Lane 10: a test's AI usage counts, marked as a test (D227, #736) #917

Merged

claude-bot merged 6 commits from build/lane10-d227-test-usage into main 2026-09-29 05:16:57 +00:00

Conversation 0 Commits 6 Files changed 12 +1017 -161

claude-bot commented 2026-09-28 17:22:48 +00:00

Collaborator

Copy link

Part of #736 (lane 10). D227: a "Test the bot" run transcribes and summarises with the campaign's own models, so it counts in usage like any other run, marked as a test.

Migration 10e5f6a7b8c9 on 8c9d0e1f2a3b (#909's)

• session_usage.session_id becomes nullable.
• New columns:
  • recording_test_id: a foreign key to recording_tests, ON DELETE SET NULL, indexed;
  • purpose (session or test, with a CHECK). Existing rows backfill to session.
• A second CHECK: a session row has a session_id, and a test row has none.
• The downgrade deletes test rows before restoring NOT NULL.
• Up → down → up with seeded session and test rows, on two fresh Postgres 16 databases. There is one head, and the pin has moved.

The writer

recording_test_service meters ASR and the LLM exactly as process_audio does. It writes a purpose='test' row on success and on every failure path, replacing the old log-only usage.

Every reader checked

• usage_service._load_rows (Admin → AI usage & speed, and its per-campaign view) already outer-joins Session, so a NULL session_id gives a null title and still counts. Tested.
• Preflight (measure_realtime_factor and measure_llm_tokens_per_run) and admin_overview_service don't join Session, so test rows count automatically.
• Hosted entitlements have no SessionUsage reader (the credits live at the payment provider). CAMP-13 and HOME-7 don't read usage today.
• Admin usage page: a test row reads "Test the bot" with a test chip.

Checks

• Backend full suite: 5832 passed, 0 failed, on a fresh Postgres 16.
• vitest: 2747 passed. eslint: 0 errors. ruff: clean.
• Changelog: a changelog.d/ fragment.

🤖 Generated with Claude Code

Part of #736 (lane 10). **D227:** a "Test the bot" run transcribes and summarises with the campaign's own models, so it counts in usage like any other run, marked as a test. ## Migration `10e5f6a7b8c9` on `8c9d0e1f2a3b` (#909's) - `session_usage.session_id` becomes nullable. - **New columns:** - `recording_test_id`: a foreign key to `recording_tests`, `ON DELETE SET NULL`, indexed; - `purpose` (`session` or `test`, with a CHECK). Existing rows backfill to `session`. - **A second CHECK:** a `session` row has a `session_id`, and a `test` row has none. - **The downgrade** deletes `test` rows before restoring NOT NULL. - Up → down → up with seeded session and test rows, on two fresh Postgres 16 databases. There is one head, and the pin has moved. ## The writer `recording_test_service` meters ASR and the LLM exactly as `process_audio` does. It writes a `purpose='test'` row on success and on every failure path, replacing the old log-only usage. ## Every reader checked - **`usage_service._load_rows`** (Admin → AI usage & speed, and its per-campaign view) already outer-joins `Session`, so a NULL `session_id` gives a null title and still counts. Tested. - **Preflight** (`measure_realtime_factor` and `measure_llm_tokens_per_run`) and **`admin_overview_service`** don't join `Session`, so test rows count automatically. - **Hosted entitlements** have no `SessionUsage` reader (the credits live at the payment provider). CAMP-13 and HOME-7 don't read usage today. - **Admin usage page:** a test row reads "Test the bot" with a `test` chip. ## Checks - **Backend full suite:** 5832 passed, 0 failed, on a fresh Postgres 16. - **vitest:** 2747 passed. **eslint:** 0 errors. **ruff:** clean. - **Changelog:** a `changelog.d/` fragment. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

claude-bot added 4 commits 2026-09-28 17:22:49 +00:00

feat(webapp): a test's AI usage counts, marked as a test (D227) ... 905a80b364

WIP checkpoint before merging origin/main: migration 10e5f6a7b8c9
makes session_usage.session_id nullable and adds recording_test_id
+ purpose columns; process_recording_test writes usage rows the same
way the session pipeline does; usage readers updated for NULL
session_id.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

Merge remote-tracking branch 'origin/main' into build/lane10-d227-test-usage 7ed8bbd756

Merge branch 'build/lane8-handout-takeback' into build/lane10-d227-test-usage ... 33748bbaff

Re-chains 10e5f6a7b8c9 (D227: a test's usage counts, marked as a
test) onto 8c9d0e1f2a3b (#896: handout reveal messages) instead of
directly onto 8b8c9d0e1f2a (the map), so the chain stays
8b8c9d0e1f2a -> 8c9d0e1f2a3b -> 10e5f6a7b8c9 with a single head.
Moves the one-head pin accordingly and extends test_notifications'
history docstring with both migrations. Verified with a seeded
up/down/up against a fresh Postgres 16.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

Merge remote-tracking branch 'origin/main' into build/lane10-d227-test-usage 88f10b0b05

rbrooks added 1 commit 2026-09-29 01:56:50 +00:00

Merge remote-tracking branch 'origin/main' into build/lane10-d227-test-usage ff444aefa3

rbrooks added 1 commit 2026-09-29 04:31:22 +00:00

Merge remote-tracking branch 'origin/main' into build/lane10-d227-test-usage 3c1ea1a6a5

claude-bot merged commit cf64d7d62f into main 2026-09-29 05:16:57 +00:00

claude-bot referenced this pull request from a commit 2026-09-29 05:17:00 +00:00

Merge PR #917: D227, a bot test counts against usage (#736)

Sign in to join this conversation.

Reviewers

No reviewers

Labels

Clear labels   

backend

Issue related to the backend

  

bot

Issue related to the bot

  

bug

Something is not working

  

duplicate

This issue or pull request already exists

  

enhancement

New feature

  

frontend

Issue related to the frontend

  

help wanted

Need some help

  

invalid

Something is wrong

  

question

More information is needed

  

recording

Issue related to the recording system

  

security

Security hardening / vulnerability work

  

wontfix

This won't be fixed

No labels

backend

bot

bug

duplicate

enhancement

frontend

help wanted

invalid

question

recording

security

wontfix

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

2 participants

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

rbrooks/Quest-Board!917

No description provided.