langchain_core.runnables.base.RunnableGenerator

class langchain_core.runnables.base.RunnableGenerator(transform: Union[Callable[[Iterator[Input]], Iterator[Output]], Callable[[AsyncIterator[Input]], AsyncIterator[Output]]], atransform: Optional[Callable[[AsyncIterator[Input]], AsyncIterator[Output]]] = None)[source]

Runnable that runs a generator function.

RunnableGenerators can be instantiated directly or by using a generator within a sequence.

RunnableGenerators can be used to implement custom behavior, such as custom output parsers, while preserving streaming capabilities. Given a generator function with a signature Iterator[A] -> Iterator[B], wrapping it in a RunnableGenerator allows it to emit output chunks as soon as they are streamed in from the previous step.

Note that if a generator function has a signature A -> Iterator[B], such that it requires its input from the previous step to be completed before emitting chunks (e.g., most LLMs need the entire prompt available to start generating), it can instead be wrapped in a RunnableLambda.

Here is an example to show the basic mechanics of a RunnableGenerator:

from typing import Any, AsyncIterator, Iterator

from langchain_core.runnables import RunnableGenerator


def gen(input: Iterator[Any]) -> Iterator[str]:
    for token in ["Have", " a", " nice", " day"]:
        yield token


runnable = RunnableGenerator(gen)
runnable.invoke(None)  # "Have a nice day"
list(runnable.stream(None))  # ["Have", " a", " nice", " day"]
runnable.batch([None, None])  # ["Have a nice day", "Have a nice day"]


# Async version:
async def agen(input: AsyncIterator[Any]) -> AsyncIterator[str]:
    for token in ["Have", " a", " nice", " day"]:
        yield token

runnable = RunnableGenerator(agen)
await runnable.ainvoke(None)  # "Have a nice day"
[p async for p in runnable.astream(None)] # ["Have", " a", " nice", " day"]

RunnableGenerator makes it easy to implement custom behavior within a streaming context. Below we show an example:

from langchain_core.prompts import ChatPromptTemplate
from langchain_core.runnables import RunnableGenerator, RunnableLambda
from langchain_openai import ChatOpenAI
from langchain_core.output_parsers import StrOutputParser


model = ChatOpenAI()
chant_chain = (
    ChatPromptTemplate.from_template("Give me a 3 word chant about {topic}")
    | model
    | StrOutputParser()
)

def character_generator(input: Iterator[str]) -> Iterator[str]:
    for token in input:
        if "," in token or "." in token:
            yield "👏" + token
        else:
            yield token


runnable = chant_chain | character_generator
assert type(runnable.last) is RunnableGenerator
"".join(runnable.stream({"topic": "waste"})) # Reduce👏, Reuse👏, Recycle👏.

# Note that RunnableLambda can be used to delay streaming of one step in a
# sequence until the previous step is finished:
def reverse_generator(input: str) -> Iterator[str]:
    # Yield characters of input in reverse order.
    for character in input[::-1]:
        yield character

runnable = chant_chain | RunnableLambda(reverse_generator)
"".join(runnable.stream({"topic": "waste"}))  # ".elcycer ,esuer ,ecudeR"

Attributes

InputType

The type of input this runnable accepts specified as a type annotation.

OutputType

The type of output this runnable produces specified as a type annotation.

config_specs

List configurable fields for this runnable.

input_schema

The type of input this runnable accepts specified as a pydantic model.

name

The name of the runnable.

output_schema

The type of output this runnable produces specified as a pydantic model.

Methods

__init__(transform[, atransform])

abatch(inputs[, config, return_exceptions])

Default implementation runs ainvoke in parallel using asyncio.gather.

ainvoke(input[, config])

Default implementation of ainvoke, calls invoke from a thread.

assign(**kwargs)

Assigns new fields to the dict output of this runnable.

astream(input[, config])

Default implementation of astream, which calls ainvoke.

astream_events(input[, config, ...])

