Add support for the `X-OpenRouter-Categories` header via a new
`app_categories` field on `ChatOpenRouter`, and extract inline client
construction into a dedicated `_build_client` method.
Automated refresh of model profile data for all in-monorepo partner
integrations via `langchain-profiles refresh`.
🤖 Generated by the `refresh_model_profiles` workflow.
Co-authored-by: mdrxy <61371264+mdrxy@users.noreply.github.com>
The `require_issue_link` workflow's maintainer override (reopen PR or
remove `missing-issue-link` to bypass enforcement) has never worked. It
calls `orgs.getMembershipForUser` to verify the sender is an org member,
but `GITHUB_TOKEN` is a GitHub App installation token — not an org
member — so the endpoint always returns 403. The catch block only
handled 404, so the unhandled 403 crashed the entire job, blocking even
the normal issue-link validation from running.
## Changes
- Replace `orgs.getMembershipForUser` with
`repos.getCollaboratorPermissionLevel` in `senderIsOrgMember()` — checks
if the event sender (the user who reopened the PR or removed the label)
has write/maintain/admin access on the repo, which works with
`GITHUB_TOKEN` and is a better proxy for "maintainer" than org
membership
Bump the minimum `langchain-core` dependency to `>=1.2.21` across all 14
partner packages in the monorepo. Aligns partner lower bounds with the
latest core release so consumers pick up recent fixes (notably the
`ModelProfile` schema drift fix from core 1.2.21).
PR #35788 added 7 new fields to the `langchain-profiles` CLI output
(`name`, `status`, `release_date`, `last_updated`, `open_weights`,
`attachment`, `temperature`) but didn't update `ModelProfile` in
`langchain-core`. Partner packages like `langchain-aws` that set
`extra="forbid"` on their Pydantic models hit `extra_forbidden`
validation errors when Pydantic encountered undeclared TypedDict keys at
construction time. This adds the missing fields, makes `ModelProfile`
forward-compatible, provides a base-class hook so partners can stop
duplicating model-profile validator boilerplate, migrates all in-repo
partners to the new hook, and adds runtime + CI-time warnings for schema
drift.
## Changes
### `langchain-core`
- Add `__pydantic_config__ = ConfigDict(extra="allow")` to
`ModelProfile` so unknown profile keys pass Pydantic validation even on
models with `extra="forbid"` — forward-compatibility for when the CLI
schema evolves ahead of core
- Declare the 7 missing fields on `ModelProfile`: `name`, `status`,
`release_date`, `last_updated`, `open_weights` (metadata) and
`attachment`, `temperature` (capabilities)
- Add `_warn_unknown_profile_keys()` in `model_profile.py` — emits a
`UserWarning` when a profile dict contains keys not in `ModelProfile`,
suggesting a core upgrade. Wrapped in a bare `except` so introspection
failures never crash model construction
- Add `BaseChatModel._resolve_model_profile()` hook that returns `None`
by default. Partners can override this single method instead of
redefining the full `_set_model_profile` validator — the base validator
calls it automatically
- Add `BaseChatModel._check_profile_keys` as a separate
`model_validator` that calls `_warn_unknown_profile_keys`. Uses a
distinct method name so partner overrides of `_set_model_profile` don't
inadvertently suppress the check
### `langchain-profiles` CLI
- Add `_warn_undeclared_profile_keys()` to the CLI (`cli.py`), called
after merging augmentations in `refresh()` — warns at profile-generation
time (not just runtime) when emitted keys aren't declared in
`ModelProfile`. Gracefully skips if `langchain-core` isn't installed
- Add guard test
`test_model_data_to_profile_keys_subset_of_model_profile` in
model-profiles — feeds a fully-populated model dict to
`_model_data_to_profile()` and asserts every emitted key exists in
`ModelProfile.__annotations__`. CI fails before any release if someone
adds a CLI field without updating the TypedDict
### Partner packages
- Migrate all 10 in-repo partners to the `_resolve_model_profile()`
hook, replacing duplicated `@model_validator` / `_set_model_profile`
overrides: anthropic, deepseek, fireworks, groq, huggingface, mistralai,
openai (base + azure), openrouter, perplexity, xai
- Anthropic retains custom logic (context-1m beta → `max_input_tokens`
override); all others reduce to a one-liner
- Add `pr_lint.yml` scope for the new `model-profiles` package
Closes#29530
---
Remove a stale BlockBuster allowlist entry in `conftest.py` referencing
`aconfig_with_context` — the function and its containing module
(`langchain_core/beta/runnables/context.py`) were deleted in `fded6c6b1`
(Sep 2025, #32850). Spotted by @antonio-mello-ai in #29530.
