NgdZddlmZddlZddlZddlZddlZddlmZddl 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@ddlAmBZBddlCmDZDddlEmFZFddlGmHZHddlImJZJddlKmLZLddlMmNZNddlOmPZPejQeRZSGdde>ZTGd d!e>ZUGd"d#e,eeefZVGd$d%e,eeeefZWGd&d'eTZXGd(d)eUZYed*eDd+,Gd-d.eTZZed*eDd+,Gd/d0eTZ[Gd1d2e:Z\eeeeefZ]eeXeYfZ^Gd3d4eLZ_dS)5zEChain that takes in an input and produces an action and action input.) annotationsN)abstractmethod)Path) Any AsyncIteratorCallableDictIteratorListOptionalSequenceTupleUnioncast) deprecated) AgentAction AgentFinish AgentStep)AsyncCallbackManagerForChainRunAsyncCallbackManagerForToolRunBaseCallbackManagerCallbackManagerForChainRunCallbackManagerForToolRun Callbacks)OutputParserException)BaseLanguageModel) BaseMessage)BaseOutputParser)BasePromptTemplate)FewShotPromptTemplate)PromptTemplate)RunnableRunnableConfig ensure_config) AddableDict)BaseTool)get_color_mapping) BaseModel ConfigDictmodel_validator)Self)AGENT_DEPRECATION_WARNING)AgentExecutorIterator) AgentType) InvalidTool)Chain)LLMChain)asyncio_timeoutceZdZdZed%dZd&dZe d'd(dZe d'd(dZ eed%dZ d)dZ e d'd*dZ ed+dZd,fd Zd-d#Zd.d$ZxZS)/BaseSingleActionAgentzBase Single Action Agent class.return List[str]cdgSReturn values of the agent.outputselfs R/var/www/html/ai-engine/env/lib/python3.11/site-packages/langchain/agents/agent.py return_valuesz#BaseSingleActionAgent.return_values@ zOptional[List[str]]cdSNr;r<s r>get_allowed_toolsz'BaseSingleActionAgent.get_allowed_toolsEstrANintermediate_stepsList[Tuple[AgentAction, str]] callbacksrkwargsrUnion[AgentAction, AgentFinish]c dS)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. Nr;r=rFrHrIs r>planzBaseSingleActionAgent.planHrAc KdS)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. Nr;rMs r>aplanzBaseSingleActionAgent.aplan[ rAcdS7Return the input keys. :meta private: Nr;r<s r> input_keysz BaseSingleActionAgent.input_keysnrOrAearly_stopping_methodstrrc X|dkrtddidStd|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. forcer:3Agent stopped due to iteration limit or time limit.'Got unsupported early_stopping_method ``r ValueErrorr=rXrFrIs r>return_stopped_responsez-BaseSingleActionAgent.return_stopped_responsevsP( !G + +PQSU R:ORRR rAllmrtoolsSequence[BaseTool]callback_managerOptional[BaseCallbackManager]c t)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)clsrerfrhrIs r>from_llm_and_toolsz(BaseSingleActionAgent.from_llm_and_toolss &"!rActz#Return Identifier of an agent type.rkr<s r> _agent_typez!BaseSingleActionAgent._agent_type "!rAr c t} |j}n#t$rd}YnwxYwt |t rt |j|d<n|||d<|S)z{Return dictionary representation of agent. Returns: Dict: Dictionary representation of agent. N_type)super model_dumprqrl isinstancer.rYvalue)r=rI_dictrt __class__s r>dictzBaseSingleActionAgent.dicts ""$$ $EE"   EEE  eY ' ' # --E'NN  "E'N s + :: file_pathUnion[Path, str]NonecXt|trt|}n|}|j}|dd|}d|vrt d|d|jdkrBt|d5}tj ||d d d d d S#1swxYwYd S|j d rBt|d5}tj ||d d d d d S#1swxYwYd 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_okrtAgent z does not support saving.jsonwindentNz.yamlz.ymlFdefault_flow_style must be json or yaml)rwrYrparentmkdirr{rlsuffixopenjsondumpendswithyamlrb)r=r| save_pathdirectory_path agent_dictfs r>savezBaseSingleActionAgent.saves i % % "YII!I")TD999YY[[ * $ $%&Mt&M&M&MNN N  w & &i%% 3 *a2222 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3   & &'8 9 9 Bi%% C *aEBBBB C C C C C C C C C C C C C C C C C C @@@AA A$ B11B58B5(D  DDciS#Return logging kwargs for tool run.r;r<s r>tool_run_logging_kwargsz-BaseSingleActionAgent.tool_run_logging_kwargss rAr5r6r5rBrDrFrGrHrrIrr5rJrXrYrFrGrIrr5r) rerrfrgrhrirIrr5r4r5rYrIrr5r r|r}r5r~r5r )__name__ __module__ __qualname____doc__propertyr?rErrNrRrWrd classmethodrnrqr{rr __classcell__rzs@r>r4r4=s~)) X $    ^ $ $    ^ $   ^X < ;? """"["("""X""!B!B!B!BFrAr4ceZdZdZeddZddZe d d!dZe d d!dZ eeddZ d"dZ ed#dZ d$fd Z d%dZd&dZxZS)'BaseMultiActionAgentzBase Multi Action Agent class.r5r6cdgSr8r;r<s r>r?z"BaseMultiActionAgent.return_valuesr@rArBcdS)z]Get allowed tools. Returns: Optional[List[str]]: Allowed tools. Nr;r<s r>rEz&BaseMultiActionAgent.get_allowed_toolss trANrFrGrHrrIr%Union[List[AgentAction], AgentFinish]c dS)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. Nr;rMs r>rNzBaseMultiActionAgent.planrOrAc KdS)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. Nr;rMs r>rRzBaseMultiActionAgent.aplan rSrAcdSrUr;r<s r>rWzBaseMultiActionAgent.input_keysrOrArXrYrc X|dkrtddidStd|d)r[r\r:z$Agent stopped due to max iterations.r^r_r`rarcs r>rdz,BaseMultiActionAgent.return_stopped_response%sG( !G + +*PQSUVV VR:ORRR rActrprkr<s r>rqz BaseMultiActionAgent._agent_typeArrrAr c t} t|j|d<n#t$rYnwxYw|S)*Return dictionary representation of agent.rt)rurvrYrqrlr=rIryrzs r>r{zBaseMultiActionAgent.dictFsX""$$  !122E'NN"    D  s; AAr|r}r~cXt|trt|}n|}|}d|vrt d|d|j}|dd|jdkrBt|d5}tj ||d d d d d S#1swxYwYd S|j d rBt|d5}tj ||d d d d d S#1swxYwYd 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") rtrz does not support saving.TrrrrrNrFrr)rwrYrr{rlrrrrrrrrrb)r=r|rrrrs r>rzBaseMultiActionAgent.saveOs" i % % "YII!IYY[[ * $ $%&Nt&N&N&NOO O")TD999  w & &i%% 3 *a2222 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3   & &'8 9 9 Bi%% C *aEBBBB C C C C C C C C C C C C C C C C C C @@@AA ArciSrr;r<s r>rz,BaseMultiActionAgent.tool_run_logging_kwargsvs  rArrrDrFrGrHrrIrr5rrrrrr)rrrrrr?rErrNrRrWrdrqr{rrrrs@r>rrsT(( X $    ^ $ $    ^ $   ^X 8"""X"%B%B%B%BNrArc*eZdZdZeddZdS) AgentOutputParserz=Base class for parsing agent output into agent action/finish.textrYr5rJcdS)z$Parse text into agent action/finish.Nr;r=rs r>parsezAgentOutputParser.parserOrAN)rrYr5rJrrrrrrr;rAr>rr|s8GG333^333rArc*eZdZdZeddZdS) MultiActionAgentOutputParserzBase class for parsing agent