[Beta] Generate a stream of events.

astream_log(input[, config, diff, ...])

Stream all output from a runnable, as reported to the callback system.

atransform(input[, config])

Default implementation of atransform, which buffers input and calls astream.

batch(inputs[, config, return_exceptions])

Default implementation runs invoke in parallel using a thread pool executor.

bind(**kwargs)

Bind arguments to a Runnable, returning a new Runnable.

config_schema(*[, include])

The type of config this runnable accepts specified as a pydantic model.

get_graph([config])

Return a graph representation of this runnable.

get_input_schema([config])

Get a pydantic model that can be used to validate input to the runnable.

get_name([suffix, name])

Get the name of the runnable.

get_output_schema([config])

Get a pydantic model that can be used to validate output to the runnable.

get_prompts([config])

invoke(input[, config])

Transform a single input into an output.

map()

Return a new Runnable that maps a list of inputs to a list of outputs, by calling invoke() with each input.

pick(keys)

Pick keys from the dict output of this runnable.

pipe(*others[, name])

Compose this runnable with another object to create a RunnableSequence.

stream(input[, config])

Default implementation of stream, which calls invoke.

transform(input[, config])

Default implementation of transform, which buffers input and then calls stream.

with_config([config])

Bind config to a Runnable, returning a new Runnable.

with_fallbacks(fallbacks, *[, ...])

Add fallbacks to a runnable, returning a new Runnable.

with_listeners(*[, on_start, on_end, on_error])

Bind lifecycle listeners to a Runnable, returning a new Runnable.

with_retry(*[, retry_if_exception_type, ...])

Create a new Runnable that retries the original runnable on exceptions.

with_types(*[, input_type, output_type])

Bind input and output types to a Runnable, returning a new Runnable.

Parameters
  • transform (Union[Callable[[Iterator[Input]], Iterator[Output]], Callable[[AsyncIterator[Input]], AsyncIterator[Output]]]) –

  • atransform (Optional[Callable[[AsyncIterator[Input]], AsyncIterator[Output]]]) –

__init__(transform: Union[Callable[[Iterator[Input]], Iterator[Output]], Callable[[AsyncIterator[Input]], AsyncIterator[Output]]], atransform: Optional[Callable[[AsyncIterator[Input]], AsyncIterator[Output]]] = None) None[source]
Parameters
  • transform (Union[Callable[[Iterator[Input]], Iterator[Output]], Callable[[AsyncIterator[Input]], AsyncIterator[Output]]]) –

  • atransform (Optional[Callable[[AsyncIterator[Input]], AsyncIterator[Output]]]) –

Return type

None

async abatch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, return_exceptions: bool = False, **kwargs: Optional[Any]) List[Output]

Default implementation runs ainvoke in parallel using asyncio.gather.

The default implementation of batch works well for IO bound runnables.

Subclasses should override this method if they can batch more efficiently; e.g., if the underlying runnable uses an API which supports a batch mode.

Parameters
  • inputs (List[Input]) –

  • config (Optional[Union[RunnableConfig, List[RunnableConfig]]]) –

  • return_exceptions (bool) –

  • kwargs (Optional[Any]) –

Return type

List[Output]

async ainvoke(input: Input, config: Optional[RunnableConfig] = None, **kwargs: Any) Output[source]

Default implementation of ainvoke, calls invoke from a thread.

The default implementation allows usage of async code even if the runnable did not implement a native async version of invoke.

Subclasses should override this method if they can run asynchronously.

Parameters
  • input (Input) –

  • config (Optional[RunnableConfig]) –

  • kwargs (Any) –

Return type

Output

assign(**kwargs: Union[Runnable[Dict[str, Any], Any], Callable[[Dict[str, Any]], Any], Mapping[str, Union[Runnable[Dict[str, Any], Any], Callable[[Dict[str, Any]], Any]]]]) RunnableSerializable[Any, Any]

Assigns new fields to the dict output of this runnable. Returns a new runnable.