Automated refresh of model profile data for all in-monorepo partner
integrations via `langchain-profiles refresh`.
🤖 Generated by the `refresh_model_profiles` workflow.
Co-authored-by: mdrxy <61371264+mdrxy@users.noreply.github.com>
Add a durable maintainer override to the "Require Issue Link" workflow.
The existing maintainer-reopen path skipped enforcement once but didn't
persist that decision — a subsequent PR edit could re-trigger closure.
Maintainers now have two override paths (reopen the PR or remove
`missing-issue-link`), both converging on `applyMaintainerBypass()`
which reopens the PR, cleans up `missing-issue-link`, and applies a
durable `bypass-issue-check` label so future triggers skip enforcement.
## Changes
- Add `unlabeled` to `pull_request_target` trigger types and gate it on
`missing-issue-link` removal + `external` label presence in the
job-level `if`
- Introduce `bypass-issue-check` as a new skip label alongside
`trusted-contributor` — scoped per-PR (not per-author) so maintainers
can override individual PRs without blanket trust
- Extract three helpers in the check-link script: `ensureAndAddLabel`
(idempotent label creation), `senderIsOrgMember` (org membership check),
and `applyMaintainerBypass` (remove label → reopen → add bypass)
- `applyMaintainerBypass` reopens the PR *before* adding the bypass
label so a failed reopen (deleted branch, permissions) leaves a more
actionable state; reopen failure is caught and surfaced via
`core.warning` instead of crashing the step
- Non-member label removal defensively re-adds `missing-issue-link` and
early-returns with failure outputs (re-add failure is non-fatal so the
downstream "Add label" step can retry)
- Replace hardcoded `'langchain-ai'` org in `senderIsOrgMember` with
`context.repo.owner` for portability
- Auto-close comments now include a maintainer override hint: *"reopen
this PR or remove the `missing-issue-link` label to bypass this check"*
- Live-label race guard also checks for `bypass-issue-check`
Automated refresh of model profile data for all in-monorepo partner
integrations via `langchain-profiles refresh`.
🤖 Generated by the `refresh_model_profiles` workflow.
Co-authored-by: mdrxy <61371264+mdrxy@users.noreply.github.com>
Fixes#35811
Updates support for the "azure_ai" provider in both chat model and
embeddings initialization, ensuring consistency across the codebase and
removing the use of deprecated classes `AzureAIChatCompletionsModel` and
`AzureAIEmbeddingsModel`. The changes primarily involve updating
provider mappings and instantiation logic for Azure AI integrations.
**Changes:**
* Updated the chat model initialization logic in `langchain_classic` and
`langchain_v1` to use `AzureAIOpenAIApiChatModel` instead of the
deprecated `AzureAIChatCompletionsModel` for the "azure_ai" provider.
[[1]](diffhunk://#diff-969731928b9d70dac2677539a711478df103ce70024b4c61e95bd7b584bb841dL395-R398)
[[2]](diffhunk://#diff-31b725e3fe9b2f92342af5cda9bf6e61c4bc0b69efd8c25743269f45c52571b2L41-R41)
* Added support for the "azure_ai" provider in embeddings
initialization, mapping it to the new `AzureAIOpenAIApiEmbeddingsModel`
class in both `langchain_classic` and `langchain_v1`.