output into agent actions/finish. This is used for agents that can return multiple actions. rrYr5rcdS)zParse text into agent actions/finish. Args: text: Text to parse. Returns: Union[List[AgentAction], AgentFinish]: List of agent actions or agent finish. Nr;rs r>rz"MultiActionAgentOutputParser.parserOrAN)rrYr5rrr;rAr>rrs>     ^    rArceZdZUdZded< gZded<gZded<dZded < ed Z e dd Z e dd Z dddZ dddZdS) RunnableAgentAgent powered by Runnables.z/Runnable[dict, Union[AgentAction, AgentFinish]]runnabler6input_keys_argreturn_keys_argTboolstream_runnablearbitrary_types_allowedr5c|jSr9rr<s r>r?zRunnableAgent.return_values ##rAc|jS)zReturn the input keys.rr<s r>rWzRunnableAgent.input_keyss ""rANrFrGrHrrIrrJc i|d|i}d}|jr,|j|d|iD] }||}||z } n|j|d|i}|SaPBased on past history and current inputs, decide what to do. Args: intermediate_steps: Steps the LLM has taken to date, along with the observations. callbacks: Callbacks to run. **kwargs: User inputs. Returns: Action specifying what tool to use. rFNrHconfigrrstreaminvoker=rFrHrIinputs final_outputchunks r>rNzRunnableAgent.plans"JFI35GHI   Y--fk9=U-VV * *'#(LL E)LL  *  =// Y?W/XXLrAcKi|d|i}d}|jr2|j|d|i23d{V}||} ||z }6n$|j|d|id{V}|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. rFNrHrrrastreamainvokers r>rRzRunnableAgent.aplans(JFI35GHI    $}44 Y7 5   * * * * * * *e '#(LL E)LL   "&!6!6 Y7"7""LArrDrrrrr__annotations__rrrr) model_configrr?rWrNrRr;rAr>rrs%%====/ "N""""!#O#### O    : $L$$$X$###X# $"""""N $(((((((rArceZdZUdZded< gZded<gZded<dZded < ed Z e dd Z e dd Z dddZ dddZdS)RunnableMultiActionAgentrz5Runnable[dict, Union[List[AgentAction], AgentFinish]]rr6rrTrrrr5c|jSrrr<s r>r?z&RunnableMultiActionAgent.return_valuesrrAc|jS)QReturn the input keys. Returns: List of input keys. rr<s r>rWz#RunnableMultiActionAgent.input_keyss ""rANrFrGrHrrIrrc i|d|i}d}|jr,|j|d|iD] }||}||z } n|j|d|i}|Srrrs r>rNzRunnableMultiActionAgent.plan(s(JFI35GHI   Y--fk9=U-VV * *'#(LL E)LL  *  =// Y?W/XXLrAcKi|d|i}d}|jr2|j|d|i23d{V}||} ||z }6n$|j|d|id{V}|Srrrs r>rRzRunnableMultiActionAgent.aplanOs(JFI35GHI    $}44 Y7 5   * * * * * * *e '#(LL E)LL   "&!6!6 Y7"7""LrrrDrrr;rAr>rrs%%CCCC0 "N""""!#O#### O    : $L$$$X$###X# $%%%%%T $)))))))rArz0.1.0z1.0)messageremovalceZdZUdZded< ded< ded< edd Zdfd Z dddZ dddZ ddZ xZ S)LLMSingleActionAgentz$Base class for single action agents.r1 llm_chainr output_parserr6stopr5cVtt|jjdhz S)rrFlistsetrrWr<s r>rWzLLMSingleActionAgent.input_keyss(C1226J5KKLLLrArIrr c Nt}|d=|Srrrur{rs r>r{zLLMSingleActionAgent.dict!  / " rANrFrGrHrrJc l|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. rFrrHr;)rrunrrrr=rFrHrIr:s r>rNzLLMSingleActionAgent.plansR"$# 1      !''///rAc|K|jjd||j|d|d{V}|j|S)rQrNr;)rarunrrrrs r>rRzLLMSingleActionAgent.aplanst"+t~* 1            !''///rAcVdt|jdkrdn |jddS)rr^r llm_prefixobservation_prefix)lenrr<s r>rz,LLMSingleActionAgent.tool_run_logging_kwargss5(+DI!