phase7cbatch.txt

Phase 7c batch — process Oregon school districts via the cert table (no online cert lookups).

Read inputs:
  • /workspace/batch.json — JSON array of districts to process: [{name, hq_city, county, enrollment, homepage}, ...]
  • /workspace/vendor-certs.json — authoritative vendor → certified standards table built in Phases 6+7. TRUST THIS TABLE. Do NOT fetch site.imsglobal.org or ed-fi.org/certified-products yourself.

For each district in /workspace/batch.json:

1. Geocode: produce lat/lng for hq_city. Use web_search if you don't already know the city's coords; otherwise use known Oregon city coordinates.

2. Software detection: web_fetch the homepage. Then check the navigation for tech/parent/student/family/IT pages and fetch a couple. Also run a `site:<homepage_host>` web_search for each vendor in this list: PowerSchool, Infinite Campus, Synergy, Skyward, Canvas, Schoology, Google Classroom, Seesaw, ClassDojo, ParentVUE, StudentVUE, ParentSquare, Clever, ClassLink, SchoolMessenger.
   For each confirmed mention (active-use context, not historical, not generic), add to the district's software array: {name, source_url, evidence: <one-line>, confidence: high|med|low}.
   Confidence rules:
     • high — page returns 200, current/active-use context, ≥1 corroborating page on the district domain
     • med — page returns 200, single mention, plausibly current
     • low — only legacy/historical mention or 404 on follow-up

3. Standards inference (from the cert table — NO online lookups):
   For each detected software, look up the vendor in vendor_certifications. Copy each of those certified standards to the district's standards array as: {name, inferred_from: <software>, source_type, source_url, confidence: "high"}.
   Then add ALL state_mandates universally to every district.
   If a software vendor is NOT in vendor_certifications (e.g. ParentSquare, Seesaw, ParentVUE/StudentVUE), record the software but don't add any standards from it.

4. ESD-hosted Synergy detection: if a Synergy source URL contains cesdk12.org / cascadetech.org / lblesd.k12.or.us / similar regional ESD domain, set district's `synergy_host = {name: "<ESD>", url: "<ESD page>"}`. Otherwise null.

Output: write the full processed array to /mnt/session/outputs/batch-results.json. Schema per district:
  {rank: null, name, hq_city, county, enrollment, enrollment_year, enrollment_source, homepage, lat, lng, software: [...], standards: [...], synergy_host: {...}|null}

Be efficient — DON'T narrate every district; just process them. Stop and summarize:
  • Total districts processed
  • Software detection counts by vendor (across this batch)
  • Districts with no software detected (likely small districts with minimal web presence)
  • Any districts where homepage was unreachable