o Zh@sdZddlmZddlZddlZddlZddlZddlmZddl m Z ddl m Z m Z mZmZmZmZmZmZmZmZmZddlZddlmZddlmZmZmZdd lmZm Z m!Z!m"Z"m#Z#m$Z$dd l%m&Z&dd l'm(Z(dd l)m*Z*dd l+m,Z,ddl-m.Z.ddl/m0Z0ddl1m2Z2ddl3m4Z4m5Z5m6Z6ddl7m8Z8ddl9m:Z:ddl;mZ>m?Z?m@Z@ddlAmBZBddlCmDZDddlEmFZFddlGmHZHddlImJZJddlKmLZLddlMmNZNddlOmPZPeQeRZSGddde>ZTGd d!d!e>ZUGd"d#d#e,eeefZVGd$d%d%e,eeeefZWGd&d'd'eTZXGd(d)d)eUZYed*eDd+d,Gd-d.d.eTZZed*eDd+d,Gd/d0d0eTZ[Gd1d2d2e:Z\eeeeefZ]eeXeYfZ^Gd3d4d4eLZ_dS)5zEChain that takes in an input and produces an action and action input.) annotationsN)abstractmethod)Path) Any AsyncIteratorCallableDictIteratorListOptionalSequenceTupleUnioncast) deprecated) AgentAction AgentFinish AgentStep)AsyncCallbackManagerForChainRunAsyncCallbackManagerForToolRunBaseCallbackManagerCallbackManagerForChainRunCallbackManagerForToolRun Callbacks)OutputParserException)BaseLanguageModel) BaseMessage)BaseOutputParser)BasePromptTemplate)FewShotPromptTemplate)PromptTemplate)RunnableRunnableConfig ensure_config) AddableDict)BaseTool)get_color_mapping) BaseModel ConfigDictmodel_validator)Self)AGENT_DEPRECATION_WARNING)AgentExecutorIterator) AgentType) InvalidTool)Chain)LLMChain)asyncio_timeoutcseZdZdZed0ddZd1ddZe d2d3ddZe d2d3ddZ eed0ddZ d4ddZ e d2d5d"d#Z ed6d$d%Zd7fd'd( Zd8d,d-Zd9d.d/ZZS):BaseSingleActionAgentzBase Single Action Agent class.return List[str]cCdgSReturn values of the agent.outputselfr9r9M/var/www/html/lang_env/lib/python3.10/site-packages/langchain/agents/agent.py return_values@z#BaseSingleActionAgent.return_valuesOptional[List[str]]cCsdSNr9r:r9r9r<get_allowed_toolsEsz'BaseSingleActionAgent.get_allowed_toolsNintermediate_stepsList[Tuple[AgentAction, str]] callbacksrkwargsrUnion[AgentAction, AgentFinish]cKdS)0Given input, decided what to do. Args: intermediate_steps: Steps the LLM has taken to date, along with observations. callbacks: Callbacks to run. **kwargs: User inputs. Returns: Action specifying what tool to use. Nr9r;rBrDrEr9r9r<planHzBaseSingleActionAgent.plancdS)6Async given input, decided what to do. Args: intermediate_steps: Steps the LLM has taken to date, along with observations. callbacks: Callbacks to run. **kwargs: User inputs. Returns: Action specifying what tool to use. Nr9rIr9r9r<aplan[zBaseSingleActionAgent.aplancCrG7Return the input keys. :meta private: Nr9r:r9r9r< input_keysnrKz BaseSingleActionAgent.input_keysearly_stopping_methodstrrcK&|dkr tddidStd|d)Return response when agent has been stopped due to max iterations. Args: early_stopping_method: Method to use for early stopping. intermediate_steps: Steps the LLM has taken to date, along with observations. **kwargs: User inputs. Returns: AgentFinish: Agent finish object. Raises: ValueError: If `early_stopping_method` is not supported. forcer83Agent stopped due