(;(;""1   rArrrDrr) rrrrrrrWr{rNrRrrrs@r>rr{s /.$$$$$)OOO% MMMXM $000008 $000002        rArceZdZUdZded< ded< dZded< d8fd Zd9dZed:dZ d;dZ ed:dZ ddZ d=d>dZd?dZed:d Zed!"d@d$ZeedAd%ZeedAd&ZeedBd*ZedCd,ZeedDd-Ze dEdFd3ZdGd6ZdHd7ZxZS)IAgentzAgent that calls the language model and deciding the action. This is driven by a LLMChain. The prompt in the LLMChain MUST include a variable called "agent_scratchpad" where the agent can put its intermediary work. r1rrrNrB allowed_toolsrIrr5r c Nt}|d=|Srrrs r>r{z Agent.dictrrAc|jS)zGet allowed tools.)rr<s r>rEzAgent.get_allowed_toolss !!rAr6cdgSr8r;r<s r>r?zAgent.return_valuesr@rArrYc td)zsFix the text. Args: text: Text to fix. Returns: str: Fixed text. z(fix_text not implemented for this agent.rbrs r> _fix_textzAgent._fix_textsCDDDrAcrd|jd|jgS)N z )r rstripr<s r>_stopz Agent._stopsB 4(//11 3 3 54*1133 5 5  rArFrGUnion[str, List[BaseMessage]]c\d}|D]&\}}||jz }|d|j|d|jz }'|S)zJConstruct the scratchpad that lets the agent continue its thought process.r^r)logr r )r=rFthoughtsaction observations r>_construct_scratchpadzAgent._construct_scratchpadsW#5 W W FK  "H VT4VkVVT_VV VHHrArHrrJc z|j|fi|}|jjdd|i|}|j|S)rLrHr;)get_full_inputsrpredictrr)r=rFrHrI full_inputs full_outputs r>rNz Agent.plansU"+d*+=HHHH ,dn,PPyPKPP !'' 444rAcK|j|fi|}|jjdd|i|d{V}|j|d{V}|S)rQrHNr;)r#rapredictraparse)r=rFrHrIr%r& agent_outputs r>rRz Agent.aplan's"+d*+=HHHH 3DN3WWiW;WWWWWWWW !/66{CCCCCCCC rADict[str, Any]c P||}||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)r!r)r=rFrIr new_inputsr%s r>r#zAgent.get_full_inputs=s<--.@AA*2DJGG ..:. rAcVtt|jjdhz S)rVr.rr<s r>rWzAgent.input_keysOs( C1226H5IIJJJrAaftermoder+cn|jj}d|jvrtd|jdt |tr|xjdz c_nEt |tr|xj dz c_ ntdt||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. r.zl`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 ) rpromptinput_variablesloggerwarningappendrwr!templater rrbtype)r=r5s r>validate_promptzAgent.validate_promptWs& V%; ; ; NN=     " ) )*< = = =&.11 O#99F$9:: O !77  !MtF||!M!MNNN rAcdS)z&Prefix to append the observation with.Nr;r<s r>r zAgent.observation_prefixtrOrAcdS)z#Prefix to append the LLM call with.Nr;r<s r>r zAgent.llm_prefixyrOrArfrgrcdS)zCreate a prompt for this class. Args: tools: Tools to use. Returns: BasePromptTemplate: Prompt template. Nr;rmrfs r> create_promptzAgent.create_prompt~rOrAr~cdS)zgValidate that appropriate tools are passed in. Args: tools: Tools to use. Nr;r@s r>_validate_toolszAgent._validate_toolss  rAc dS)z)Get default output parser for this class.Nr;)rmrIs r>_get_default_output_parserz Agent._get_default_output_parserrOrArerrhriOptional[AgentOutputParser]c ||t||||}d|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. )rer5rhcg|] }|j Sr;name.0tools r> z,Agent.from_llm_and_tools..s222Ddi222rA)rrrr;)rCr1rArE) rmrerfrhrrIr tool_names_output_parsers