fi hdZddlmZddlZddlmZmZmZmZm Z m Z m Z ddl m Z ddlmZmZmZmZddlmZddlmZdd lmZmZdd lmZdd lmZmZdd lm Z m!Z!dd l"m#Z#m$Z$m%Z%ddl&m'Z'm(Z(m)Z)m*Z*m+Z+m,Z,m-Z-m.Z.m/Z/m0Z0ddl1m2Z2m3Z3m4Z4m5Z5m6Z6m7Z7m8Z8m9Z9m:Z:ddl;mddl=m>Z>m?Z?m@Z@ddlAmBZBmCZCddlDmEZEddlmFZFddlGmHZHddlImJZJddlmKZKddlLmMZMddl&mNZNmOZOdZPdjZRgdZSd2dZT d3dZU d4dZVd5d6d ZWd7d!ZXd8d"ZY d5 d9d#ZZ d:d$Z[ d;d%Z\ dd+Z_ d?d,Z` d@d-Za dAd.Zb dBd/Zc dCd0Zdd1gZey)Dz:Agent factory for creating agents with middleware support.) annotationsN) TYPE_CHECKING AnnotatedAnycastget_args get_originget_type_hints) BaseChatModel) AIMessage AnyMessage SystemMessage ToolMessage)BaseTool)RunnableCallable)ENDSTART) StateGraph)ToolCallWithContextToolNode)CommandSend) NotRequiredRequired TypedDict) AgentMiddleware AgentStateJumpTo ModelRequest ModelResponseOmitFromSchema ResponseT StateT_co_InputAgentState_OutputAgentState) AutoStrategyMultipleStructuredOutputsErrorOutputToolBindingProviderStrategyProviderStrategyBindingResponseFormatStructuredOutputErrorStructuredOutputValidationError ToolStrategy)init_chat_model) AwaitableCallableSequence)RunnableRunnableConfig) BaseCache)CompiledStateGraph)Runtime) BaseStore) Checkpointer)ContextT)ToolCallRequestToolCallWrapperz)Error: {error} Please fix your mistakes.a Middleware added tools that the agent doesn't know how to execute. Unknown tools: {unknown_tool_names} Registered tools: {available_tool_names} This happens when middleware modifies `request.tools` in `wrap_model_call` to include tools that weren't passed to `create_agent()`. How to fix this: Option 1: Register tools at agent creation (recommended for most cases) Pass the tools to `create_agent(tools=[...])` or set them on `middleware.tools`. This makes tools available for every agent invocation. Option 2: Handle dynamic tools in middleware (for tools created at runtime) Implement `wrap_tool_call` to execute tools that are added dynamically: class MyMiddleware(AgentMiddleware): def wrap_tool_call(self, request, handler): if request.tool_call["name"] == "dynamic_tool": # Execute the dynamic tool yourself or override with tool instance return handler(request.override(tool=my_dynamic_tool)) return handler(request) )grokzgpt-5zgpt-4.1zgpt-4ozgpt-osszo3-prozo3-minicBt|trt|gdS|S)z3Normalize middleware return value to ModelResponse.Nresultstructured_response) isinstancer r )r@s b/home/obispo/Crisostomo_bridge/mision_env/lib/python3.12/site-packages/langchain/agents/factory.py_normalize_to_model_responserDis &)$VH$GG Mc|syt|dk(r|d dfd }|S d d}|dt|ddD] }||  dfd }|S) aCompose multiple `wrap_model_call` handlers into single middleware stack. Composes handlers so first in list becomes outermost layer. Each handler receives a handler callback to execute inner layers. Args: handlers: List of handlers. First handler wraps all others. Returns: Composed handler, or `None` if handlers empty. Example: ```python # handlers=[auth, retry] means: auth wraps retry # Flow: auth calls retry, retry calls base handler def auth(req, state, runtime, handler): try: return handler(req) except UnauthorizedError: refresh_token() return handler(req) def retry(req, state, runtime, handler): for attempt in range(3): try: return handler(req) except Exception: if attempt == 2: raise handler = _chain_model_call_handlers([auth, retry]) ``` Nrc,||}t|SNrDrequesthandlerr@single_handlers rCnormalized_singlez5_chain_model_call_handlers..normalized_singles$GW5F/7 7rEc$ dfd }|S)z-Compose two handlers where outer wraps inner.c<dfd }||}t|S)Nc,|}t|SrIrJreq inner_resultrMinners rC inner_handlerzX_chain_model_call_handlers..compose_two..composed..inner_handlers$S'2 3LAArErTrreturnr rJrLrMrW outer_resultrVouters ` rCcomposedzA_chain_model_call_handlers..compose_two..composeds"  B !