`docs/src/python/python.md` is the whole Python API reference, but it is
maintained by hand and had drifted from the public API. Anything not
listed there simply doesn't get rendered, so a number of public,
documented, tested APIs were invisible to users — most notably branch
management, where `diff` and `merge` live.
I audited every public symbol reachable from `lancedb` and its
subpackages against the `:::` directives on the page. This adds the
missing ones:
- **Branching** — `Branches`, `AsyncBranches` (`list` / `create` /
`checkout` / `delete` / `diff` / `merge`)
- **Tables** — `TableStatistics` (returned by `Table.stats()`; the
fragment-level stats classes were already listed)
- **Full text queries** — `FullTextQuery`, `MatchQuery`, `PhraseQuery`,
`BoostQuery`, `MultiMatchQuery`, `BooleanQuery`, `FullTextOperator`,
`Occur`
- **Querying** — `LanceEmptyQueryBuilder`, `LanceTakeQueryBuilder`,
`AsyncTakeQuery`
- **Indices** — `Fm` (the FM-index for substring search), `IndexConfig`
- **Blobs** — `blob`, `BlobType`, `BlobFile`
- **Namespaces** — `connect_namespace`, `connect_namespace_async`, and
both namespace connection classes
- **Remote config** — `TlsConfig`, `HeaderProvider`, `OAuthConfig`,
`OAuthFlowType`
- **Rerankers** — the `Reranker` base class plus `JinaReranker`,
`RRFReranker`, `MRRReranker`, `AnswerdotaiRerankers`,
`VoyageAIReranker`, `WatsonxReranker` (5 of 12 were listed)
- **Embeddings** — `get_registry`, `register`, and the 14 embedding
functions that were missing (3 of 17 were listed)
- **PyTorch** — `StreamingDataset` and the permutation API it is built
on
- **Misc** — `Session`, `tokenize`, `FtsToken`, `pydantic.Vector`,
`pydantic.MultiVector`, `instrument_lancedb_metrics`, and the two
exception types
It also repairs cross-references in docstrings that no longer resolve:
links into guide pages that have since moved to lancedb.com
(`querying-an-ann-index`, `experimental-full-text-search`),
`lance.dataset` references with no inventory behind them, and the
relative targets `[Table](Table)` and `[PyArrow Table](pyarrow.Table)`.
Deliberately left out: concrete implementation classes reached through
their abstract base (`LanceTable`, `LanceDBConnection`,
`RemoteDBConnection`), query base classes already covered by
`inherited_members: true`, and internal plumbing such as
`FullTextSearchQuery` and `ColumnOrdering`.
## Testing
The docs job only runs on pushes to `main`, so I built the site locally
and compared against a build of `upstream/main`: every added entry
resolves, and no symbol that was rendered before stopped being rendered
when the four packages moved to automodule. `mkdocs build --strict`
exits 0 on this branch, against 61 warnings on `main`.
## Also in this PR
`lancedb.index`, `lancedb.embeddings`, `lancedb.remote` and
`lancedb.rerankers` are now rendered by a single mkdocstrings directive
each, driven by the module's `__all__`, rather than a hand-maintained
list. These four are where most of the drift was, and `__all__` is
harder to forget than a docs page. `lancedb.embeddings` had no
`__all__`; without one mkdocstrings renders no members at all for a
re-export package, so one is added. AGENTS.md gains a section on how the
page is wired up and how to build the docs locally.
Rendering all that code for the first time surfaced ~100 more build
warnings, which would have made #3707 (turning on `mkdocs build
--strict`) harder to land, so the warning backlog is cleared here too.
97 of the 158 warnings were one systematic false positive — griffe
cannot see the generated `__init__` of a pydantic dataclass, so every
documented parameter looks unknown — switched off via
`warn_unknown_params`. The remaining 61 came from 15 docstrings with
real bugs: prose trailing a `Parameters` section (we were rendering
parameters called `The`, `you` and `To`), types dropped because numpydoc
needs spaces around the colon, `num_partitions, default sqrt(num_rows)`
parsing as a list of names and inventing a `default` parameter, and one
parameter indented five spaces. `mkdocs build --strict` now exits 0.
