Jun-te Kim
|
708f1b5d18
|
repositories: UserAddressRepository + UserAddressCsvS3Repository (CSV-on-S3 adapter)
Adds the persistence layer for UserAddress batches:
- Abstract UserAddressRepository with load_batch / save_batch.
- Concrete UserAddressCsvS3Repository over CsvS3Client:
- load_batch reads canonical upload columns (Address 1/2/3, Postcode,
Internal Reference), comma-joins non-empty address parts, and
passes Internal Reference through (None when missing/empty).
- save_batch writes a 3-column CSV (user_address,postcode,
internal_reference) to {path_prefix}/{ISO datetime}_{uuid8}.csv
and returns the s3://bucket/key URI.
- Postcode sanitisation flows through UserAddress.__post_init__; the
repo never calls sanitise_postcode directly.
Tests (moto-backed) cover: three-line address load, Address-1-only
load, missing Internal Reference, save->reload round trip, and
unique-filename-per-save. pyright --strict clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
2026-05-19 17:37:02 +00:00 |
|