Expose configurables through RemoteRunnable #47

New Issue

2026-02-16T00:18:18-05:00

yindo commented

2026-02-16 00:18:18 -05:00

Originally created by @Vinno97 on GitHub (Nov 23, 2023).

For local runnables, you can use runnable.config_specs to get a list of all configurable fields. The RemoteRunnable doesn't expose these, even though it technically has the correct information via /config_schema endpoint. The same can be said for Input and Output with input_schema and output_schema, respectively.

Loading these endpoints (lazily) would enable better compatibility with applications that may rely on this. In my case, I automatically expose some of my chain's options through a UI when they're configurable. This now doesn't work when I use langserve.

Originally created by @Vinno97 on GitHub (Nov 23, 2023). For local runnables, you can use `runnable.config_specs` to get a list of all configurable fields. The `RemoteRunnable` doesn't expose these, even though it technically has the correct information via `/config_schema` endpoint. The same can be said for `Input` and `Output` with `input_schema` and `output_schema`, respectively. Loading these endpoints (lazily) would enable better compatibility with applications that may rely on this. In my case, I automatically expose some of my chain's options through a UI when they're configurable. This now doesn't work when I use langserve.

yindo commented

2026-02-16 00:18:19 -05:00

@Vinno97 commented on GitHub (Nov 23, 2023):

I don't know of an elegant way to map the "annotations" field back to Python types, but this seems to work for my chains:

def json_annotation_to_type(annotation: str) -> type:
    if annotation == "string":
        return str
    elif annotation == "boolean":
        return bool
    elif annotation == "integer":
        return int
    elif annotation == "number":
        return float
    elif annotation == "array":
        return list
    raise NotImplementedError(f'Can\'t convert annotation "{annotation}" to type')


class ConfigurableRemoteRunnable(RemoteRunnable):
    @property
    def config_specs(self) -> list[ConfigurableFieldSpec]:
        if getattr(self, "_config_specs", None) is None:
            schema = self.sync_client.get("/config_schema").json()

            fields = schema["definitions"]["Configurable"]["properties"]
            self._config_specs = []
            for field_id, field_props in fields.items():
                self._config_specs.append(
                    ConfigurableFieldSpec(
                        id=field_id,
                        name=field_props["title"],
                        description=field_props["description"],
                        default=field_props.get("default", None),
                        annotation=json_annotation_to_type(field_props["type"]),
                    )
                )

        return self._config_specs

@Vinno97 commented on GitHub (Nov 23, 2023): I don't know of an elegant way to map the "annotations" field back to Python types, but this seems to work for my chains: ```py def json_annotation_to_type(annotation: str) -> type: if annotation == "string": return str elif annotation == "boolean": return bool elif annotation == "integer": return int elif annotation == "number": return float elif annotation == "array": return list raise NotImplementedError(f'Can\'t convert annotation "{annotation}" to type') class ConfigurableRemoteRunnable(RemoteRunnable): @property def config_specs(self) -> list[ConfigurableFieldSpec]: if getattr(self, "_config_specs", None) is None: schema = self.sync_client.get("/config_schema").json() fields = schema["definitions"]["Configurable"]["properties"] self._config_specs = [] for field_id, field_props in fields.items(): self._config_specs.append( ConfigurableFieldSpec( id=field_id, name=field_props["title"], description=field_props["description"], default=field_props.get("default", None), annotation=json_annotation_to_type(field_props["type"]), ) ) return self._config_specs ```

yindo referenced this issue

2026-02-16 00:20:06 -05:00

[PR #47] [MERGED] Temporary workaround for generating unique model names #304

Sign in to join this conversation.

Branches Tags

main

dependabot/pip/pip-82fdfb001f

dependabot/pip/jupyterlab-4.5.6

dependabot/pip/fastapi-0.136.0

dependabot/pip/pydantic-2.13.3

dependabot/pip/sse-starlette-3.3.4

dependabot/pip/pytest-asyncio-1.3.0

sam/bump-langchain

joaquin/test-pushing-sk-to-langserver-should-block

joaquin/bump-again-due-to-failed-release

eyurtsev-patch-5

wfh/relax_sse_starlette

eyurtsev-patch-4

jacob/docs

nc/22apr/serve

add-playground-test

eugene/playground_security

eyurtsev-patch-3

jacob/chat_widget

eyurtsev-patch-2

eyurtsev-patch-1

eugene/add_is_instance_check

eugene/agent_executor_playground

revert-209-2023-11-10-format-langsmith-runs

2023-11-10-format-langsmith-runs

harrison/dont-do-validation

nc/4nov

eugene/playground

eugene/exmaples

dqbd/playground-autotype-tuple-fix

erick/packages-bugfix

erick/assign-input-params

jacob/playground

jacob/allowlist

erick/langservehub

eugene/add_interpreptors

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: langchain-ai/langserve#47

Expose configurables through RemoteRunnable #47

Expose configurables through `RemoteRunnable` #47