to iteration limit or time limit.'Got unsupported early_stopping_method ``r ValueErrorr;rSrBrEr9r9r<return_stopped_responsevs z-BaseSingleActionAgent.return_stopped_responsellmrtoolsSequence[BaseTool]callback_managerOptional[BaseCallbackManager]cKt)a.Construct an agent from an LLM and tools. Args: llm: Language model to use. tools: Tools to use. callback_manager: Callback manager to use. kwargs: Additional arguments. Returns: BaseSingleActionAgent: Agent object. NotImplementedError)clsr`rarcrEr9r9r<from_llm_and_toolssz(BaseSingleActionAgent.from_llm_and_toolscCrez#Return Identifier of an agent type.rfr:r9r9r< _agent_typez!BaseSingleActionAgent._agent_typerc s\t}z|j}Wn tyd}Ynwt|tr$t|j|d<|S|dur,||d<|S)z{Return dictionary representation of agent. Returns: Dict: Dictionary representation of agent. N_type)super model_dumprkrg isinstancer-rTvalue)r;rE_dictrm __class__r9r<dicts    zBaseSingleActionAgent.dict file_pathUnion[Path, str]NonecCst|tr t|}n|}|j}|jddd|}d|vr&td|d|jdkrKt|d}t j ||dd Wd d S1sDwYd S|j d rqt|d}t j ||d d Wd d S1sjwYd St |d)zSave the agent. Args: file_path: Path to file to save the agent to. Example: .. code-block:: python # If working with agent executor agent.agent.save(file_path="path/agent.yaml") Tparentsexist_okrmAgent z does not support saving.jsonwindentNz.yamlz.ymlFZdefault_flow_style must be json or yaml)rprTrparentmkdirrurgsuffixopenjsondumpendswithyamlr])r;rv save_pathdirectory_path agent_dictfr9r9r<saves"   "  "zBaseSingleActionAgent.savecCiS#Return logging kwargs for tool run.r9r:r9r9r<tool_run_logging_kwargssz-BaseSingleActionAgent.tool_run_logging_kwargsr3r4r3r?r@rBrCrDrrErr3rFrSrTrBrCrErr3r) r`rrarbrcrdrErr3r2r3rTrErr3rrvrwr3rxr3r)__name__ __module__ __qualname____doc__propertyr=rArrJrNrRr_ classmethodrirkrurr __classcell__r9r9rsr<r2=s.     #r2cseZdZdZed(ddZd)ddZe d*d+ddZe d*d+ddZ eed(ddZ d,ddZ ed-ddZ d.fdd Z d/d$d%Zd0d&d'ZZS)1BaseMultiActionAgentzBase Multi Action Agent class.r3r4cCr5r6r9r:r9r9r<r=r>z"BaseMultiActionAgent.return_valuesr?cCrG)z]Get allowed tools. Returns: Optional[List[str]]: Allowed tools. Nr9r:r9r9r<rAsz&BaseMultiActionAgent.get_allowed_toolsNrBrCrDrrEr%Union[List[AgentAction], AgentFinish]cKrG)a5Given input, decided what to do. Args: intermediate_steps: Steps the LLM has taken to date, along with the observations. callbacks: Callbacks to run. **kwargs: User inputs. Returns: Actions specifying what tool to use. Nr9rIr9r9r<rJrKzBaseMultiActionAgent.plancrL)a;Async given input, decided what to do. Args: intermediate_steps: Steps the LLM has taken to date, along with the observations. callbacks: Callbacks to run. **kwargs: User inputs. Returns: Actions specifying what tool to use. Nr9rIr9r9r<rN rOzBaseMultiActionAgent.aplancCrGrPr9r:r9r9r<rRrKzBaseMultiActionAgent.input_keysrSrTrcKrU)rVrWr8z$Agent