r>rnzAgent.from_llm_and_toolss* E"""$$U++-   32E222 &J#*H*H*J*Js $(      rArXrc |dkrtddidS|dkrd}|D]&\}}||jz }|d|j|d|jz }'|dz }||jd}i||}|jjd i|} |j| } t| tr| Std| i| Std |) aReturn response when agent has been stopped due to max iterations. Args: early_stopping_method: Method to use for early stopping. intermediate_steps: Steps the LLM has taken to date, along with observations. **kwargs: User inputs. Returns: AgentFinish: Agent finish object. Raises: ValueError: If `early_stopping_method` is not in ['force', 'generate']. r\r:r]r^generaterzB I now need to return a final answer based on the previous steps:r-zBearly_stopping_method should be one of `force` or `generate`, got Nr;) rrr r rrr$rrrwrb) r=rXrFrIrrr r/r%r& parsed_outputs r>rdzAgent.return_stopped_responses;( !G + +PQSU #j 0 0H'9  # FJ&R0R+RRRR V H/7 KKJ2V2z2K0$.0??;??K .44[AAM-55 I$$#Hk#:KHHH/,// rAc |j|jdS)rr r r<s r>rzAgent.tool_run_logging_kwargss/"&"9   rArrr)rrYr5rY)rFrGr5rrDr)rFrGrIrr5r+r5r+r)rfrgr5r)rfrgr5r~)rIrr5r)NN) rerrfrgrhrirrFrIrr5rrr)rrrrrrr{rErr?rrr!rNrRr#rWr*r<rr r rrArCrErnrdrrrs@r>rrs $$$$$))-M----F """"X E E E E   X  $555550 $,$KKKXK_'"""#"8555^X5222^X2   ^[    [ 888^[8 ;?59 ! ! ! ! [! F5555n        rArcLeZdZUdZdZded< dZded< ddd Z dddZdS) ExceptionToolz!Tool that just returns the query. _ExceptionrYrJzException tool descriptionNquery run_manager#Optional[CallbackManagerForToolRun]r5c|SrDr;r=rZr[s r>_runzExceptionTool._runs  rA(Optional[AsyncCallbackManagerForToolRun]c K|SrDr;r^s r>_arunzExceptionTool._arun s  rArD)rZrYr[r\r5rY)rZrYr[r`r5rY) rrrrrJrrYr_rbr;rAr>rWrWs++D'K''''" <@AErArWc4eZdZUdZded< ded< dZded< d Zd ed < d Zd ed< dZded< dZ ded< dZ ded< e didjdZ e ddkd Ze d!e dld$Zedmd&Zdnd*Zdnd+Z diddd,dod1Zedpd3Zedpd4Zdqd7Zdrd<Z didsdDZ didtdFZdudIZ didvdOZ didwdQZ didxdUZ didydVZ didzdXZ did{dYZ! did|dZZ" did}d[Z#d~d_Z$dd`Z% diddfZ& diddhZ'd S) AgentExecutorzAgent that is using tools.from_agent_and_toolsz"AgentExecutor.from_agent_and_toolsGs5&s       rAr1r2r+c|j}|j}|}|Jt|td|Dkr t d|dd|Dd|S)aValidate that tools are compatible with agent. Args: values: Values to validate. Returns: Dict: Validated values. Raises: ValueError: If allowed tools are different than provided tools. Ncg|] }|j Sr;rIrKs r>rNz0AgentExecutor.validate_tools..rs)F)F)F$))F)F)FrAzAllowed tools (z!) different than provided tools (cg|] }|j Sr;rIrKs r>rNz0AgentExecutor.validate_tools..us'D'D'Dd 'D'D'DrA))rfrfrErrb)r=rfrfrs r>validate_toolszAgentExecutor.validate_toolsas  //11  $=!!S)F)F)F)F)F%G%GGG HmHH'D'De'D'D'DHHH rAbeforevaluesr ct|d}|rt|tr |j}|tt t tfk}n#t$r }d}Yd}~nd}~wwxYw| dd}|rt|||d<nt|||d<|S)zConvert runnable to agent if passed in. Args: values: Values to validate. Returns: Dict: Validated values. rfFNrT)rr) getrwr" OutputTyperr rr Exceptionpoprr)rmrvrf output_type multi_action_rs r>validate_runnable_agentz%AgentExecutor.validate_runnable_agentys 7##  Zx00  T#.  +eD4E{4R.SS  % % %$  % %jj):DAAO ":"O###w#0"O###w sA A4(A//A42Union[BaseSingleActionAgent, BaseMultiActionAgent]cxt|jtrtt|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. )rwrfr"rRunnableAgentTyper<s r> _action_agentzAgentExecutor._action_agents3 dj( + + )4:66 6: rAr|r}r~c td)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=r|s r>rzAgentExecutor.saves !   rAc6|j|S)zZSave the underlying agent. Args: file_path: Path to save to. )rrrs r> save_agentzAgentExecutor.save_agents !&&y111rA)include_run_infoasync_rrrr-c4t||||j|S)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. )tagsr)r-r)r=rrHrrs r>iterzAgentExecutor.iters,&%   -     rAr6c|jjS)rV)rrWr<s r>rWzAgentExecutor.input_keyss !,,rAcH|jr|jjdgzS|jjS)z@Return the singular output key. :meta private: rF)rgrr?r<s r> output_keyszAgentExecutor.output_keyss0  ) 4%37K6LL L%3 3rArJr&c0d|jD|S)zLookup tool by name. Args: name: Name of tool. Returns: BaseTool: Tool object. ci|] }|j| Sr;rIrKs r> z-AgentExecutor.lookup_tool..s777D 4777rA)rf)r=rJs r> lookup_toolzAgentExecutor.lookup_tools87DJ777==rA iterationsint time_elapsedfloatcV|j ||jkrdS|j ||jkrdSdS)NFT)rirj)r=rrs r>_should_continuezAgentExecutor._should_continues>   *zT=P/P/P5  # / 7775trAr:rrFrr[$Optional[CallbackManagerForChainRun]r+cj|r||d|j|j}|jr||d<|SNgreen)colorverboserFon_agent_finishrr?rgr=r:rFr[rs r>_returnzAgentExecutor._returnsO  U  ' 'gt| ' T T T+  ) D1CL- .rA)Optional[AsyncCallbackManagerForChainRun]czK|r#||d|jd{V|j}|jr||d<|Srrrs r>_areturnzAgentExecutor._areturns{  --gt|.       +  ) D1CL- .rANextStepOutput1Union[AgentFinish, List[Tuple[AgentAction, str]]]ct|dtrt|dksJ|dSd|DS)NrlcTg|]%}t|t|j|jf&Sr;)rwrrr rLas r>rNz4AgentExecutor._consume_next_step..%s@./Z9=U=U1=)rA)rwrr )r=rvs r>_consume_next_stepz AgentExecutor._consume_next_steps\ fRj+ . . v;;!####": 39 rAname_to_tool_mapDict[str, BaseTool] color_mappingDict[str, str]rGc n|d||||||DS)Ncg|]}|Sr;r;rs r>rNz1AgentExecutor._take_next_step..2s(      rA)r_iter_next_stepr=rrrrFr[s r>_take_next_stepzAgentExecutor._take_next_step)sW&&  --$!&       rA4Iterator[Union[AgentFinish, AgentAction, AgentStep]]c#K ||}|jj|fd|r|ndi|}n#t$r}t |jtr |j }nd}|rtdt|t|} t |jtr3|j r)t|j } t|j } n^d} n[t |jtr|j} n9t|jr||} ntdtd| | }|r||d |j} t#j|jf|jd|r|ndd | } t+|| VYd}~dSd}~wwxYwt |t,r|VdSt |tr|g} n|} | D]} | V| D]} |||| |VdS) Take a single step in the thought-action-observation loop. Override this to take control of how the agent makes and acts on choices. rHNFAn 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`rXrrrrrHrr )_prepare_intermediate_stepsrrN get_childrrwrkrrbrY send_to_llmr  llm_outputcallableron_agent_actionrrWr tool_inputrrr_perform_agent_action)r=rrrrFr[r:e raise_errorrr tool_run_kwargsactions agent_actions r>rzAgentExecutor._iter_next_step>s. !