[[1]](diffhunk://#diff-9490c8d3daa7c5b66d10ce2961f2bc1a6a35005150125c905316adb728bf80d6R9)
[[2]](diffhunk://#diff-9490c8d3daa7c5b66d10ce2961f2bc1a6a35005150125c905316adb728bf80d6R209-R212)
[[3]](diffhunk://#diff-d37aa14c07863cc184800713b88a9c25fdb43efd1a4daa31df3ecabb2f70f177R16)
---------
Co-authored-by: Mason Daugherty <github@mdrxy.com>
Co-authored-by: Mason Daugherty <mason@langchain.dev>
When a maintainer manually reopens a PR that was auto-closed by the
`require-issue-link` workflow, skip enforcement so it stays open. Scoped
to PRs carrying the `missing-issue-link` label (i.e. only those closed
by this workflow). Non-org-members reopening their own PRs still go
through normal enforcement.
A PR that only touches `uv.lock` currently gets the label of its' dir
because the file rule matches on the file prefix. This is misleading —
lockfile-only changes aren't meaningful package changes. The
`excludedFiles` list already existed in config (for size calculations),
but file rules didn't consult it.
## Changes
- Add `skipExcludedFiles` option to file rules in
`pr-labeler-config.json`, enabled for the four package rules
(`deepagents`, `cli`, `acp`, `evals`) so lockfile-only PRs don't trigger
package labels
- `matchFileLabels` in `pr-labeler.js` now filters out files whose
basename appears in the top-level `excludedFiles` list (currently just
`uv.lock`) before testing rules that opt in via `skipExcluded`
- Non-package rules (`github_actions`, `dependencies`) are unaffected —
they don't set the flag
Automated refresh of model profile data for all in-monorepo partner
integrations via `langchain-profiles refresh`.
🤖 Generated by the `refresh_model_profiles` workflow.
Co-authored-by: mdrxy <61371264+mdrxy@users.noreply.github.com>
Reduce CI log noise by suppressing pytest's per-test dot/verbose
streaming output. The `_test.yml` workflow now passes `PYTEST_EXTRA=-q`
to `make test`, which overrides the default verbosity with quiet mode —
failures still print in full, but the thousands of `.......` progress
lines are gone. Local `make test` is unaffected since `PYTEST_EXTRA`
defaults empty.
## Changes
- Add `PYTEST_EXTRA ?=` variable to all 21 package Makefiles and inject
it into each `test` target's pytest invocation
- Pass `PYTEST_EXTRA=-q` in `_test.yml` for both the main test step and
the min-version retest step
`--no-group test --no-group dev --no-group lint` only controls which
groups get *installed*; uv still resolves the full dependency graph (all
groups) and tries to generate metadata for editable sources like
`../langchain_v1`. In a sparse checkout these paths don't exist.
`--frozen` skips re-resolution entirely and uses the existing lock, so
the missing editable paths are never accessed.
The reusable refresh_model_profiles workflow sparse-checks out only
libs/model-profiles from the langchain monorepo. `uv sync` fails because
the test/dev/lint dependency groups reference sibling editable packages
(../langchain_v1, ../core) that aren't present in the sparse checkout.
Restrict to the default dependency group so only the runtime deps
(httpx, tomli, typing-extensions) are installed — which is all the CLI
needs.
`pr-labeler.js` used `require('@actions/core')` to access GitHub Actions
logging/failure helpers, but that module is bundled inside
`actions/github-script`'s dist — it's not resolvable via Node's
`require()` from a checked-out file on disk. Two of the three call sites
were in rarely-hit error branches, so the bug was latent. The third
(`applyTierLabel`) ran unconditionally, crashing the tier-label step on
every external PR. Because the tier step runs *before* the "add external
label" step, the crash prevented the `external` label from ever being
applied — which meant `require_issue_link.yml` never triggered and
unapproved external PRs stayed open.
