github/langchain
1
0
Fork 0
mirror of https://github.com/hwchase17/langchain.git synced 2025-06-22 06:39:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

langchain[patch]: Extract _aperform_agent_action from _aiter_next_step from AgentExecutor (#15707)

Browse Source 
- **Description:** extreact the _aperform_agent_action in the
AgentExecutor class to allow for easier overriding. Extracted logic from
_iter_next_step into a new method _perform_agent_action for consistency
and easier overriding.
- **Issue:** #15706

Closes #15706
This commit is contained in:
Gianfranco Demarco 2024-01-24 03:22:09 +01:00 committed by GitHub
parent 95ee69a301
commit c69f599594
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 89 additions and 69 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

158
libs/langchain/langchain/agents/agent.py
View File

@ -1179,37 +1179,48 @@ class AgentExecutor(Chain):
for agent_action in actions: for agent_action in actions:
yield agent_action yield agent_action
for agent_action in actions: for agent_action in actions:
if run_manager: yield self._perform_agent_action(
run_manager.on_agent_action(agent_action, color="green") name_to_tool_map, color_mapping, agent_action, run_manager
# Otherwise we lookup the tool )
if agent_action.tool in name_to_tool_map:
tool = name_to_tool_map[agent_action.tool] def _perform_agent_action(
return_direct = tool.return_direct self,
color = color_mapping[agent_action.tool] name_to_tool_map: Dict[str, BaseTool],
tool_run_kwargs = self.agent.tool_run_logging_kwargs() color_mapping: Dict[str, str],
if return_direct: agent_action: AgentAction,
tool_run_kwargs["llm_prefix"] = "" run_manager: Optional[CallbackManagerForChainRun] = None,
# We then call the tool on the tool input to get an observation ) -> AgentStep:
observation = tool.run( if run_manager:
agent_action.tool_input, run_manager.on_agent_action(agent_action, color="green")
verbose=self.verbose, # Otherwise we lookup the tool
color=color, if agent_action.tool in name_to_tool_map:
callbacks=run_manager.get_child() if run_manager else None, tool = name_to_tool_map[agent_action.tool]
**tool_run_kwargs, return_direct = tool.return_direct
) color = color_mapping[agent_action.tool]
else: tool_run_kwargs = self.agent.tool_run_logging_kwargs()
tool_run_kwargs = self.agent.tool_run_logging_kwargs() if return_direct:
observation = InvalidTool().run( tool_run_kwargs["llm_prefix"] = ""
{ # We then call the tool on the tool input to get an observation
"requested_tool_name": agent_action.tool, observation = tool.run(
"available_tool_names": list(name_to_tool_map.keys()), agent_action.tool_input,
}, verbose=self.verbose,
verbose=self.verbose, color=color,
color=None, callbacks=run_manager.get_child() if run_manager else None,
callbacks=run_manager.get_child() if run_manager else None, **tool_run_kwargs,
**tool_run_kwargs, )
) else:
yield AgentStep(action=agent_action, observation=observation) tool_run_kwargs = self.agent.tool_run_logging_kwargs()
observation = InvalidTool().run(
{
"requested_tool_name": agent_action.tool,
"available_tool_names": list(name_to_tool_map.keys()),
},
verbose=self.verbose,
color=None,
callbacks=run_manager.get_child() if run_manager else None,
**tool_run_kwargs,
)
return AgentStep(action=agent_action, observation=observation)
async def _atake_next_step( async def _atake_next_step(
self, self,
@ -1303,52 +1314,61 @@ class AgentExecutor(Chain):
for agent_action in actions: for agent_action in actions:
yield agent_action yield agent_action
async def _aperform_agent_action(
agent_action: AgentAction,
) -> AgentStep:
if run_manager:
await run_manager.on_agent_action(
agent_action, verbose=self.verbose, color="green"
)
# Otherwise we lookup the tool
if agent_action.tool in name_to_tool_map:
tool = name_to_tool_map[agent_action.tool]
return_direct = tool.return_direct
color = color_mapping[agent_action.tool]
tool_run_kwargs = self.agent.tool_run_logging_kwargs()
if return_direct:
tool_run_kwargs["llm_prefix"] = ""
# We then call the tool on the tool input to get an observation
observation = await tool.arun(
agent_action.tool_input,
verbose=self.verbose,
color=color,
callbacks=run_manager.get_child() if run_manager else None,
**tool_run_kwargs,
)
else:
tool_run_kwargs = self.agent.tool_run_logging_kwargs()
observation = await InvalidTool().arun(
{
"requested_tool_name": agent_action.tool,
"available_tool_names": list(name_to_tool_map.keys()),
},
verbose=self.verbose,
color=None,
callbacks=run_manager.get_child() if run_manager else None,
**tool_run_kwargs,
)
return AgentStep(action=agent_action, observation=observation)
# Use asyncio.gather to run multiple tool.arun() calls concurrently # Use asyncio.gather to run multiple tool.arun() calls concurrently
result = await asyncio.gather( result = await asyncio.gather(
*[_aperform_agent_action(agent_action) for agent_action in actions] *[
self._aperform_agent_action(
name_to_tool_map, color_mapping, agent_action, run_manager
)
for agent_action in actions
],
) )
# TODO This could yield each result as it becomes available # TODO This could yield each result as it becomes available
for chunk in result: for chunk in result:
yield chunk yield chunk
async def _aperform_agent_action(
self,
name_to_tool_map: Dict[str, BaseTool],
color_mapping: Dict[str, str],
agent_action: AgentAction,
run_manager: Optional[AsyncCallbackManagerForChainRun] = None,
) -> AgentStep:
if run_manager:
await run_manager.on_agent_action(
agent_action, verbose=self.verbose, color="green"
)
# Otherwise we lookup the tool
if agent_action.tool in name_to_tool_map:
tool = name_to_tool_map[agent_action.tool]
return_direct = tool.return_direct
color = color_mapping[agent_action.tool]
tool_run_kwargs = self.agent.tool_run_logging_kwargs()
if return_direct:
tool_run_kwargs["llm_prefix"] = ""
# We then call the tool on the tool input to get an observation
observation = await tool.arun(
agent_action.tool_input,
verbose=self.verbose,
color=color,
callbacks=run_manager.get_child() if run_manager else None,
**tool_run_kwargs,
)
else:
tool_run_kwargs = self.agent.tool_run_logging_kwargs()
observation = await InvalidTool().arun(
{
"requested_tool_name": agent_action.tool,
"available_tool_names": list(name_to_tool_map.keys()),
},
verbose=self.verbose,
color=None,
callbacks=run_manager.get_child() if run_manager else None,
**tool_run_kwargs,
)
return AgentStep(action=agent_action, observation=observation)
def _call( def _call(
self, self,
inputs: Dict[str, str], inputs: Dict[str, str],