The agent engineering platform.
Python agentsaiai-agentsanthropicchatgptdeepagentsenterpriseframeworkgeminigenerative-ailangchainlanggraphllmmultiagentopen-sourceopenaipydanticpythonragtypescript
docs(core): clarify `Runnable` pipe coercion [closes #39075] (#39707)
ccurme4f355f38dec5
diff --git a/libs/core/langchain_core/runnables/base.py b/libs/core/langchain_core/runnables/base.pyindex c6a31a9a79..446de50b90 100644--- a/libs/core/langchain_core/runnables/base.py+++ b/libs/core/langchain_core/runnables/base.py@@ -663,6 +663,12 @@ class Runnable(ABC, Generic[Input, Output]):Returns:A new `Runnable`.++ Raises:+ TypeError: If `other` cannot be coerced to a `Runnable`. This method does+ not return `NotImplemented`, to avoid delegating LCEL composition to+ unrelated reflected `|` implementations. To interoperate, implement+ `Runnable`, make the object callable, or define `__or__` on the left."""return RunnableSequence(self, coerce_to_runnable(other))@@ -706,6 +712,12 @@ class Runnable(ABC, Generic[Input, Output]):Returns:A new `Runnable`.++ Raises:+ TypeError: If `other` cannot be coerced to a `Runnable`. This method does+ not return `NotImplemented`, to avoid delegating LCEL composition to+ unrelated reflected `|` implementations. To interoperate, implement+ `Runnable`, make the object callable, or define `__or__` on the left."""return RunnableSequence(coerce_to_runnable(other), self)