stopped due to max iterations.rYrZr[r\r^r9r9r<r_%s  z,BaseMultiActionAgent.return_stopped_responsecCrerjrfr:r9r9r<rkArlz BaseMultiActionAgent._agent_typerc s4t}z t|j|d<W|StyY|Sw)*Return dictionary representation of agent.rm)rnrorTrkrgr;rErrrsr9r<ruFs  zBaseMultiActionAgent.dictrvrwrxcCst|tr t|}n|}|}d|vrtd|d|j}|jddd|jdkrKt|d}t j ||dd Wd d S1sDwYd S|j d rqt|d}t j ||d d Wd d S1sjwYd St |d)aSave the agent. Args: file_path: Path to file to save the agent to. Raises: NotImplementedError: If agent does not support saving. ValueError: If file_path is not json or yaml. Example: .. code-block:: python # If working with agent executor agent.agent.save(file_path="path/agent.yaml") rmr|z does not support saving.Tryr}r~rrNrFrr)rprTrrurgrrrrrrrrr])r;rvrrrrr9r9r<rOs"    "  "zBaseMultiActionAgent.savecCrrr9r:r9r9r<rvrlz,BaseMultiActionAgent.tool_run_logging_kwargsrrr@rBrCrDrrErr3rrrrrr)rrrrrr=rArrJrNrRr_rkrurrrr9r9rsr<rs(     'rc@eZdZdZed ddZdS) AgentOutputParserz=Base class for parsing agent output into agent action/finish.textrTr3rFcCrG)z$Parse text into agent action/finish.Nr9r;rr9r9r<parserKzAgentOutputParser.parseN)rrTr3rFrrrrrrr9r9r9r<r|src@r) MultiActionAgentOutputParserzBase class for parsing agent output into agent actions/finish. This is used for agents that can return multiple actions. rrTr3rcCrG)zParse text into agent actions/finish. Args: text: Text to parse. Returns: Union[List[AgentAction], AgentFinish]: List of agent actions or agent finish. Nr9rr9r9r<rrKz"MultiActionAgentOutputParser.parseN)rrTr3rrr9r9r9r<rsrc@eZdZUdZded< gZded<gZded<dZded < edd Z e dd d Z e dddZ ddddZ ddddZdS) RunnableAgentAgent powered by Runnables.z/Runnable[dict, Union[AgentAction, AgentFinish]]runnabler4input_keys_argreturn_keys_argTboolstream_runnableZarbitrary_types_allowedr3cC|jSr7rr:r9r9r<r=r>zRunnableAgent.return_valuescCr)zReturn the input keys.rr:r9r9r<rRr>zRunnableAgent.input_keysNrBrCrDrrErrFcKfi|d|i}d}|jr'|jj|d|idD] }|dur |}q||7}q|S|jj|d|id}|SaPBased on past history and current inputs, decide what to do. Args: intermediate_steps: Steps the LLM has taken to date, along with the observations. callbacks: Callbacks to run. **kwargs: User inputs. Returns: Action specifying what tool to use. rBNrDconfigrrstreamZinvoker;rBrDrEinputs final_outputchunkr9r9r<rJs zRunnableAgent.plancxi|d|i}d}|jr-|jj|d|id2z3dHW}|dur%|}q||7}q6|S|jj|d|idIdH}|SaRAsync based on past history and current inputs, decide what to do. Args: intermediate_steps: Steps the LLM has taken to date, along with observations. callbacks: Callbacks to run. **kwargs: User inputs. Returns: Action specifying what tool to use. rBNrDrrrastreamZainvokerr9r9r<rNs    zRunnableAgent.aplanrr@rrrrr__annotations__rrrr(Z