akvo / akvo-mis / #1411

91%

push

coveralls-python

Epic/478 viz global question filter (#532)

* [#478] Plan the global dashboard question filter, with pending tests

- Split the VIZ-027 design into a parent spec and backend/frontend
  parts, each task with user and technical acceptance criteria
- Check the design against the code (findings A1-A7) and prototype it
  on local Rural Water Project data
- Record decisions D-13 (filter out only), D-14 (same-named monitoring
  questions read as one, latest answer wins, merged options) and D-15
  (one global_criteria parameter per value; new option values may not
  contain ':' ',' '|', enforced by the backend validators only, so the
  form editor library does not change)
- Compare with the WAI SDG portal's advanced filter (prior art)
- Add backend and frontend tests written before the implementation.
  Backend feature tests carry @pending("<task>"): a failing assertion is
  reported as a skip, so CI stays green, and a test that passes while
  still marked fails the run until its marker is removed

Tests only and docs; no application code changes.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

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

* [#520] Implement the global dashboard question filter backend

- Parse repeated `global_criteria` (option_not_in:<qid>:<value>, split at
  most twice) in the four views after check_ids and the tenant-scoped
  form lookup; never via a serializer, so `global_criteria[0]` cannot
  bypass the public allowlist
- Family check and same-name groups across monitoring forms (D-14);
  400 for malformed entries, more than 50 values, or foreign questions
- Exclude filtered-out registrations with an SQL subquery on id or
  parent_id; "latest answered" is one DISTINCT ON (parent_id) pass,
  dates matched by question name
- Apply the exclusion to values, escalation, map, formula and table
- Validate and snapshot default_filters.questions (merged option labels),
  keep a separate filter_questions allowlist for public da... (continued)

8696 of 9812 branches covered (88.63%)

Branch coverage included in aggregate %.

15785 of 17164 relevant lines covered (91.97%)

0.92 hits per line