Skip to content

feat(indexing)!: fold per-format inline tools into index-documents - #197

Open
adityamparikh wants to merge 3 commits into
apache:mainfrom
adityamparikh:feat/index-documents-single-tool
Open

feat(indexing)!: fold per-format inline tools into index-documents#197
adityamparikh wants to merge 3 commits into
apache:mainfrom
adityamparikh:feat/index-documents-single-tool

Conversation

@adityamparikh

Copy link
Copy Markdown
Contributor

Summary

Replaces the four inline indexing tools (index-json-documents, index-csv-documents, index-xml-documents, index-markdown-documents) with one tool:

index-documents(collection, content, format)

format is required and accepts json, csv, xml, markdown or the alias md, case-insensitively. The four per-format methods stay as Java entry points; the new tool dispatches to them.

Why one tool

  • Same shape as file ingestion. feat(indexing): stream local JSON, CSV, XML and Markdown files #194 adds index-file(collection, path, format). With this change inline and file indexing share one calling convention instead of one tool plus four differently named siblings.
  • Smaller tool catalog. Tool schemas are re-sent every session; three near-identical schemas go away.
  • One home for guidance. Schema-first, sanitisation and payload advice live in one description.
  • Explicit format, not sniffing. Inline payloads have no filename, and CSV and Markdown are both plain text with no safe distinguishing prefix, so the model sets a four-value argument instead of the server guessing. index-file keeps extension inference because a filename is real signal.

What changed

  • IndexingService: new @McpTool index-documents; the per-format methods lose their @McpTool; resolveIndexTool becomes normalizeFormat; the index-data prompt tells the model to call index-documents with the matching format.
  • README tool table and prompt row, THREAT_MODEL XML note, observability test README.
  • Tests: McpToolRegistrationTest asserts index-documents is the only index-* tool and that all three arguments are required; IndexingServiceTest covers each format, the md alias, case and whitespace normalisation, and rejection of unknown or missing formats before Solr is touched; McpClientIntegrationTestBase and SampleClient call the new tool over MCP and assert the old names are gone.

Breaking change

Clients and saved prompts that call the four per-format tools must call index-documents with a format argument. At 0.0.2-SNAPSHOT this is the cheapest point to make the rename.

Not done here

A typed JSON documents array (native JSON in the tool call instead of an escaped string) would save tokens for JSON payloads. It needs a second parameter or a second tool and is left for a follow-up.

Verification

./gradlew build on Java 25: 411 tests, 0 failures, 7 skipped (the OTLP suite, skipped on main too). All Testcontainers-based integration classes ran. Spotless applied.

Related

🤖 Generated with Claude Code

https://claude.ai/code/session_019zJ9WA8nNyA5Yxb8ueM7vV

adityamparikh and others added 3 commits September 12, 2026 12:39
Replace index-json-documents, index-csv-documents, index-xml-documents and
index-markdown-documents with a single index-documents tool that takes
collection, content and a required format (json, csv, xml, markdown or the
md alias; case-insensitive). The four per-format methods remain as Java
entry points used by tests and by the new tool's dispatch.

Why one tool: it mirrors the shape of file ingestion (collection, payload,
format) so clients learn one calling convention, keeps a single home for
indexing guidance, and removes three near-identical schemas from every
session's tool catalog. The format is an explicit argument rather than
sniffed: inline payloads have no filename, and CSV and Markdown are both
plain text with no safe distinguishing prefix.

The index-data prompt now instructs the model to call index-documents with
the matching format. README, THREAT_MODEL and the observability test README
are updated; the MCP client, sample client, registration and unit tests
cover the new tool and assert the old names are gone.

BREAKING CHANGE: MCP clients and prompts that call the four per-format
indexing tools must call index-documents with a format argument.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019zJ9WA8nNyA5Yxb8ueM7vV
Signed-off-by: Aditya Parikh <aditya.m.parikh@gmail.com>
Match the text-block style SearchService already uses instead of a
concatenated string.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019zJ9WA8nNyA5Yxb8ueM7vV
Signed-off-by: Aditya Parikh <aditya.m.parikh@gmail.com>
The four per-format methods are now only self-invoked from indexDocuments,
so their @PreAuthorize never ran; the gate on the tool method is the real
one. Also: no commit stubs in the new format tests (mocks return null
anyway, and strict stubs would flag them once the commit overload changes),
prompt tests renamed to match what they assert, one duplicated description
assertion removed (SampleClient covers it), the tool javadoc halved, and
THREAT_MODEL's tool counts updated for the single index tool.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CiUHyyXLTo9ATdgg8eRFZJ
Signed-off-by: Aditya Parikh <aditya.m.parikh@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant