akvo / akvo-mis / #468 / 1

89%

main: 89%

push

coveralls-python

[#146] feat(akvo-flow): fuzzy Flow→MIS question matching and mapping CSV export

- Add fuzzy text-only matching pipeline for mapping Flow questions to MIS questions
  using `rapidfuzz` (partial_ratio) with normalization helpers for robust comparisons.
- Provide `flatten_flow_form_pd` to normalize Flow/MIS JSON into tabular form for matching.
- Implement `match_questions_by_text` and `find_best_match_by_text` with configurable
  similarity threshold (default 80) and confidence categorization (high/medium/low).
- Ensure deduplication and stable outputs:
  - Aggregates matched MIS IDs/labels where multiple children match.
  - Drops duplicates and writes mapping CSVs to `backend/source/akvo-flow/<flow_id>_forms_mapping.csv`.
- Helps automation and downstream seeding by producing deterministic mappings and
  a concise matching summary for review.

Closes #146

Signed-off-by: Your Name <your.email@example.com>

3503 of 4069 branches covered (86.09%)

Branch coverage included in aggregate %.

7239 of 8051 relevant lines covered (89.91%)

0.9 hits per line