model_configrr=rRrJrNr9r9r9r<rs&       'rc@r)RunnableMultiActionAgentrz5Runnable[dict, Union[List[AgentAction], AgentFinish]]rr4rrTrrrr3cCrrrr:r9r9r<r=r>z&RunnableMultiActionAgent.return_valuescCr)QReturn the input keys. Returns: List of input keys. rr:r9r9r<rRsz#RunnableMultiActionAgent.input_keysNrBrCrDrrErrcKrrrrr9r9r<rJ(s zRunnableMultiActionAgent.plancrrrrr9r9r<rNOs    zRunnableMultiActionAgent.aplanrr@rrr9r9r9r<rs&       *rz0.1.0z1.0)messageZremovalcsveZdZUdZded< ded< ded< edd d Zdfdd Z ddddZ ddddZ d ddZ Z S)!LLMSingleActionAgentz$Base class for single action agents.r0 llm_chainr output_parserr4stopr3cCtt|jjdhS)rrBlistsetrrRr:r9r9r<rRszLLMSingleActionAgent.input_keysrErrc t}|d=|Srrrnrurrsr9r<ru zLLMSingleActionAgent.dictNrBrCrDrrFcKs(|jjd||j|d|}|j|S)a4Given input, decided what to do. Args: intermediate_steps: Steps the LLM has taken to date, along with the observations. callbacks: Callbacks to run. **kwargs: User inputs. Returns: Action specifying what tool to use. rBrrDNr9)rrunrrrr;rBrDrEr8r9r9r<rJs zLLMSingleActionAgent.plancs0|jjd||j|d|IdH}|j|S)rMrNr9)rarunrrrrr9r9r<rNs zLLMSingleActionAgent.aplancCs&dt|jdkr ddS|jddS)rrYr llm_prefixobservation_prefix)lenrr:r9r9r<rs z,LLMSingleActionAgent.tool_run_logging_kwargsrrr@rr) rrrrrrrRrurJrNrrr9r9rsr<r{s    rcsFeZdZUdZded< ded< dZded< dKfd d ZdLddZedMddZ dNddZ edMddZ dOddZ dPdQd"d#Z dPdQd$d%ZdRd'd(ZedMd)d*Zed+d,dSd.d/ZeedTd0d1ZeedTd2d3ZeedUd7d8ZedVd:d;ZeedWdzAgent.return_valuesrrTcCtd)zsFix the text. Args: text: Text to fix. Returns: str: Fixed text. z(fix_text not implemented for this agent.r]rr9r9r< _fix_texts zAgent._fix_textcCs d|jd|jgS)N z )rrstripr:r9r9r<_stopsz Agent._stoprBrCUnion[str, List[BaseMessage]]cCs<d}|D]\}}||j7}|d|j|d|j7}q|S)zJConstruct the scratchpad that lets the agent continue its thought process.rYr)logrr)r;rBthoughtsaction observationr9r9r<_construct_scratchpads   zAgent._construct_scratchpadrDrrFcKs4|j|fi|}|jjdd|i|}|j|S)rHrDNr9)get_full_inputsrpredictrr)r;rBrDrE full_inputs full_outputr9r9r<rJs z Agent.plancsF|j|fi|}|jjdd|i|IdH}|j|IdH}|S)rMrDNr9)rrZapredictrZaparse)r;rBrDrErrZ agent_outputr9r9r<rN's z Agent.aplanDict[str, Any]cKs&||}||jd}i||}|S)a1Create the full inputs for the LLMChain from intermediate steps. Args: intermediate_steps: Steps the LLM has taken to date, along with observations. **kwargs: User inputs. Returns: Dict[str, Any]: Full inputs for the LLMChain. agent_scratchpadr)rr)r;rBrEr new_inputsrr9r9r<r=s  zAgent.get_full_inputscCr)rQrrr:r9r9r<rROszAgent.input_keysaftermoder*cCsv|jj}d|jvr9td|jdt|tr"|jd7_|St|t r0|j d7_ |St dt ||S)aFValidate that prompt matches format. Args: values: Values to validate. Returns: Dict: Validated values. Raises: ValueError: If `agent_scratchpad` is not in prompt.input_variables and prompt is not a FewShotPromptTemplate or a PromptTemplate. rzl`agent_scratchpad` should be a variable in prompt.input_variables. Did not find it, so adding it at the end.z {agent_scratchpad}zGot unexpected prompt type ) rpromptZinput_variablesloggerwarningappendrpr templaterrr]type)r;rr9r9r<validate_promptWs    zAgent.validate_promptcCrG)z&Prefix to append the observation with.Nr9r:r9r9r<rtrKzAgent.observation_prefixcCrG)z#Prefix to append the LLM call with.Nr9r:r9r9r<ryrKzAgent.llm_prefixrarbrcCrG)zCreate a prompt for this class. Args: tools: Tools to use. Returns: BasePromptTemplate: Prompt template. Nr9rhrar9r9r< create_prompt~rKzAgent.create_promptrxcCrG)zgValidate that appropriate tools are passed in. Args: tools: Tools to use. Nr9rr9r9r<_validate_toolsszAgent._validate_toolscKrG)z)Get default output parser for this class.Nr9)rhrEr9r9r<_get_default_output_parserrKz Agent._get_default_output_parserr`rrcrdOptional[AgentOutputParser]c KsN||t||||d}dd|D}|p|}|d|||d|S)aOConstruct an agent from an LLM and tools. Args: llm: Language model to use. tools: Tools to use. callback_manager: Callback manager to use. output_parser: Output parser to use. kwargs: Additional arguments. Returns: Agent: Agent object. )r`rrccSg|]}|jqSr9name.0toolr9r9r< z,Agent.from_llm_and_tools..)rrrNr9)rr0rr) rhr`rarcrrErZ tool_namesZ_output_parserr9r9r<ris  zAgent.from_llm_and_toolsrSrc Ks|dkr tddidS|dkrXd}|D]\}}||j7}|d|j|d|j7}q|d7}||jd}i||}|jjd i|} |j| } t | trQ| Std| i| St d |) aReturn response when agent has been stopped due to max iterations. Args: early_stopping_method: Method to use for early stopping. intermediate_steps: Steps the LLM has taken to date, along with observations. **kwargs: User inputs. Returns: AgentFinish: Agent finish object. Raises: ValueError: If `early_stopping_method` is not in ['force', 'generate']. rWr8rXrYgeneraterzB I now need to return a final answer based on the previous steps:rzBearly_stopping_method should be one of `force` or `generate`, got Nr9) rrrrrrrrrrpr]) r;rSrBrErrrrrrZ parsed_outputr9r9r<r_s4      zAgent.return_stopped_responsecCs|j|jdS)rrrr:r9r9r<rszAgent.tool_run_logging_kwargsrrr)rrTr3rT)rBrCr3rr@r)rBrCrErr3rr3r*r)rarbr3r)rarbr3rx)rErr3r)NN) r`rrarbrcrdrr rErr3rrr)rrrrrrrurArr=rrrrJrNrrRr)rrrrrrrrrir_rrr9r9rsr<rsZ              #7rc@sJeZdZUdZdZded< dZded< ddd d Z ddddZdS) ExceptionToolz!Tool that just returns the query. _ExceptionrTr zException tool descriptionNquery run_manager#Optional[CallbackManagerForToolRun]r3cCs|Sr@r9r;rrr9r9r<_runszExceptionTool._run(Optional[AsyncCallbackManagerForToolRun]cs|Sr@r9rr9r9r<_arun szExceptionTool._arunr@)rrTrrr3rT)rrTrrr3rT) rrrrr rrrrr9r9r9r<rs     rc@seZdZUdZded< ded< dZded< d Zd ed < d Zd ed< dZded< dZ ded< dZ ded< e ddddZ e dddd!d"Ze d#de dd&d'Zedd)d*Zdd.d/Zdd0d1Z