feat: tag topics and lessons to CCSS focus standards; load grades 7-8

[jcasimir]

What

This is prerequisite one of Unit 10 in the Interstandard standards-translator plan (CommonMath's lessons need CCSS tags before they can be retargeted to state codes). It closes the "Standards parsing (quick win)" todo in plans/todos.md.

• Engageny::FocusStandardTagger (lib/engageny/focus_standard_tagger.rb) parses the ny-list-focusstandards markup out of a Topic's overview_html — EngageNY lists Focus Standards once per topic overview, not per lesson, using the same Aspose list-style class convention as ny-h1 headings (see Engageny::Importer#extract_topic_title). Codes are normalized to the dotted form standards.code uses (an embedded cluster letter like the A in 5.NBT.A.1 is stripped, since our imported Standard rows never carry one — confirmed against the existing grade 4-6 data). It creates polymorphic StandardTagging rows on the Topic and, with the same set of standards, on every Lesson beneath it (union of the topic's tags). It's idempotent: existing taggings are left alone.
• Codes that don't resolve to a Standard row (grade 7-8 codes that weren't imported yet, or bare cluster references like 5.NBT.A that were never imported at all) are reported, not silently dropped.
• lib/tasks/standards.rake adds standards:tag_focus, which runs the tagger over every Topic with overview HTML and prints topic/lesson tagging counts plus an unresolved-code frequency report (also written to tmp/standards_tag_focus_unresolved.txt).
• scripts/fetch_ccss_standards.rb now pulls CCSS grades 4-8 (was 4-6) from the same SirFizX/standards-data source. Regenerated data/ccss-math-grades-4-8.json (180 standards, up from 110) and pointed ccss:import at it. This is what makes grade 7-8 focus-standard codes resolve once topics for those grades are imported.
• Enabled rails/test_unit/railtie (it was commented out) and added test/test_helper.rb — the repo had no test infrastructure at all despite bin/rails test being documented in CLAUDE.md. Pinned .ruby-version via a mise.toml.

Where the focus-standard markup actually lives

The real EngageNY HTML (data/engageny/grade-*-html/) is gitignored and not present in this checkout — it only exists in the production database and possibly a local dev checkout. So I could not run the tagger against real data or confirm the exact DOM shape (e.g. whether ny-list-focusstandards decorates <p> tags per line, as ny-h1 does for headings, or real <ul>/<li> markup) beyond what lib/engageny/*.rb implies. I built the tagger and its tests (test/lib/engageny/focus_standard_tagger_test.rb, test/tasks/standards_tag_focus_test.rb) against hand-made fixtures under test/fixtures/files/ that mirror the class-on-paragraph convention documented in the importer, and used a CSS-class selector (.ny-list-focusstandards) rather than a specific tag so it matches regardless of which element carries the class.

Before this ships to lessons that actually have this markup, run bin/rails standards:tag_focus against a copy of the production data and check the printed counts and unresolved list — if the real markup differs from what I assumed (e.g. codes formatted differently, or not text-only), the extraction regex or CSS selector may need a quick follow-up adjustment. I'd treat that as a fast, low-risk fix rather than a blocker on merging this scaffolding.

How to run

bin/rails standards:tag_focus

Grades 7-8 standards import (needed before those topics' codes resolve):

bin/rails ccss:import

Post-Deploy Monitoring & Validation

• Run bin/rails standards:tag_focus after deploy and check the printed counts (topics tagged, lessons tagged) and the unresolved-code report.
• No schema change — standard_taggings and standards already existed and are polymorphic/ready for this.

Test plan

• bin/rails test (new tests only — 10 runs, 41 assertions, 0 failures)
• bin/rubocop on changed files — no offenses
• bundle exec brakeman -q — no warnings (repo's bin/brakeman --ensure-latest fails on version drift unrelated to this change, per known repo issue)
• Run standards:tag_focus against real production data once available and confirm the markup assumptions above

🤖 Generated with Claude Code

https://claude.ai/code/session_011EqZeVdG5q6KDS1Q8sbcxu