Org Web Adapter

Editing journals/2026_06_11.org

:PROPERTIES:
:ID:       1405bd64-956c-460f-84b9-60360392b5b7
:END:
#+title: 2026-06-11

* Meetings :sprint112:work:hungryroot:meeting:
** DONE Hungryroot Book Club - What the River Knows :bookclub:
:PROPERTIES:
:ID:       16c84423-97d7-9add-2546-5ace17ef1f03
:END:
<2026-06-11 Thu 12:30-13:30>

* Tasks :work:hungryroot:sprint112:
** DONE Figure out why migration checks are failing on cookbooks sharing branch :be7895:chore:cookbooks:favorites:
:PROPERTIES:
:ID:       9a5c83c3-e6eb-959a-6e28-52e07f58b7dd
:END:
<2026-06-11 Thu>

** DONE Test backfill script on staging BE-7859 :be7859:investigation:exp242:cookbooks:data:
:PROPERTIES:
:ID:       85c4e338-2a83-70ab-0f27-7e88c6d01074
:END:
<2026-06-09 Tue>
*** Description

This has taken so long ... not even sure it's totally worthwhile, but I feel
like I should keep plugging away.

*** Notes
- Note taken on [2026-06-11 Thu 18:38]
  Decision: meal_frequency backfill — skip for now
  
  The exp_236 audit found 175 meal frequency mismatches and 25 rows (5
  customers) with no quiz row at all. The question was whether to add a meal
  frequency backfill to backfill_exp236_tags.py.
  
  The CSV is a snapshot from June 4. Customers can update meal frequency anytime
  via PATCH /api/customers/<id>/quizzes/. The Quiz model has no change history —
  individual field updates aren't audited. This means mismatches between the CSV
  and the DB are ambiguous: they could be a missing write (backfill needed) or a
  legitimate customer update since June 4 (DB is correct).
  
  Overwriting mismatches risks clobbering valid customer preferences. Creating
  quiz rows for the 5 missing customers is also risky Quiz has many fields
  beyond frequency; a partial row would be incomplete.

The backfill command stays as-is: only customer_tag rows are backfilled.

#+begin_src bash
claude --resume f0cb9d52-88c4-47b8-860b-1cd54d0e95e8
#+end_src