github/langchain
1
0
Fork 0
mirror of https://github.com/hwchase17/langchain.git synced 2026-02-21 06:33:41 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs improvements

Browse Source
This commit is contained in:
Sydney Runkle
2025-09-16 11:12:55 -04:00
parent e8802c60de
commit 4f33433ce5
1 changed files with 12 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
libs/langchain_v1/langchain/agents/middleware/human_in_the_loop.py
View File

@@ -15,10 +15,12 @@ class HumanInterruptConfig(TypedDict):
This controls the available interaction options when the graph is paused for human input.
Attributes:
allow_approve: Whether the human can accept/approve the current state
allow_ignore: Whether the human can choose to ignore/skip the current step
allow_response: Whether the human can provide a text response/feedback
allow_edit: Whether the human can edit the provided content/state
allow_approve: Whether the human can approve a given tool call
allow_ignore: Whether the human can choose to ignore a given tool call.
Results in an artificial tool message added to the `messages` list.
allow_response: Whether the human can provide a response to a given tool call.
Results in an artificial tool message added to the `messages` list.
allow_edit: Whether the human can edit a tool call (name and args)
"""
allow_approve: bool
@@ -30,7 +32,8 @@ class HumanInterruptConfig(TypedDict):
class HumanInterrupt(TypedDict):
"""Represents an interrupt triggered by the graph that requires human intervention.
This is passed to the `interrupt` function when execution is paused for human input.
For each tool call requiring human input, a `HumanInterrupt` is created and
is accessible in the agent state when execution is paused for human input.
Attributes:
action: The action being requested (a tool name)
@@ -106,16 +109,19 @@ class HumanInTheLoopMiddleware(AgentMiddleware):
def __init__(
self,
tool_configs: dict[str, bool | HumanInterruptConfig],
*,
message_prefix: str = "Tool execution requires approval",
) -> None:
"""Initialize the human in the loop middleware.
Args:
tool_configs: Mapping of tool name to interrupt config (bool | HumanInterruptConfig).
If a tool doesn't have an entry, it's auto-approved by default.
* `True` indicates all interrupt config options are allowed.
* `False` indicates that the tool is auto-approved.
* `HumanInterruptConfig` indicates the specific interrupt config options to use.
message_prefix: The message prefix to use when constructing interrupt content.
message_prefix: The prefix to use when constructing interrupts (requests for input).
This is used to provide context about the tool call and the action being requested.
"""
super().__init__()
resolved_tool_configs = {}