Add aug 21 release

[shuchenliu]

add more benchmark testing for Publications queries

[Copilot]

Pull request overview

Adds a new bulk-search lookup strategy for the PubMed /publications endpoint and updates the benchmarking harness to compare current vs bulk-search, along with expanded unit/integration tests and refreshed documentation.

Changes:

• Implement bulk-search alternative-identifier resolution (single terms _search + safe reverse-mapping with fallback to exact per-ID _msearch).
• Add Elasticsearch client helpers: querymany_exact(...) (root-field term _msearch) and search_terms(...) (bounded terms _search with hit-total reporting).
• Update benchmark CLI/tests/docs to replace the prior two-phase experiment with bulk-search.

Reviewed changes

Copilot reviewed 13 out of 13 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
tests/test_publications_load.py	Updates benchmark harness tests to use bulk-search strategy and labels.
tests/test_elasticsearch.py	Adds coverage for querymany_exact request shape and search_terms behavior/result formatting.
tests/test_document_metadata.py	Reworks metadata service tests for bulk-search, adds fallback/ordering/concurrency cases, and expands live integration coverage across strategies.
README.md	Documents the new bulk-search experiment, updated index-shape expectations, and refreshed benchmark instructions.
biothings_annotator/webapp/openapi.json	Updates endpoint/schema descriptions to reflect alternate identifiers living in pubmed.identifiers.
biothings_annotator/application/views/document_metadata.py	Switches header-driven A/B selector from two-phase to bulk-search using strategy constants.
biothings_annotator/annotator/settings.py	Removes the deprecated DOCUMENT_METADATA_TWO_PHASE_LOOKUP setting.
biothings_annotator/annotator/elasticsearch.py	Introduces querymany_exact and search_terms plus shared exact-batch helper.
biothings_annotator/annotator/document_metadata.py	Implements bulk-search strategy, strategy selection, and safer reverse-mapping with fallback.
benchmarks/publications/workload.py	Updates supported lookup strategies to include bulk-search.
benchmarks/publications/runner.py	Renames paired observation arm and comparison logic from two-phase to bulk-search.
benchmarks/publications/report.py	Updates reporting keys/labels and delta semantics to bulk-search minus current.
benchmarks/publications/main.py	Updates CLI help text to reflect bulk-search paired comparisons.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.