Parameters

kwargs (Union[Runnable[Dict[str, Any], Any], Callable[[Dict[str, Any]], Any], Mapping[str, Union[Runnable[Dict[str, Any], Any], Callable[[Dict[str, Any]], Any]]]]) –

Return type

RunnableSerializable[Any, Any]

astream(input: Input, config: Optional[RunnableConfig] = None, **kwargs: Any) AsyncIterator[Output][source]

Default implementation of astream, which calls ainvoke. Subclasses should override this method if they support streaming output.

Parameters
  • input (Input) –

  • config (Optional[RunnableConfig]) –

  • kwargs (Any) –

Return type

AsyncIterator[Output]

astream_events(input: Any, config: Optional[RunnableConfig] = None, *, version: Literal['v1'], include_names: Optional[Sequence[str]] = None, include_types: Optional[Sequence[str]] = None, include_tags: Optional[Sequence[str]] = None, exclude_names: Optional[Sequence[str]] = None, exclude_types: Optional[Sequence[str]] = None, exclude_tags: Optional[Sequence[str]] = None, **kwargs: Any) AsyncIterator[StreamEvent]

[Beta] Generate a stream of events.

Use to create an iterator over StreamEvents that provide real-time information about the progress of the runnable, including StreamEvents from intermediate results.

A StreamEvent is a dictionary with the following schema:

  • event: str - Event names are of the

    format: on_[runnable_type]_(start|stream|end).

  • name: str - The name of the runnable that generated the event.

  • run_id: str - randomly generated ID associated with the given execution of

    the runnable that emitted the event. A child runnable that gets invoked as part of the execution of a parent runnable is assigned its own unique ID.

  • tags: Optional[List[str]] - The tags of the runnable that generated

    the event.

  • metadata: Optional[Dict[str, Any]] - The metadata of the runnable

    that generated the event.

  • data: Dict[str, Any]

Below is a table that illustrates some evens that might be emitted by various chains. Metadata fields have been omitted from the table for brevity. Chain definitions have been included after the table.

event

name

chunk

input

output

on_chat_model_start

[model name]

{“messages”: [[SystemMessage, HumanMessage]]}

on_chat_model_stream

[model name]

AIMessageChunk(content=”hello”)

on_chat_model_end

[model name]

{“messages”: [[SystemMessage, HumanMessage]]}

{“generations”: […], “llm_output”: None, …}

on_llm_start

[model name]

{‘input’: ‘hello’}

on_llm_stream

[model name]

‘Hello’

on_llm_end

[model name]

‘Hello human!’

on_chain_start

format_docs

on_chain_stream

format_docs

“hello world!, goodbye world!”

on_chain_end

format_docs

[Document(…)]

“hello world!, goodbye world!”

on_tool_start

some_tool

{“x”: 1, “y”: “2”}

on_tool_stream

some_tool

{“x”: 1, “y”: “2”}

on_tool_end

some_tool

{“x”: 1, “y”: “2”}

on_retriever_start

[retriever name]

{“query”: “hello”}

on_retriever_chunk

[retriever name]

{documents: […]}

on_retriever_end

[retriever name]

{“query”: “hello”}

{documents: […]}

on_prompt_start

[template_name]

{“question”: “hello”}

on_prompt_end

[template_name]

{“question”: “hello”}

ChatPromptValue(messages: [SystemMessage, …])

Here are declarations associated with the events shown above:

format_docs:

def format_docs(docs: List[Document]) -> str:
    '''Format the docs.'''
    return ", ".join([doc.page_content for doc in docs])

format_docs = RunnableLambda(format_docs)

some_tool:

@tool
def some_tool(x: int, y: str) -> dict:
    '''Some_tool.'''
    return {"x": x, "y": y}

prompt:

template = ChatPromptTemplate.from_messages(
    [("system", "You are Cat Agent 007"), ("human", "{question}")]
).with_config({"run_name": "my_template", "tags": ["my_template"]})