dddd2dd7d8Zedd:d;Zedd.zAllowed tools (z!) different than provided tools (cSr r9r r r9r9r<rur))r rarArr])r;r rarr9r9r<validate_toolsas  zAgentExecutor.validate_toolsbeforevaluesrc Cs|d}|rJt|trJz|j}Wnty%}zd}WYd}~nd}~ww|ttttfk}| dd}|rBt ||d|d<|St ||d|d<|S)zConvert runnable to agent if passed in. Args: values: Values to validate. Returns: Dict: Validated values. r FNrT)rr) getrpr!Z OutputType Exceptionrr rrpoprr)rhr,r  output_type_Z multi_actionrr9r9r<validate_runnable_agentys&     z%AgentExecutor.validate_runnable_agent2Union[BaseSingleActionAgent, BaseMultiActionAgent]cCst|jtr tt|jS|jS)aCType cast self.agent. If the `agent` attribute is a Runnable, it will be converted one of RunnableAgentType in the validate_runnable_agent root_validator. To support instantiating with a Runnable, here we explicitly cast the type to reflect the changes made in the root_validator. )rpr r!rRunnableAgentTyper:r9r9r< _action_agents zAgentExecutor._action_agentrvrwrxcCr)zRaise error - saving not supported for Agent Executors. Args: file_path: Path to save to. Raises: ValueError: Saving not supported for agent executors. zpSaving not supported for agent executors. If you are trying to save the agent, please use the `.save_agent(...)`rr;rvr9r9r<rs zAgentExecutor.savecCs |j|S)zZSave the underlying agent. Args: file_path: Path to save to. )r5rr6r9r9r< save_agents zAgentExecutor.save_agent)include_run_infoasync_rr8r9r,cCst||||j|dS)agEnables iteration over steps taken to reach final output. Args: inputs: Inputs to the agent. callbacks: Callbacks to run. include_run_info: Whether to include run info. async_: Whether to run async. (Ignored) Returns: AgentExecutorIterator: Agent executor iterator object. )tagsr8)r,r:)r;rrDr8r9r9r9r<iterszAgentExecutor.iterr4cCs|jjS)rQ)r5rRr:r9r9r<rRszAgentExecutor.input_keyscCs|jr |jjdgS|jjS)z@Return the singular output key. :meta private: rB)r!r5r=r:r9r9r< output_keysszAgentExecutor.output_keysr r%cCsdd|jD|S)zLookup tool by name. Args: name: Name of tool. Returns: BaseTool: Tool object. cSi|]}|j|qSr9r r r9r9r< z-AgentExecutor.lookup_tool..)ra)r;r r9r9r< lookup_tools zAgentExecutor.lookup_tool iterationsint time_elapsedfloatcCs4|jdur ||jkr dS|jdur||jkrdSdS)NFT)r#r$)r;rArCr9r9r<_should_continues   zAgentExecutor._should_continuer8rrBrr$Optional[CallbackManagerForChainRun]rcCs.|r |j|d|jd|j}|jr||d<|SNgreen)colorverboserBZon_agent_finishrJr=r!r;r8rBrrr9r9r<_returns zAgentExecutor._return)Optional[AsyncCallbackManagerForChainRun]cs6|r|j|d|jdIdH|j}|jr||d<|SrGrKrLr9r9r<_areturns zAgentExecutor._areturnNextStepOutput1Union[AgentFinish, List[Tuple[AgentAction, str]]]cCs4t|dtrt|dksJ|dSdd|DS)Nr&cSs"g|] }t|tr|j|jfqSr9)rprrrrar9r9r<r%s   z4AgentExecutor._consume_next_step..)rprr)r;r,r9r9r<_consume_next_steps