTP 400 on
+# inputs that exceed this). chars/token ratio varies wildly:
+#   prose:                ~3.5 chars/token
+#   numeric trial tables: ~2.4 chars/token (GH plot reports with
+#                                          full ranking tables)
+# Empirically: GH plot reports failed at 5,261+ chars; agripro
+# trials at 5,552 chars sometimes failed. Cap at 4,500 chars =
+# ~2.2 chars/token worst-case for 2,048 tokens, leaving safe
+# headroom across all source types. The FULL text stays in the
+# on-disk .md so get_page returns it verbatim regardless.
+MAX_EMBED_CHARS = 4500
+
+
+def _truncate_for_embed(text: str) -> tuple[str, bool]:
+    """Cap chunk text to fit nomic-embed-text's 2,048-token context.
+
+    Returns ``(maybe_truncated_text, was_truncated)``. The head is
+    preserved because high-signal content (variety identity, top
+    performers, ratings preamble) sits at the start of every chunk
+    type we produce.
+    """
+    if len(text) <= MAX_EMBED_CHARS:
+        return text, False
+    suffix = "\n…(truncated for embedding; full text via get_page)\n"
+    body = text[: MAX_EMBED_CHARS - len(suffix)].rstrip()
+    return body + suffix, True
+
+
 def chunks_from_trial(
     sidecar_path: Path | str,
     *,
@@ -548,7 +582,10 @@ def chunks_from_trial(
         md_text = md_p.read_text(encoding="utf-8")
 
     text = _render_trial_chunk(sidecar, md_text=md_text)
+    text, truncated = _truncate_for_embed(text)
     meta = _flat_trial_metadata(sidecar)
+    if truncated:
+        meta["embed_truncated"] = True
     chunk_id = f"{meta['source']}::{meta['source_key']}::0"
     yield {
         "id": chunk_id,