Example:

from langchain_core.runnables import RunnableLambda

async def reverse(s: str) -> str:
    return s[::-1]

chain = RunnableLambda(func=reverse)

events = [
    event async for event in chain.astream_events("hello", version="v1")
]

# will produce the following events (run_id has been omitted for brevity):
[
    {
        "data": {"input": "hello"},
        "event": "on_chain_start",
        "metadata": {},
        "name": "reverse",
        "tags": [],
    },
    {
        "data": {"chunk": "olleh"},
        "event": "on_chain_stream",
        "metadata": {},
        "name": "reverse",
        "tags": [],
    },
    {
        "data": {"output": "olleh"},
        "event": "on_chain_end",
        "metadata": {},
        "name": "reverse",
        "tags": [],
    },
]
Parameters
  • input (Any) – The input to the runnable.

  • config (Optional[RunnableConfig]) – The config to use for the runnable.

  • version (Literal['v1']) – The version of the schema to use. Currently only version 1 is available. No default will be assigned until the API is stabilized.

  • include_names (Optional[Sequence[str]]) – Only include events from runnables with matching names.

  • include_types (Optional[Sequence[str]]) – Only include events from runnables with matching types.

  • include_tags (Optional[Sequence[str]]) – Only include events from runnables with matching tags.

  • exclude_names (Optional[Sequence[str]]) – Exclude events from runnables with matching names.

  • exclude_types (Optional[Sequence[str]]) – Exclude events from runnables with matching types.

  • exclude_tags (Optional[Sequence[str]]) – Exclude events from runnables with matching tags.

  • kwargs (Any) – Additional keyword arguments to pass to the runnable. These will be passed to astream_log as this implementation of astream_events is built on top of astream_log.

Returns

An async stream of StreamEvents.

Return type

AsyncIterator[StreamEvent]

Notes

async astream_log(input: Any, config: Optional[RunnableConfig] = None, *, diff: bool = True, with_streamed_output_list: bool = True, include_names: Optional[Sequence[str]] = None, include_types: Optional[Sequence[str]] = None, include_tags: Optional[Sequence[str]] = None, exclude_names: Optional[Sequence[str]] = None, exclude_types: Optional[Sequence[str]] = None, exclude_tags: Optional[Sequence[str]] = None, **kwargs: Any) Union[AsyncIterator[RunLogPatch], AsyncIterator[RunLog]]

Stream all output from a runnable, as reported to the callback system. This includes all inner runs of LLMs, Retrievers, Tools, etc.

Output is streamed as Log objects, which include a list of jsonpatch ops that describe how the state of the run has changed in each step, and the final state of the run.

The jsonpatch ops can be applied in order to construct state.

Parameters
  • input (Any) – The input to the runnable.

  • config (Optional[RunnableConfig]) – The config to use for the runnable.

  • diff (bool) – Whether to yield diffs between each step, or the current state.

  • with_streamed_output_list (bool) – Whether to yield the streamed_output list.

  • include_names (Optional[Sequence[str]]) – Only include logs with these names.

  • include_types (Optional[Sequence[str]]) – Only include logs with these types.

  • include_tags (Optional[Sequence[str]]) – Only include logs with these tags.

  • exclude_names (Optional[Sequence[str]]) – Exclude logs with these names.

  • exclude_types (Optional[Sequence[str]]) – Exclude logs with these types.

  • exclude_tags (Optional[Sequence[str]]) – Exclude logs with these tags.

  • kwargs (Any) –

Return type

Union[AsyncIterator[RunLogPatch], AsyncIterator[RunLog]]

atransform(input: AsyncIterator[Input], config: Optional[RunnableConfig] = None, **kwargs: Any) AsyncIterator[Output][source]

Default implementation of atransform, which buffers input and calls astream. Subclasses should override this method if they can start producing output while input is still being generated.