## Changes
- Thread the `core` object (provided by `actions/github-script` at eval
time) through `loadAndInit()` → `init()` instead of calling
`require('@actions/core')` from the checked-out script — fixes the
`MODULE_NOT_FOUND` crash on all three call sites (`ensureLabel`,
`getContributorInfo`, `applyTierLabel`)
- Add a console-based fallback in `loadAndInit` so callers that don't
need `core.setFailed` still work without passing it
- Update all 9 `loadAndInit(github, owner, repo)` call sites across
`pr_labeler.yml`, `pr_labeler_backfill.yml`, and
`tag-external-issues.yml` to pass `core`
Remove `opened` from the `require_issue_link.yml` trigger list. New PRs
never have labels at creation time, so the job-level `if` (which
requires the `external` label) always evaluated to skip — producing a
noisy "Skipped" check on every internal PR. Enforcement chains through
`pr_labeler.yml` → `labeled` event, making `opened` a no-op.
## Changes
- Drop `opened` from `pull_request_target.types` in
`require_issue_link.yml` — only `edited`, `reopened`, and `labeled`
remain
- Update header comment to document why `opened` is intentionally
excluded
CodSpeed benchmarks shared the main CI workflow's concurrency group,
which has `cancel-in-progress: true`. On push-to-master (post-merge),
back-to-back merges would cancel in-flight CodSpeed jobs — failing the
`ci_success` gate and marking the merge commit red, even though the PR
passed all checks before merging. The cancelled jobs also meant baseline
benchmark data was never uploaded, silently breaking CodSpeed's
regression detection for subsequent PRs.
## Changes
- Extract CodSpeed into a standalone `codspeed.yml` workflow with its
own concurrency group — push-to-master runs use `github.sha` as the
group key (unique per commit, never cancelled), while PR runs continue
to cancel stale runs via `github.ref`
- Remove the `codspeed` job, its output, and its `ci_success` gate
dependency from `check_diffs.yml` — benchmark results are informational
and shouldn't block merges
## Review
- The new workflow duplicates the `build` (change detection) job from
`check_diffs.yml` since it can't reference jobs across workflows. This
means an extra `check_diff.py` run per trigger, but it's lightweight
(~2s) and the tradeoff is worth the isolation.
Speed up CodSpeed benchmarks for partners with heavy SDK inits by
switching them to walltime mode. `fireworks` takes ~328s and `openai` ~6
min under CPU simulation (Valgrind-based) — walltime is noisier but more
than adequate for detecting init-time regressions on these packages.
## Changes
- Add `CODSPEED_WALLTIME_DIRS` set in `_get_configs_for_single_dir` that
routes `libs/core`, `libs/partners/fireworks`, and
`libs/partners/openai` to walltime mode; all other partners default to
`simulation`
- Emit a `codspeed-mode` field in the CodSpeed matrix config and consume
it as `${{ matrix.job-configs.codspeed-mode }}` in the workflow,
replacing the inline ternary
- Replace deprecated `instrumentation` runner mode with `simulation` in
CodSpeed benchmarks (purely a rename, identical behavior)
- Scope partner benchmarks to `tests/unit_tests/ -m benchmark` instead
of `tests/` — avoids collecting ~80-90 non-benchmark tests per partner
just to deselect them
- Integration benchmark (`test_stream_time`) is always skipped anyway
(no VCR cassettes in CI), so no loss from excluding `integration_tests/`
Resolves: `Warning: The 'instrumentation' runner mode is deprecated and
will be removed in a future version. Please use 'simulation' instead.`
CI lint jobs use `uv run --all-groups` for all tools, but ruff doesn't
need dependency resolution — only mypy does. By splitting into
`UV_RUN_LINT` (ruff) and `UV_RUN_TYPE` (mypy), the CI-facing targets run
ruff with `--group lint` only, giving fast-fail feedback before mypy
triggers the full environment sync.
For packages where source code only conditionally imports heavy deps
(text-splitters, huggingface), `lint_package` also overrides
`UV_RUN_TYPE` to `--group lint --group typing`, skipping the ~3.5GB
`test_integration` download entirely. `lint_tests` keeps `--all-groups`
since test code legitimately imports those deps.
