justin/seed-mcp
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

agripro + nk scrapers — 146 Syngenta varieties added (760 total in corpus) #6

Merged
justin merged 1 commits from agripro-nk-scrapers into main 2026-05-25 14:17:19 -04:00
Conversation 0 Commits 1 Files Changed 296
+23233 -60
justin commented 2026-05-25 14:17:08 -04:00
Owner
Copy Link
Copy Source

Summary

  • agripro (24 wheat varieties): Drupal Views form scrape via /search-agripro-brand-varieties?title=&variety_type_value=All. Three rated sections per variety (Agronomics / Grain / Disease) with <div class="row"> label/value pairs. Wheat class distribution: 12 HRS, 7 SWW, 3 HRW, 1 HWS, 1 Barley. Closes the Northern Plains HRS gap WestBred didn't cover.
  • nk (122 varieties — recon's "29" was outdated; current NK seed-finder lists 41 corn + 81 soy): ASP.NET WebForms endpoint POST /NKSeeds/{Corn,Soy}ProductFinder.aspx/GetProducts returning {"d": "<html>"} with one <div class="sf-result"> per variety. pdfplumber text extraction on tech-sheet PDFs pulls disease ratings (corn), Phytophthora source genes, SCN race coverage, agronomic numeric ratings (soy), and soil-type adaptation labels (soy).
  • lessons.md scale-direction correction: NK and AgriPro both use 1 = best, lower = more resistant (REVERSED vs Bayer/Golden Harvest). My earlier lesson incorrectly grouped them with Bayer. NK's PDF footer literally prints "1-9 Scale: 1 = Best, 9 = Worst". The chunker reads _scale_direction from each variety's sidecar, so the chunk preamble correctly tells the LLM how to interpret the numbers — anti-hallucination guarantee holds across the reversed-scale vendors.
  • requirements.txt: pdfplumber>=0.11 added for NK tech-sheet text extraction.

Coverage after merge

Source Count Brands
bayer_seeds 475 DEKALB / Asgrow / WestBred
golden_harvest 139 Golden Harvest
nk 122 NK
agripro 24 AgriPro
Total 760 6 brands across 2 vendors

109 wheat varieties (up from 85) — now includes HRS/SWW/HRW/HWS classes across both AgriPro and WestBred.

Known v1 limitations (documented, not blockers)

  • NK corn agronomic chart bars (Drought, Stalk Strength, Emergence) aren't text-extractable from the PDF. Chunk records the label list with an explicit "rated in tech-sheet PDF chart, see source" annotation; full numeric values would need pdf2image+OCR or pdfplumber geometric layout parsing.
  • AgriPro non-code variety lookup ("AP Iliad" search may not pin Iliad at #1 because "Iliad" is not a code-token pattern). Workaround: lookup_variety('agripro-ap-iliad') for exact lookup; brand='AgriPro' filter for branded search.
  • Cross-vendor result diversity: Dense embedding favors Bayer for broad queries because Bayer chunks are longer/richer. A reranker (Phase 6, deferred) would help; for now the agent can use brand filters for vendor-specific results.

