feat(b2b_learner_records): serve activity from the learner-records MVs

[blarghmatey]

What are the relevant tickets?

N/A. Stacked on #60. Reads the columns added in mitodl/ol-data-platform#2693.

Description (What does it do?)

Replaces the NULL placeholders for last_active_on, days_active, videos_watched, problems_attempted and chatbot_interactions on /enrollments, and last_active_on and courses_in_progress on /learners, with the MV columns. videos_watched is the MV's videos_played. outcomes_consent_on stays NULL.

completion_status = in_progress now means a nonzero grade or any tracked activity. mv_b2b_learner.courses_in_progress uses that definition. Without this change, the default /learners count and the /enrollments statuses would disagree. The recomputed rollup (contract_id / include_inactive) counts courses_in_progress with the same CASE and takes MAX(last_active_on) over the enrollments already in scope. By default that's active enrollments only, like the MV. Under include_inactive it's every enrollment.

Spec and model descriptions:

• The counters were described as distinct blocks. The MV sums per-day distinct counts, so a block used on two days counts twice. organization_administration_report counts videos_watched per day too, and the b2b_dashboard totals sum it.
• x-data-readiness on the activity fields moves off pending-model.
• Activity doesn't move updated_since. A day's activity first shows up at a refresh after that day began, so a cursor derived from it would already sort below the as_of a partner passes. Partners get activity from a full reload. Real change tracking is a separate piece of work.

How can this be tested?

• uv run pytest: 241 passed. New cases cover the enrollment projection and status CASE, the default learner projection, the recomputed courses_in_progress/last_active_on, and a shared enrollment keeping its activity fields.
• ruff check, ruff format --check and mypy src are clean.
• Ran the generated SQL for /enrollments (with and without include_inactive) and all three /learners paths against DuckDB stub tables shaped like the MVs. Statuses, last_active_on and courses_in_progress agree across the paths. Not run on StarRocks: the columns don't exist there until #2693 merges and the MVs rebuild.

Checklist:

• Merge feat(b2b_learner_records): serve /courses from mv_b2b_contract_courserun #60 (and feat(b2b_learner_records): add the learner-records tenant #56) first
• Deploy only after DESC b2b_learner_records.mv_b2b_learner_enrollment and DESC b2b_learner_records.mv_b2b_learner list the new columns. Until then, the queries fail on unknown columns.

🤖 Generated with Claude Code

https://claude.ai/code/session_01XMuwnLQwHtHuH7aECS7sf3

[Copilot]

🔵 Needs a closer look

Deployment depends on an unmerged cross-repository MV schema change that has not yet been validated in StarRocks.

Pull request overview

Populates learner-record activity fields from the new materialized-view columns and aligns progress calculations across endpoints.

Changes:

• Projects enrollment and learner activity metrics.
• Includes tracked activity in in_progress calculations.
• Updates tests and API documentation.

File summaries

File	Description
queries.py	Reads activity columns and updates rollups/statuses.
models.py	Documents populated activity fields.
test_learner_records.py	Tests activity projections and calculations.
b2b-learner-records-v1.yaml	Updates the public API contract.
b2b-learner-records-onepager.md	Marks the activity model available.
b2b-learner-records-design.md	Documents the implemented activity pipeline.

Review details

• Files reviewed: 6/6 changed files
• Comments generated: 0
• Review effort level: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.