Additionally, `lint_imports.sh` was inconsistently wired — most packages
had the script but weren't calling it.
## Changes
**Makefile optimization**
- Introduce `UV_RUN_LINT` and `UV_RUN_TYPE` Make variables, both
defaulting to `uv run --all-groups`. For `lint_package` and
`lint_tests`, `UV_RUN_LINT` is overridden to `uv run --group lint` so
ruff runs instantly without syncing heavy deps
- For `text-splitters` and `huggingface`, override `UV_RUN_TYPE` on
`lint_package` to `uv run --group lint --group typing` — mypy runs
without downloading torch, CUDA, spacy, etc.
**mypy config for lean groups**
- Add `transformers` and `transformers.*` to `ignore_missing_imports` in
`text-splitters` pyproject.toml (conditional `try/except` import, same
treatment as existing `konlpy`/`nltk` entries)
- Add `torch`, `torch.*`, `langchain_community`, `langchain_community.*`
to `ignore_missing_imports` in `huggingface` pyproject.toml
- Add dual `# type: ignore[unreachable, unused-ignore]` in
`text-splitters/base.py` to handle the `PreTrainedTokenizerBase`
isinstance check that behaves differently depending on whether
transformers is installed
**lint_imports.sh consistency**
- Add `./scripts/lint_imports.sh` to the lint recipe in every package
that wasn't calling it (standard-tests, model-profiles, all 15
partners), and create the script for the two packages missing it
entirely (`model-profiles`, `openrouter`)
- Update all `lint_imports.sh` scripts to allow `from langchain.agents`
and `from langchain.tools` imports (legitimate v1 middleware
dependencies used by `langchain-anthropic` and `langchain-openai`)
Consolidate four separate PR labeling workflows (`pr_labeler_file.yml`,
`pr_labeler_title.yml`, `pr_size_labeler.yml`, and the PR-handling half
of `tag-external-contributions.yml`) into a single `pr_labeler.yml`
workflow. The old workflows raced against each other — concurrent label
mutations could drop or duplicate labels depending on execution order. A
unified workflow with concurrency grouping eliminates that class of bug.
Automated refresh of model profile data for all in-monorepo partner
integrations via `langchain-profiles refresh`.
🤖 Generated by the `refresh_model_profiles` workflow.
Co-authored-by: mdrxy <61371264+mdrxy@users.noreply.github.com>
Extract the model profile refresh logic into a reusable `workflow_call`
workflow so external repos like `langchain-google` and `langchain-aws`
can run the same daily profile refresh and get auto-PRs without
duplicating the pipeline. The in-monorepo caller becomes a thin wrapper
passing provider JSON.
## Changes
- Add `_refresh_model_profiles.yml` as a reusable `workflow_call`
workflow — accepts a `providers` JSON array of `{provider, data_dir}`
pairs, optional `cli-path` (skips cloning the CLI repo when the caller
already has it), and configurable PR metadata inputs
- External callers get the `langchain-profiles` CLI via sparse checkout
of `langchain-ai/langchain` at a configurable `cli-ref`; the in-monorepo
caller short-circuits with `cli-path: libs/model-profiles`
- Add input validation step using `jq` — rejects non-array JSON and
entries missing `provider`/`data_dir` keys with `::error::` annotations
- Replace the piped `while read` loop with `mapfile`/`for` +
per-provider error handling: one provider failure no longer kills the
rest, and all failures are collected and reported at the end
- Route all `${{ inputs.* }}` expressions through `env:` bindings in
`run:` blocks to prevent script injection from caller-controlled values
- Validate `cli-path` existence before use, with a clear error if the
directory is missing
- Summary step now runs with `if: always()` and handles
failure/success/no-op states separately
- Refactor `refresh_model_profiles.yml` into a thin caller that passes
the 10 in-monorepo providers as JSON
