[keith991001]

Important

1. Read CONTRIBUTING.md.
2. Associated issue: rewrite property to accept session: Session as parameter #40372 (claimed the slice in this comment).

Summary

Part of #40372, follow-up to #41370 / #41440.

Removes five legacy @property accessors on Dataset in api/models/dataset.py that still reach for the global db.session internally. Each of them is a thin wrapper around an existing session-aware get_* twin, and every call site has already been migrated to the twins (DatasetDetailResponseSource from fields/dataset_fields.py covers the response-validation reads for all dataset detail endpoints):

• Dataset.total_documents → get_total_documents(session=...)
• Dataset.total_available_documents → get_total_available_documents(session=...)
• Dataset.author_name → get_author_name(session=...)
• Dataset.latest_process_rule → get_latest_process_rule(session=...)
• Dataset.external_knowledge_info → get_external_knowledge_info(session=...)

I verified there are no remaining production reads of these five as properties (including implicit from_attributes=True reads — all three dataset-detail validation sites go through dataset_detail_response_source).

Not touching the other Dataset properties (some still have property readers), nor Document.* / DocumentSegment.* / the slices claimed by other contributors on the issue.

Tests

The session-aware get_* twins already have unit coverage in tests/unit_tests/models/test_dataset_models.py (test_session_aware_dataset_getters_use_caller_session, test_dataset_detail_getters_use_caller_session), which keeps passing. Updated the two container-integration tests that still read total_documents / total_available_documents as properties to call the get_* twins explicitly.

Screenshots

N/A (backend-only refactor).

Checklist

• This change requires a documentation update, included: Dify Document
• I understand that this PR may be closed in case there was no previous discussion or issues. (This doesn't apply to typos!)
• I've added a test for each change that was introduced, and I tried as much as possible to make a single atomic change.
• I've updated the documentation accordingly.
• I ran make lint && make type-check (backend) and vp staged (frontend) to appease the lint gods

From Claude Code

[github-actions]

Pyrefly Type Coverage

Metric	Base	PR	Delta
Type coverage	61.78%	61.77%	-0.00%
Strict coverage	61.38%	61.37%	-0.00%
Typed symbols	42,994	42,989	-5
Untyped symbols	26,776	26,776	0
Modules	3290	3290	0

[keith991001]

CI caught a fourth validation site my call-site sweep missed: CreateEmptyRagPipelineDatasetApi validates DatasetDetailResponse from a raw Dataset indirectly through the dump_response helper, so it wasn't found when grepping for model_validate. The three removed detail fields (author_name, total_documents, total_available_documents) then came up missing.

Fixed in 81ffc00 by routing that endpoint through dataset_detail_response_source(dataset, session=session) — the same wrapper the other three dataset-detail endpoints already use, reusing the session the endpoint already creates. Swept the repo for any other dump_response(DatasetDetail*, <raw dataset>) sites: this was the only one.