Parameters
  • input (AsyncIterator[Input]) –

  • config (Optional[RunnableConfig]) –

  • kwargs (Any) –

Return type

AsyncIterator[Output]

batch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, return_exceptions: bool = False, **kwargs: Optional[Any]) List[Output]

Default implementation runs invoke in parallel using a thread pool executor.

The default implementation of batch works well for IO bound runnables.

Subclasses should override this method if they can batch more efficiently; e.g., if the underlying runnable uses an API which supports a batch mode.

Parameters
  • inputs (List[Input]) –

  • config (Optional[Union[RunnableConfig, List[RunnableConfig]]]) –

  • return_exceptions (bool) –

  • kwargs (Optional[Any]) –

Return type

List[Output]

bind(**kwargs: Any) Runnable[Input, Output]

Bind arguments to a Runnable, returning a new Runnable.

Parameters

kwargs (Any) –

Return type

Runnable[Input, Output]

config_schema(*, include: Optional[Sequence[str]] = None) Type[BaseModel]

The type of config this runnable accepts specified as a pydantic model.

To mark a field as configurable, see the configurable_fields and configurable_alternatives methods.

Parameters

include (Optional[Sequence[str]]) – A list of fields to include in the config schema.

Returns

A pydantic model that can be used to validate config.

Return type

Type[BaseModel]

get_graph(config: Optional[RunnableConfig] = None) Graph

Return a graph representation of this runnable.

Parameters

config (Optional[RunnableConfig]) –

Return type

Graph

get_input_schema(config: Optional[RunnableConfig] = None) Type[BaseModel]

Get a pydantic model that can be used to validate input to the runnable.

Runnables that leverage the configurable_fields and configurable_alternatives methods will have a dynamic input schema that depends on which configuration the runnable is invoked with.

This method allows to get an input schema for a specific configuration.

Parameters

config (Optional[RunnableConfig]) – A config to use when generating the schema.

Returns

A pydantic model that can be used to validate input.

Return type

Type[BaseModel]

get_name(suffix: Optional[str] = None, *, name: Optional[str] = None) str

Get the name of the runnable.

Parameters
  • suffix (Optional[str]) –

  • name (Optional[str]) –

Return type

str

get_output_schema(config: Optional[RunnableConfig] = None) Type[BaseModel]

Get a pydantic model that can be used to validate output to the runnable.

Runnables that leverage the configurable_fields and configurable_alternatives methods will have a dynamic output schema that depends on which configuration the runnable is invoked with.

This method allows to get an output schema for a specific configuration.

Parameters

config (Optional[RunnableConfig]) – A config to use when generating the schema.

Returns

A pydantic model that can be used to validate output.

Return type

Type[BaseModel]

get_prompts(config: Optional[RunnableConfig] = None) List[BasePromptTemplate]
Parameters

config (Optional[RunnableConfig]) –

Return type

List[BasePromptTemplate]

invoke(input: Input, config: Optional[RunnableConfig] = None, **kwargs: Any) Output[source]

Transform a single input into an output. Override to implement.

Parameters
  • input (Input) – The input to the runnable.

  • config (Optional[RunnableConfig]) – A config to use when invoking the runnable. The config supports standard keys like ‘tags’, ‘metadata’ for tracing purposes, ‘max_concurrency’ for controlling how much work to do in parallel, and other keys. Please refer to the RunnableConfig for more details.

  • kwargs (Any) –

Returns

The output of the runnable.

Return type

Output

map() Runnable[List[Input], List[Output]]

Return a new Runnable that maps a list of inputs to a list of outputs, by calling invoke() with each input.

Return type

Runnable[List[Input], List[Output]]

pick(keys: Union[str, List[str]]) RunnableSerializable[Any, Any]

Pick keys from the dict output of this runnable. Returns a new runnable.

Parameters

keys (Union[str, List[str]]) –

Return type

RunnableSerializable[Any, Any]

pipe(*others: Union[Runnable[Any, Other], Callable[[Any], Other]], name: Optional[str] = None) RunnableSerializable[Input, Other]

Compose this runnable with another object to create a RunnableSequence.

