[Paul Lizer (paullizer)]

Summary

• V2 discovery and search-cache setup still enumerated source containers even when the maintained document access index (DAI) could serve them. Build on the restored DAI-first source picker and extend indexed discovery to named group/public browsers and Assigned Knowledge, preserving scoped source fallback, current authorization, screening, and pending-publication checks.
• Carry file size and legacy/XSD identity through the projection, qualify revision families by their owner, and remove the editor's 1,000-document display cap from runtime Assigned Knowledge resolution. Inventory responses disclose bounded display counts.
• Read all-revision cache fingerprints from usable DAI partitions, skip fingerprinting when result caching is disabled, and check valid cache hits before repeated revision repair. Reuse existing generation invalidation rather than adding a new freshness/completeness system.
• Orchestration planning and document-search execution already share this backend, so no separate DAI layer is added. Restore the Document Search action's missing service import and add real-owner/SDK regression coverage and documentation.

Linked issue

N/A. No GitHub issue was requested for this work.

Release Notes & Latest Features

• New Feature
• Bug Fix
• UI Enhancement
• Breaking Change
• Internal only

Is this visible to end users?

• Yes
• No

Is this admin-facing (Admin Settings, governance, deployment, config)?

• Yes
• No

Only the existing DAI maintenance/backfill workflow is affected; no new setting or container is introduced.

Should this become a Latest Feature card?

• Yes
• No
• Already added

Screenshot needed for the card?

• Yes
• No
• Attached

Version bump

• application/single_app/config.py VERSION third segment bumped, or not needed because this is docs-only
• deployers/version.txt bumped, or not needed because deployers/ was not changed

Application version is 0.261.407, following the base branch's 0.261.406. The existing 0.261.404 source-menu fix is now in the base. This PR adds no deployer-file changes relative to that base.

Rollout: DAI projection schema advances to 4. Existing scoped source queries remain the fallback until the normal backfill/repair workflow makes the updated projection ready. Healthy empty index results are trusted; there is no full-source verification scan or rebuild-on-read.

Testing / validation

Network-blocked runs used the application and functional-test directories on PYTHONPATH, with the pinned tokenizer assets prepared as in the existing Cosmos CI workflow.

After synchronizing base 860c27c84:

• python -m pytest -q --disable-warnings --tb=short -p test_support.cosmos_network_guard functional_tests\test_v2_document_index_coverage.py functional_tests\test_v2_document_source_selection.py functional_tests\test_v2_document_source_discovery_imports.py functional_tests\test_v2_document_source_runtime.py functional_tests\test_orchestration_context_picker.py functional_tests\test_orchestration_search_result_integrity.py - 95 passed.
• python -m pytest -q --disable-warnings --tb=short -p test_support.cosmos_network_guard functional_tests\test_action_credentials_bootstrap.py functional_tests\test_stale_active_group_recovery.py functional_tests\test_m365_cosmos_sdk_contract.py::test_real_sdk_point_documents_are_wrapped_but_nested_json_and_query_rows_are_plain - 16 passed.
• python .\scripts\check_cosmos_response_boundaries.py - 11 inherited base review errors remain across nine untouched owners in functions_settings.py and deployers/bicep/postconfig.py. The base itself already has 39 stale owner reviews. The combined chat module's 30 reviews were renewed after checking identical storage-function ASTs and point/query/batch consumers against both parents, preserving real-owner runtime evidence. The unrelated settings/postconfiguration reviews and missing runtime evidence were not repaired or exempted in this PR.
• python .\functional_tests\test_docs_app_surface_coverage.py - 7/7 passed, including inventory freshness.
• python .\functional_tests\test_docs_site_quality.py - 6/6 passed.

Original implementation validation, before base synchronization:

• python -m pytest -q --disable-warnings --tb=short -p test_support.cosmos_network_guard functional_tests\test_v2_document_index_coverage.py functional_tests\test_cosmos_wave5a_document_access_read_switch.py functional_tests\test_cosmos_wave6_document_access_cache.py functional_tests\test_group_document_read_apis.py functional_tests\test_public_document_read_apis.py functional_tests\test_v2_document_source_selection.py functional_tests\test_v2_document_source_discovery_imports.py - 427 passed.
• python -m pytest -q --disable-warnings --tb=short -p test_support.cosmos_network_guard functional_tests\test_orchestration_internal_analysis.py functional_tests\test_orchestration_retained_analysis_publication.py functional_tests\test_orchestration_search_result_integrity.py - 101 passed.
• python -m pytest -q --disable-warnings --tb=short -p test_support.cosmos_network_guard functional_tests\test_orchestration_context_picker.py functional_tests\test_v2_chat_context_picker.py - 22 passed.
• python .\scripts\check_cosmos_response_boundaries.py --run-contracts - 0 review gaps; 1,934 tests and 431 subtests passed, with 14 initial Analyze fixture setup errors. The missing get_cache_settings test seam was then corrected; all 14 cases pass in the clean 101-test follow-up above. The complete sweep was not rerun after that fixture-only correction.
• git -c core.whitespace=blank-at-eol,blank-at-eof,space-before-tab,cr-at-eol --no-pager diff HEAD --check - passed, including after base synchronization.

A wider Analyze-fixture selection recorded 563 passed and seven unrelated failures. All seven reproduce on an isolated, unchanged b6c1c8acd snapshot: six saved-analysis report tests patch a removed token-limit helper, and one workflow retry test omits completeness_policy from its extracted namespace. Those pre-existing harness issues were not changed. Details and base-sync validation are recorded in docs/explanation/fixes/V2_DOCUMENT_ACCESS_INDEX_COVERAGE_FIX.md.

No live Azure RU or latency benchmark was performed.

Documentation

• Release notes updated, or not needed
• Feature documentation updated, or not needed
• Fix documentation updated, or not needed

Release notes were offered but deferred without approval. The source-picker, group/public read and browsing, Assigned Knowledge, and fix documentation reflect the implemented behavior and rollout.

Security checklist

• New Flask routes include @swagger_route(security=get_auth_security())
• Settings sent to non-admin frontends use sanitize_settings_for_user()
• Browser JavaScript is served from local SimpleChat static assets only; no CDN-hosted JS
• No secrets, keys, connection strings, or local-only artifacts are included

No new Flask routes, settings exports, or browser assets are introduced. DAI is a discovery/read model, not an authorization grant; current scoped source hydration and final response checks remain in place.