%!A!ABT!U!U -T',"5@J+//111dFF %% % % $4d;; $"&"<< #   3+.a&&33 q66D$4d;; S=C"%am"4"4Kq|,,DD"BKKD6<< S"8 $455 S"88;;  !QRRR {DAAF C++F'+BBB"0HHJJO--//-! 5@J+//111d  " K6{CCC C C C FFFFFK% P fk * * LLL F fk * * hGGG#  L    #  L,, -{      sAA G-FG((G-rrrcB|r||d|j|vrt||j}|j}||j}|j}|rd|d<|j|jf|j||r|ndd|} ny|j}tj|jt| df|jd|r|ndd|} t|| S)Nrrr^r rrequested_tool_nameavailable_tool_namesr) rrM return_directrrrrrrr/rkeysr r=rrrr[rMrrrr s r>rz#AgentExecutor._perform_agent_actionso  E  ' ' G ' D D D   0 0 0#L$56D .M!,"34E"0HHJJO 302 -"$(' 5@J+//111d  " KK#0HHJJO++--++7+<,01A1F1F1H1H,I,I   5@J+//111d  "  K +FFFFrAc ~K|d||||||2d{VS)Nc$Kg|3d{V }| 6SrDr;rs r>rNz2AgentExecutor._atake_next_step..sO             s)r_aiter_next_steprs r>_atake_next_stepzAgentExecutor._atake_next_stepsy&&  #44$!&               rA9AsyncIterator[Union[AgentFinish, AgentAction, AgentStep]]cK |}jj|fdrndi|d{V}n#t$r}t jtr j }nd}|rtdt|t|} t jtr3|j r)t|j } t|j } n^d} n[t jtrj} n9tjr|} ntdtd| | }j} t!j|jfjdrndd| d{V} t)|| WVYd}~dSd}~wwxYwt |t*r|WVdSt |tr|g} n|} | D]} | WVt-jfd | Dd{V}|D]}|WVdS) rrHNFrrrrXrrc@g|]}|Sr;)_aperform_agent_action)rLrrrr[r=s r>rNz2AgentExecutor._aiter_next_step..sD!++$m\;rA)rrrRrrrwrkrrbrYrr rrrrrWrrrrrasynciogather)r=rrrrFr[r:rrrr rrrresultrs``` ` r>rzAgentExecutor._aiter_next_stepsk, !%!A!ABT!U!U 44-3"5@J+//111dFF %# # # $4d;; $"&"<< #   3+.a&&33 q66D$4d;; S=C"%am"4"4Kq|,,DD"BKKD6<< S"8 $455 S"88;;  !QRRR {DAAF"0HHJJO 4  4!! 5@J+//111d !! " !!K6{CCC C C C C FFFFFG# L fk * * LLLL F fk * * hGGG#  L     ~%,          EKKKKK  sAA G%E?G  G%cvK|r#|||jdd{V|j|vrz||j}|j}||j}|j}|rd|d<|j|jf|j||r|ndd|d{V} n|j}tj|jt| df|jd|r|ndd|d{V} t|| S)Nr)rrr^r rrr) rrrMrrrrrrr/rrrrs r>rz$AgentExecutor._aperform_agent_actions  --dl'.          0 0 0#L$56D .M!,"34E"0HHJJO 302 - ) '! 5@J+//111d !! " !!KK#0HHJJO 2 2+7+<,01A1F1F1H1H,I,I !  5@J+//111d ! !" ! !      K +FFFFrAcd|jD}td|jDddg}g}d}d}tj}|||r||||||} t | t r|| ||S|| t| d kr7| d} | | } | || ||S|d z }tj|z }||||j j |j |fi|} || ||S) z(Run text through and get agent response.ci|] }|j| Sr;rIrKs r>rz'AgentExecutor._call..LCCCDItCCCrAcg|] }|j Sr;rIrKs r>rNz'AgentExecutor._call..O . . .4TY . . .rArredexcluded_colorsrr[r)rfr'timerrrwrrextendr _get_tool_returnrrdrX r=rr[rrrFrr start_timenext_step_outputnext_step_action tool_returnr:s r>_callzAgentExecutor._callEsDC CCC) . .4: . . .%@P   =?  Y[[ ##J == 4#33 "' 4   *K88 ||$&8k$  % %&6 7 7 7#$$))#3A#6 "334DEE *<<#%7[( !OJ9;;3L/##J == 40<#;  &(:  >D  ||F$6K|PPPrAc|Kd|jD}td|jDdg}g}d}d}tj} t|j4d{V|||r||||||d{V} t| tr0| | ||d{Vcdddd{VS| | t| d krO| d} | | } | 0| | ||d{Vcdddd{VS|d z }tj|z }||||j j|j|fi|} | | ||d{Vcdddd{VS#1d{VswxYwYdS#t t"jf$r:|j j|j|fi|} | | ||d{VcYSwxYw) z.Async run text through and get agent response.ci|] }|j| Sr;rIrKs r>rz(AgentExecutor._acall..{rrAcg|] }|j Sr;rIrKs r>rNz(AgentExecutor._acall..~rrArrrrNrr)rfr'rr2rjrrrwrrrr rrrdrX TimeoutErrorrrs r>_acallzAgentExecutor._acalltsIDC CCC) . .4: . . .    =?  Y[[ * &t'>??! ! ! ! ! ! ! ! ++J EE<-1-B-B(%*$/ .C..(((((($""2K@@%)]],.(3&3&&      ! ! ! ! ! ! ! ! ! ! ! ! ! !  '--.>???+,,11+;A+>(&*&;&;BHF* '     sRG+)A*GG+&A$G G+A(GG+ G""G+%G"&G++A H;:H;rTuple[AgentAction, str]Optional[AgentFinish]c|\}}d|jD}d}t|jjdkr|jjd}|j|vr$||jjrt ||idSdS)z&Check if the tool is a returning tool.ci|] }|j| Sr;rIrKs r>rz2AgentExecutor._get_tool_return..rrAr:rr^N)rfr rr?rMrr)r=rrr rreturn_value_keys r>rzAgentExecutor._get_tool_returns%5! kCC CCC# t!/ 0 01 4 4#1?B    0 0 0 12@ "%{3trAct|jtr|jdkr||j dSt|jr||S|S)Nr)rwrmrr)r=rFs r>rz)AgentExecutor._prepare_intermediate_stepssl t3S 9 9 &,q00%t'C&C&E&EF F d2 3 3 &//0BCC C% %rAinputUnion[Dict[str, Any], Any]rOptional[RunnableConfig]Iterator[AddableDict]c +"Kt|}t|||df|d|d|d|ddd|}|D]}|VdS) 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. rHrmetadatarun_namerun_idTrrrr yield_actionsNr$r-rxr=rrrIiteratorsteps r>rzAgentExecutor.streams v&&(   JJ{ # #  F##ZZ ++ZZ ++::h''        DJJJJ  rAAsyncIterator[AddableDict]c 0Kt|}t|||df|d|d|d|ddd|}|23d{V}|WV 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. rHrrrrTrNr r s r>rzAgentExecutor.astreams"v&&(   JJ{ # #  F##ZZ ++ZZ ++::h''      #       $JJJJJ#((sBrD) rfrerfrgrHrrIrr5rdrU)rvr r5r)r5rr) rrrHrrrrrr5r-r)rJrYr5r&)rrrrr5r)r:rrFrr[rr5r+)r:rrFrr[rr5r+)rvrr5r) rrrrrrrFrGr[rr5r) rrrrrrrFrGr[rr5r) rrrrrrr[rr5r) rrrrrrrFrGr[rr5r) rrrrrrrFrGr[rr5r) rrrrrrr[rr5r)rrr[rr5r+)rrr[rr5r)rrr5r)rFrGr5rG)rrrrrIrr5r)rrrrrIrr5r )(rrrrrrgrirjrXrkrmrror*rtrrrrrrrWrrrrrrrrrrrrrrrrrrr;rAr>rdrds$$$GGGG3-&+++++3$&N&&&&9+/....")((((       $     [ 2_'"""#"._(###[$#:   X      2222 $ "'       6---X-444X4 > > > >    =A     "BF         "=A      6=A KKKKKd=A %G%G%G%G%GZBF      6BF SSSSStBF 'G'G'G'G'GX=A-Q-Q-Q-Q-QdBF<<<<<|$ & & & & ,0D,0rArd)`r __future__rrrloggingrabcrpathlibrtypingrrrr r r r r rrrrlangchain_core._apirlangchain_core.agentsrrrlangchain_core.callbacksrrrrrrlangchain_core.exceptionsrlangchain_core.language_modelsrlangchain_core.messagesrlangchain_core.output_parsersrlangchain_core.promptsrlangchain_core.prompts.few_shotr langchain_core.prompts.promptr!langchain_core.runnablesr"r#r$langchain_core.runnables.utilsr%langchain_core.toolsr&langchain_core.utils.inputr'pydanticr(r)r*typing_extensionsr+langchain._api.deprecationr,langchain.agents.agent_iteratorr-langchain.agents.agent_typesr.langchain.agents.toolsr/langchain.chains.baser0langchain.chains.llmr1langchain.utilities.asyncior2 getLoggerrr7r4rrrrrrrrWrrrdr;rAr>r,sKK""""""                             ******EEEEEEEEEE<;;;;;<<<<<<//////::::::555555AAAAAA888888LLLLLLLLLL666666))))))888888;;;;;;;;;;""""""@@@@@@AAAAAA222222......''''''))))))777777  8 $ $gggggIgggTRRRRR9RRRj33333({K/G)HI333     U4 ,k9:;   *jjjjj)jjjZrrrrr3rrrj  %  P P P P P 0P P  P f  %  a a a a a !a a  a H H.eKi?@A-)AABw w w w w Ew w w w w rA