Test plan

  • agripro smoke: all 24 wheat varieties parsed; AP Iliad has Stripe Rust 1, Eyespot 2 (1 = best for AgriPro).
  • nk smoke: 122 cards parsed (41 corn / 81 soy); 77 of 81 soy varieties have Phytophthora gene extracted; corn disease ratings extract correctly (e.g. NK8005 Tar Spot 2 = excellent).
  • Re-index produced 760 chunks in 26s on the 3-node Ollama pool.
  • list_versions reports 4 sources / 2 vendors / 6 brands / 3 crops with correct counts.
  • crop_seed_api_lessons(topic='rating-scales') returns the corrected guidance with NK + AgriPro in the reversed-scale bucket.
## Summary - **agripro** (24 wheat varieties): Drupal Views form scrape via `/search-agripro-brand-varieties?title=&variety_type_value=All`. Three rated sections per variety (Agronomics / Grain / Disease) with `<div class="row">` label/value pairs. Wheat class distribution: 12 HRS, 7 SWW, 3 HRW, 1 HWS, 1 Barley. **Closes the Northern Plains HRS gap** WestBred didn't cover. - **nk** (122 varieties — recon's "29" was outdated; current NK seed-finder lists 41 corn + 81 soy): ASP.NET WebForms endpoint `POST /NKSeeds/{Corn,Soy}ProductFinder.aspx/GetProducts` returning `{"d": "<html>"}` with one `<div class="sf-result">` per variety. pdfplumber text extraction on tech-sheet PDFs pulls disease ratings (corn), Phytophthora source genes, SCN race coverage, agronomic numeric ratings (soy), and soil-type adaptation labels (soy). - **lessons.md scale-direction correction**: NK and AgriPro both use **1 = best, lower = more resistant** (REVERSED vs Bayer/Golden Harvest). My earlier lesson incorrectly grouped them with Bayer. NK's PDF footer literally prints *"1-9 Scale: 1 = Best, 9 = Worst"*. The chunker reads `_scale_direction` from each variety's sidecar, so the chunk preamble correctly tells the LLM how to interpret the numbers — anti-hallucination guarantee holds across the reversed-scale vendors. - **requirements.txt**: `pdfplumber>=0.11` added for NK tech-sheet text extraction. ## Coverage after merge | Source | Count | Brands | |---|---|---| | `bayer_seeds` | 475 | DEKALB / Asgrow / WestBred | | `golden_harvest` | 139 | Golden Harvest | | `nk` | 122 | NK | | `agripro` | 24 | AgriPro | | **Total** | **760** | **6 brands across 2 vendors** | 109 wheat varieties (up from 85) — now includes HRS/SWW/HRW/HWS classes across both AgriPro and WestBred. ## Known v1 limitations (documented, not blockers) - **NK corn agronomic chart bars** (Drought, Stalk Strength, Emergence) aren't text-extractable from the PDF. Chunk records the label list with an explicit "rated in tech-sheet PDF chart, see source" annotation; full numeric values would need pdf2image+OCR or pdfplumber geometric layout parsing. - **AgriPro non-code variety lookup** ("AP Iliad" search may not pin Iliad at #1 because "Iliad" is not a code-token pattern). Workaround: `lookup_variety('agripro-ap-iliad')` for exact lookup; `brand='AgriPro'` filter for branded search. - **Cross-vendor result diversity**: Dense embedding favors Bayer for broad queries because Bayer chunks are longer/richer. A reranker (Phase 6, deferred) would help; for now the agent can use brand filters for vendor-specific results. ## Test plan - [x] `agripro` smoke: all 24 wheat varieties parsed; AP Iliad has Stripe Rust 1, Eyespot 2 (1 = best for AgriPro). - [x] `nk` smoke: 122 cards parsed (41 corn / 81 soy); 77 of 81 soy varieties have Phytophthora gene extracted; corn disease ratings extract correctly (e.g. NK8005 Tar Spot 2 = excellent). - [x] Re-index produced 760 chunks in 26s on the 3-node Ollama pool. - [x] `list_versions` reports 4 sources / 2 vendors / 6 brands / 3 crops with correct counts. - [x] `crop_seed_api_lessons(topic='rating-scales')` returns the corrected guidance with NK + AgriPro in the reversed-scale bucket.
justin added 1 commit 2026-05-25 14:17:08 -04:00
agripro + nk scrapers — 146 Syngenta varieties added (wheat + corn/soy) 9ce920f622 
agripro (24 varieties)
- Drupal Views form scrape via /search-agripro-brand-varieties with
  explicit GET params (sidesteps the AJAX-only-on-load default that
  returns an empty form skeleton).
- Per-variety parse: <h1>, .field--node--variety-type--variety,
  .field--node--tag-line--variety, .field--node--body, plus the
  three rated sections (Agronomics / Grain / Disease) with their
  <div class="row"><div class="label">label</div><div>value</div>
  pairs.
- Wheat-class distribution: 12 HRS, 7 SWW, 3 HRW, 1 HWS, 1 Barley
  — provides the Northern Plains HRS coverage WestBred lacks.

nk (122 varieties — recon's "29" was outdated; the current NK seed
finder lists 41 corn + 81 soy)
- ASP.NET WebForms endpoint:
  POST /NKSeeds/{Corn,Soy}ProductFinder.aspx/GetProducts returns
  {"d": "<html>"} where the inner HTML is one <div class="sf-result">
  per variety. BeautifulSoup tokenizes the whole blob.
- Per-card: product code (NK8005, NK008-P8XF), RM/MG from the
  title <span>, "Brands Available" trait variants, marketing
  positioning + bullet strengths, tech-sheet PDF URL.
- pdfplumber text extraction on the tech-sheet PDFs adds:
  * corn disease ratings (Gray Leaf Spot, NCLB, Goss's Wilt,
    Anthracnose, Tar Spot, Fusarium, etc.) where the PDF prints
    "Label N" lines (text-extractable)
  * soybean Phytophthora source genes (Rps1c, Rps3a, ...)
  * soybean SCN race coverage
  * soybean agronomic ratings (Emergence, Standability, Shatter
    Tolerance, Green Stem) with text-extractable 1-9 values
  * soybean soil-type adaptation (Best/Good/Fair/Poor) for drought
    prone / high pH / poorly drained / etc.
- Agronomic rating BARS for corn (Emergence, Stalk Strength,
  Drought) are not text-extractable; we record the labels with an
  explicit "rated in PDF chart, see tech sheet" value so the agent
  can direct the farmer at the source for those numbers.

Scale-direction correction in lessons.md:
- NK and AgriPro both use 1 = best, lower = more resistant — the
  REVERSED convention vs Bayer / Golden Harvest. NK's tech-sheet
  footer literally prints "1-9 Scale: 1 = Best, 9 = Worst".
  AgriPro positioning on stripe-rust-resistant varieties (AP Iliad
  with Stripe Rust 1, Eyespot 2) confirms the same direction.
- sources-not-yet-indexed section trimmed to just Beck's PFR +
  Beck's products — everything else IS now in the corpus.

Cross-vendor coverage after this PR: 760 varieties.
  bayer_seeds     475 (DEKALB 288 / Asgrow 102 / WestBred 85)
  golden_harvest  139
  nk              122  (41 corn / 81 soy)
  agripro          24  (12 HRS / 7 SWW / 3 HRW / 1 HWS / 1 Barley)
Vendors: Bayer, Syngenta. Brands: 6. Crops: corn, soy, wheat (109
wheat now, up from 85).

requirements.txt: pdfplumber>=0.11 for NK tech-sheet parsing.

Co-Authored-By: Claude Opus 4.7 (1M context) <[email protected]>
justin merged commit 7b3da908e0 into main 2026-05-25 14:17:19 -04:00
Sign in to join this conversation.
Reviewers
No Reviewers
Labels
No items
No labels
Milestone
No items
No Milestone
Projects
Clear projects
No projects
Assignees
Clear assignees
justin (Justin Paul)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: justin/seed-mcp#6