z AgentExecutor._consume_next_stepname_to_tool_mapDict[str, BaseTool] color_mappingDict[str, str]rCc Cs"|dd||||||DS)NcSsg|]}|qSr9r9rSr9r9r<r2sz1AgentExecutor._take_next_step..)rU_iter_next_stepr;rVrXrrBrr9r9r<_take_next_step)szAgentExecutor._take_next_step4Iterator[Union[AgentFinish, AgentAction, AgentStep]]c csz||}|jj|fd|r|ndi|}Wnty}zt|jtr-|j }nd}|r:tdt |t |} t|jtrU|j rRt |j } t |j } nd} nt|jt r_|j} nt |jrj||} ntdtd| | }|r}|j|dd |j} tj|jf|jd|r|ndd | } t|| d VWYd}~dSd}~wwt|tr|VdSt|tr|g} n|} | D]} | Vq| D] } |||| |VqdS) Take a single step in the thought-action-observation loop. Override this to take control of how the agent makes and acts on choices. rDNFAn output parsing error occurred. In order to pass this error back to the agent and have it try again, pass `handle_parsing_errors=True` to the AgentExecutor. This is the error: Invalid or incomplete response.Got unexpected type of `handle_parsing_errors`rrHrIrJrIrDrr)_prepare_intermediate_stepsr5rJ get_childrrpr%rr]rT send_to_llmr llm_outputcallableron_agent_actionrrr tool_inputrJrr_perform_agent_action)r;rVrXrrBrr8e raise_errorrrtool_run_kwargsactions agent_actionr9r9r<rZ>sz             ( zAgentExecutor._iter_next_steprqrrc Cs|r |j|dd|j|vr<||j}|j}||j}|j}|r&d|d<|j|jf|j||r4|ndd|} n"|j}t j|jt | df|jd|rW|ndd|} t || dS)NrHrbrYrrcZrequested_tool_nameZavailable_tool_namesrd) rjr return_directr5rrrkrJrfr.rkeysr r;rVrXrqrrrsrIrorr9r9r<rls>       z#AgentExecutor._perform_agent_actionc s*|dd||||||2IdHS)Ncsg|z3dHW}|q6Sr@r9rSr9r9r<rs z2AgentExecutor._atake_next_step..)rU_aiter_next_stepr[r9r9r<_atake_next_steps zAgentExecutor._atake_next_step9AsyncIterator[Union[AgentFinish, AgentAction, AgentStep]]c sz|}jj|fdrndi|IdH}Wnty}z~tjtr0j }nd}|r=tdt |t |} tjtrX|j rUt |j } t |j } nd} ntjt rbj} nt jrm|} ntdtd| | }j} tj|jfjdrndd| IdH} t|| d VWYd}~dSd}~wwt|tr|VdSt|tr|g} n|} | D]} | Vqtjfd d | DIdH}|D]}|VqdS) r^rDNFr_r`rarrcrdcsg|] }|qSr9)_aperform_agent_action)rrqrXrVrr;r9r<rs z2AgentExecutor._aiter_next_step..)rer5rNrfrrpr%rr]rTrgrrhrirrrrrkrJrrasynciogather)r;rVrXrrBrr8rmrnrrrorprqresultrr9rzr<rvs|             &   zAgentExecutor._aiter_next_stepc s|r|j||jddIdH|j|vrE||j}|j}||j}|j}|r,d|d<|j|jf|j||r:|ndd|IdH} n%|j}t j|jt | df|jd|r`|ndd|IdH} t || dS)NrH)rJrIrYrrcrrrd) rjrJrrsr5rrrkrfr.rrtrrur9r9r<rysD       z$AgentExecutor._aperform_agent_actionc Csdd|jD}tdd|jDddgd}g}d}d }t}|||rm|j|||||d } t| tr=|j| ||d S|| t | d kr]| d} | | } | d ur]|j| ||d S|d 7}t|}|||s&|j j |j |fi|} |j| ||d S) z(Run text through and get agent response.cSr=r9r r r9r9r<r>Lr?z'AgentExecutor._call..cSr r9r r r9r9r<rOrz'AgentExecutor._call..rHredZexcluded_colorsrrrRN)rar&timerEr\rprrMextendr_get_tool_returnr5r_rS r;rrrVrXrBrArC start_timenext_step_outputZnext_step_actionZ tool_returnr8r9r9r<_callEsJ       