akvo / iwsims / #60

88%

main: 88%

Pull #22

coveralls-python

[#20] Rewrite Task 6 Part H to match implementation reality

The original Part H over-promised a mechanical -40% values_functions collapse.
Updated to record what actually happened and why the rest was deferred:

- Reality check: the MV optimization's core goal (query complexity) is already
  met (MV:base ratios 36:4 / 12:7 / 8:2 / 7:1); residual line count is feature
  breadth, not removable duplication.
- Status (done): criteria->indexed-MV matching, utils _typed_value extraction,
  values_functions _date_answer_sq + _finalize_month/_date.
- What is NOT a safe win (investigated, deferred): get_options_by_qname share
  (different option shapes); the matrix collapse (cells differ intentionally:
  FormData-vs-MV source, distinct-vs-plain count, registration fallback, dynamic
  date windows) -> needs a redesign, not a merge; extend-MV-to-delete-fallbacks
  has a bad tradeoff (fallbacks are non-materializable / necessary).
- Reduction table reframed as original-estimate vs actual-outcome with current
  line counts.

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

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

5183 of 6053 branches covered (85.63%)

Branch coverage included in aggregate %.

9977 of 11170 relevant lines covered (89.32%)

0.89 hits per line