Parameters
  • others (Union[Runnable[Any, Other], Callable[[Any], Other]]) –

  • name (Optional[str]) –

Return type

RunnableSerializable[Input, Other]

stream(input: Input, config: Optional[RunnableConfig] = None, **kwargs: Any) Iterator[Output][source]

Default implementation of stream, which calls invoke. Subclasses should override this method if they support streaming output.

Parameters
  • input (Input) –

  • config (Optional[RunnableConfig]) –

  • kwargs (Any) –

Return type

Iterator[Output]

transform(input: Iterator[Input], config: Optional[RunnableConfig] = None, **kwargs: Any) Iterator[Output][source]

Default implementation of transform, which buffers input and then calls stream. Subclasses should override this method if they can start producing output while input is still being generated.

Parameters
  • input (Iterator[Input]) –

  • config (Optional[RunnableConfig]) –

  • kwargs (Any) –

Return type

Iterator[Output]

with_config(config: Optional[RunnableConfig] = None, **kwargs: Any) Runnable[Input, Output]

Bind config to a Runnable, returning a new Runnable.

Parameters
Return type

Runnable[Input, Output]

with_fallbacks(fallbacks: Sequence[Runnable[Input, Output]], *, exceptions_to_handle: Tuple[Type[BaseException], ...] = (<class 'Exception'>,), exception_key: Optional[str] = None) RunnableWithFallbacksT[Input, Output]

Add fallbacks to a runnable, returning a new Runnable.

Parameters
  • fallbacks (Sequence[Runnable[Input, Output]]) – A sequence of runnables to try if the original runnable fails.

  • exceptions_to_handle (Tuple[Type[BaseException], ...]) – A tuple of exception types to handle.

  • exception_key (Optional[str]) – If string is specified then handled exceptions will be passed to fallbacks as part of the input under the specified key. If None, exceptions will not be passed to fallbacks. If used, the base runnable and its fallbacks must accept a dictionary as input.

Returns

A new Runnable that will try the original runnable, and then each fallback in order, upon failures.

Return type

RunnableWithFallbacksT[Input, Output]

with_listeners(*, on_start: Optional[Listener] = None, on_end: Optional[Listener] = None, on_error: Optional[Listener] = None) Runnable[Input, Output]

Bind lifecycle listeners to a Runnable, returning a new Runnable.

on_start: Called before the runnable starts running, with the Run object. on_end: Called after the runnable finishes running, with the Run object. on_error: Called if the runnable throws an error, with the Run object.

The Run object contains information about the run, including its id, type, input, output, error, start_time, end_time, and any tags or metadata added to the run.

Parameters
  • on_start (Optional[Listener]) –

  • on_end (Optional[Listener]) –

  • on_error (Optional[Listener]) –

Return type

Runnable[Input, Output]

with_retry(*, retry_if_exception_type: ~typing.Tuple[~typing.Type[BaseException], ...] = (<class 'Exception'>,), wait_exponential_jitter: bool = True, stop_after_attempt: int = 3) Runnable[Input, Output]

Create a new Runnable that retries the original runnable on exceptions.

Parameters
  • retry_if_exception_type (Tuple[Type[BaseException], ...]) – A tuple of exception types to retry on

  • wait_exponential_jitter (bool) – Whether to add jitter to the wait time between retries

  • stop_after_attempt (int) – The maximum number of attempts to make before giving up

Returns

A new Runnable that retries the original runnable on exceptions.

Return type

Runnable[Input, Output]

with_types(*, input_type: Optional[Type[Input]] = None, output_type: Optional[Type[Output]] = None) Runnable[Input, Output]

Bind input and output types to a Runnable, returning a new Runnable.

Parameters
  • input_type (Optional[Type[Input]]) –

  • output_type (Optional[Type[Output]]) –

Return type

Runnable[Input, Output]