Add support for published consumption readings

[WilliamCesbron]

Summary

Add explicit support for GRDF published consumption readings without changing the existing informative-reading default.

Context

Some PCEs return no usable history (or only the latest reading) from the informatives endpoint even though published consumption history is available in the GRDF customer portal. This is the situation reported in ssenart/gazpar2haws#117.

PyGazpar already models the publiees endpoint through ConsumptionType.PUBLISHED, but JsonWebDataSource always requests informative readings. In addition, published readings have a null journeeGaziere and therefore fail in JsonParser.

Changes

• allow callers to select the ConsumptionType used by JsonWebDataSource;
• preserve ConsumptionType.INFORMATIVE as the default;
• export ConsumptionType from the public package API;
• parse published readings by assigning their consumption to the last day of the reporting period (dateFinReleve - 1 day);
• ignore malformed readings that provide neither journeeGaziere nor dateFinReleve;
• add regression tests using the existing donnees_publiees.json fixture.

Important behavior

Published readings may cover more than one day. This change does not invent a daily distribution. The complete consumption is assigned to the final day covered by the published period, preserving the GRDF total while making the temporal granularity explicit.

Validation

• pytest tests/test_jsonparser.py tests/test_datafileparser.py -q — 7 passed
• Black check — passed
• isort check — passed
• Flake8 on changed Python files — passed

[ssenart]

Hello, thanks for your contribution.

I will publish an alpha release shortly.

[ssenart]

Done, draft release available here: https://pypi.org/project/pygazpar/1.4.0a1/

Thanks a lot !