zAgentExecutor._callc sdd|jD}tdd|jDdgd}g}d}d}t}zt|j4Id H|||r|j|||||d Id H} t| trY|j | ||d Id HWd Id HWS| | t | d kr| d} | | } | d ur|j | ||d Id HWd Id HWS|d 7}t|}|||s1|j j|j|fi|} |j | ||d Id HWd Id HWS1Id HswYWd Sttjfy|j j|j|fi|} |j | ||d Id HYSw) z.Async run text through and get agent response.cSr=r9r r r9r9r<r>{r?z(AgentExecutor._acall..cSr r9r r r9r9r<r~rz(AgentExecutor._acall..rHrrrNrrR)rar&rr1r$rErwrprrOrrrr5r_rS TimeoutErrorr{rr9r9r<_acalltsr          4"zAgentExecutor._acallrTuple[AgentAction, str]Optional[AgentFinish]cCs`|\}}dd|jD}d}t|jjdkr|jjd}|j|vr.||jjr.t||idSdS)z&Check if the tool is a returning tool.cSr=r9r r r9r9r<r>r?z2AgentExecutor._get_tool_return..r8rrYN)rarr5r=rrsr)r;rrqrrVZreturn_value_keyr9r9r<rs   zAgentExecutor._get_tool_returncCs>t|jtr|jdkr||j dSt|jr||S|S)Nr)rpr'rBri)r;rBr9r9r<res    z)AgentExecutor._prepare_intermediate_stepsinputUnion[Dict[str, Any], Any]rOptional[RunnableConfig]Iterator[AddableDict]cks\t|}t|||df|d|d|d|ddd|}|D]}|Vq&dS) aEnables streaming over steps taken to reach final output. Args: input: Input to the agent. config: Config to use. kwargs: Additional arguments. Yields: AddableDict: Addable dictionary. rDr:metadatarun_namerun_idTr:rrrZ yield_actionsNr#r,r-r;rrrEiteratorstepr9r9r<rs$  zAgentExecutor.streamAsyncIterator[AddableDict]cKsft|}t|||df|d|d|d|ddd|}|2z 3dHW}|Vq&6dS) a Async enables streaming over steps taken to reach final output. Args: input: Input to the agent. config: Config to use. kwargs: Additional arguments. Yields: AddableDict: Addable dictionary. rDr:rrrTrNrrr9r9r<rs$  zAgentExecutor.astreamr@) r rrarbrDrrErr3rr)r,rr3r)r3r3r) rrrDrr8rr9rr3r,r)r rTr3r%)rArBrCrDr3r)r8rrBrrrFr3r)r8rrBrrrNr3r)r,rPr3rQ) rVrWrXrYrrYrBrCrrFr3rQ) rVrWrXrYrrYrBrCrrFr3r]) rVrWrXrYrqrrrFr3r) rVrWrXrYrrYrBrCrrNr3rQ) rVrWrXrYrrYrBrCrrNr3rx) rVrWrXrYrqrrrNr3r)rrYrrFr3r)rrYrrNr3rY)rrr3r)rBrCr3rC)rrrrrErr3r)rrrrrErr3r)(rrrrrr!r#r$rSr%r'rr(r)r*r2rr5rr7r;rRr<r@rErMrOrUr\rZrlrwrvryrrrrerrr9r9r9r<rs                   R -  Z , 2 >  "r)`r __future__rr{rloggingrabcrpathlibrtypingrrrrr r r r r rrrZlangchain_core._apirZlangchain_core.agentsrrrZlangchain_core.callbacksrrrrrrZlangchain_core.exceptionsrZlangchain_core.language_modelsrZlangchain_core.messagesrZlangchain_core.output_parsersrZlangchain_core.promptsrZlangchain_core.prompts.few_shotrZlangchain_core.prompts.promptr Zlangchain_core.runnablesr!r"r#Zlangchain_core.runnables.utilsr$Zlangchain_core.toolsr%Zlangchain_core.utils.inputr&Zpydanticr'r(r)Ztyping_extensionsr*Zlangchain._api.deprecationr+Zlangchain.agents.agent_iteratorr,Zlangchain.agents.agent_typesr-Zlangchain.agents.toolsr.Zlangchain.chains.baser/Zlangchain.chains.llmr0Zlangchain.utilities.asyncior1 getLoggerrrr2rrrrrrrrrPr4rr9r9r9r<s|   4                     + muS&