---
#3747 (the coverage test that keeps this from happening again) is
stacked on this branch, so review it after this one.
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
## Summary
Switch the nodejs bindings and examples package from npm to pnpm 11 to
pick up its stronger supply-chain defaults:
- `minimumReleaseAge` defaults to 1 day, so newly-published (potentially
compromised) versions aren't resolved into installs for at least 24h.
- Install lifecycle scripts (`preinstall`/`install`/`postinstall`) are
no longer run for arbitrary transitive deps; only an explicit allowlist
may run them, and unapproved scripts cause install to fail
(`strictDepBuilds: true`).
- Audit uses GHSA IDs and `--fix=update` to add patched versions to
`minimumReleaseAgeExclude`.
This is the same class of protection that would have blunted the recent
TanStack/`@uipath`/etc. compromise discussed in the [Aikido
write-up](https://www.aikido.dev/blog/mini-shai-hulud-is-back-tanstack-compromised).
## Changes
- Replace `nodejs/package-lock.json` and
`nodejs/examples/package-lock.json` with `pnpm-lock.yaml`.
- Pin pnpm via `packageManager: pnpm@11.1.1` in both `package.json`s.
- Add `pnpm-workspace.yaml` with the four build-script packages we
actually need: `@biomejs/biome`, `onnxruntime-node`, `protobufjs`,
`sharp`. Everything else is blocked from running install scripts.
- Update package.json scripts (`npm run X` → `pnpm X`).
- Update workflows: `.github/workflows/nodejs.yml`,
`.github/workflows/npm-publish.yml`, and
`.github/workflows/codex-fix-ci.yml` — install pnpm via
`pnpm/action-setup@v4` and switch `setup-node` caches to
`pnpm-lock.yaml`.
- Refresh `nodejs/AGENTS.md`, `nodejs/CLAUDE.md`, and
`nodejs/CONTRIBUTING.md`.
`docs/package-lock.json` is **not** touched — out of scope for this PR.
## Test plan
- [ ] `Lint` job (lint Rust/TS + examples lint) passes on CI.
- [ ] `Linux (NodeJS 18/20)` build+test passes, including the examples
test step.
- [ ] `macos` build+test passes.
- [ ] `NPM Publish` workflow's PR dry-run completes (build matrix + test
matrix + dry `npm publish`).
- [ ] No new install-script approvals are required at install time.
## Follow-ups
- `update_package_lock_run_nodejs.yml` references a composite action
path that doesn't exist
(`./.github/workflows/update_package_lock_nodejs`); it was already
broken pre-PR. We may want to either delete this workflow or rewrite it
for pnpm in a follow-up.
- Consider migrating `docs/` to pnpm in a separate PR.
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* Add `ci` profile for smaller build caches. This had a meaningful
impact in Lance, and I expect a similar impact here.
https://github.com/lancedb/lance/pull/5236
* Get caching working in Rust. Previously was not working due to
`workspaces: rust`.
* Get caching working in NodeJs lint job. Previously wasn't working
because we installed the toolchain **after** we called `- uses:
Swatinem/rust-cache@v2`, which invalidates the cache locally.
* Fix broken pytest from async io transition
(`pytest.PytestRemovedIn9Warning`)
* Altered `get_num_sub_vectors` to handle bug in case of 4-bit PQ. This
was cause of `rust future panicked: unknown error`. Raised an issue
upstream to change panic to error:
https://github.com/lancedb/lance/issues/5257
* Call `npm run docs` to fix doc issue.
* Disable flakey Windows test for consistency. It's just an OS-specific
timer issue, not our fault.
* Fix Windows absolute path handling in namespaces. Was causing CI
failure `OSError: [WinError 123] The filename, directory name, or volume
label syntax is incorrect: `
This PR does the following:
- Pare down the docs to only what's needed (Python, JS/TS API docs and a
pointer to Rust docs)
- Styling changes to be more in line with the main website theme
The relative URLs remain unchanged, so assuming CI passes, there should
be no breaking changes from the main docs site that points back here.
Will JonesandClaude Opus 5
Prashanth Rao