-8L/ = =rErLrrMz'Callable[[ModelRequest], ModelResponse]rYr r\rVr]s`` rC compose_twoz/_chain_model_call_handlers..compose_twos) >! >< > >rEc,||}t|SrIrJrLrM final_resultr@s rCfinal_normalizedz4_chain_model_call_handlers..final_normalizeds gw/ +L99rEr^)r\\Callable[[ModelRequest, Callable[[ModelRequest], ModelResponse]], ModelResponse | AIMessage]rVrgrYzPCallable[[ModelRequest, Callable[[ModelRequest], ModelResponse]], ModelResponse]lenreversedhandlersrOrarMrfr@rNs @@rC_chain_model_call_handlersrmpsf  8}!! 8! 8< 8 8!     >b\FHSbM*.Wf-.::8: : rEc|syt|dk(r|d dfd }|S d d}|dt|ddD] }||  dfd }|S) aCompose multiple async `wrap_model_call` handlers into single middleware stack. Args: handlers: List of async handlers. First handler wraps all others. Returns: Composed async handler, or `None` if handlers empty. NrGrcHK||d{}t|S7wrIrJrKs rCrOz;_chain_async_model_call_handlers..normalized_singles(*'7;;F/7 7< " "c$ dfd }|S)z3Compose two async handlers where outer wraps inner.cXKdfd }||d{}t|S7w)NcHK|d{}t|S7wrIrJrSs rCrWz^_chain_async_model_call_handlers..compose_two..composed..inner_handlers&%*3%88 3LAA 9rprXrJrZs ` rCr]zG_chain_async_model_call_handlers..compose_two..composeds.  B "'w !>>L/ = =?s *(*rLrrMz2Callable[[ModelRequest], Awaitable[ModelResponse]]rYr r_r`s`` rCraz5_chain_async_model_call_handlers..compose_two s) >! >G > >rErbcHK||d{}t|S7wrIrJrds rCrfz:_chain_async_model_call_handlers..final_normalized.s( $GW55 +L996rprt)r\rCallable[[ModelRequest, Callable[[ModelRequest], Awaitable[ModelResponse]]], Awaitable[ModelResponse | AIMessage]]rVrvrYzfCallable[[ModelRequest, Callable[[ModelRequest], Awaitable[ModelResponse]]], Awaitable[ModelResponse]]rhrks @@rC _chain_async_model_call_handlersrws0  8}!! 8! 8G 8 8!     >b\FHSbM*.Wf-.::C: : rEci}|D]h}t|d}|jD]F\}}d}|r5t|} | D]%} t| tst | |dus#d}n|rB|||<Hjt ||S)aiResolve schema by merging schemas and optionally respecting `OmitFromSchema` annotations. Args: schemas: List of schema types to merge schema_name: Name for the generated `TypedDict` omit_flag: If specified, omit fields with this flag set (`'input'` or `'output'`) Returns: Merged schema as `TypedDict` T)include_extrasF)r items_extract_metadatarBr!getattrr) schemas schema_name omit_flagall_annotationsschemahints field_name field_type should_omitmetadatametas rC_resolve_schemar9sO9vd;&+kkm 9 "J K,Z8$D!$7GD)z._supports_provider_strategy..s Z4DJ$$& & ZsF) rBstrr r|profilegetrany&FALLBACK_MODELS_WITH_STRUCTURED_OUTPUT)rtools model_profilers @rC_supports_provider_strategyrs"J% E= ) E< . .ugt, .uj"-    %!!"56z*c:x:K[K[K]?]  Z3Y ZZrEct|tsy|j}|dury|dur!dtj t fSt|t rd|fSt|t r>t|tr-t|r!dtj t fSyt|tr6tfd|Dr!dtj t fSyd|fS)zWHandle structured output error. Returns `(should_retry, retry_tool_message)`. )FrFT)errorc36K|]}t|ywrI)rB)rexc_type exceptions rCrz2_handle_structured_output_error..sM8z)X.Ms) rBr. handle_errors STRUCTURED_OUTPUT_ERROR_TEMPLATEformatrtype issubclass Exceptiontupler)rresponse_formatrs` rC_handle_structured_output_errorrs o| 4#11M5<<3y><RRR-%]""-& mY /Jy-4X9@@s9~@VV V-' M}M M9@@s9~@VV V y) ))rEc||syt|dk(r|dSdd}|d}t|ddD] }|||} |S)aCompose wrappers into middleware stack (first = outermost). Args: wrappers: Wrappers in middleware order. Returns: Composed wrapper, or `None` if empty. Example: ```python wrapper = _chain_tool_call_wrappers([auth, cache, retry]) # Request flows: auth -> cache -> retry -> tool # Response flows: tool -> retry -> cache -> auth ``` NrGrc$ dfd }|S)z-Compose two wrappers where outer wraps inner.c&dfd }||S)Nc|SrIr_rTexecuterVs rC call_innerzT_chain_tool_call_wrappers..compose_two..composed..call_innersS'**rErTr;rYToolMessage | Command[Any]r_rLrrrVr\s ` rCr]z@_chain_tool_call_wrappers..compose_two..composeds  +*- -rE)rLr;rz7Callable[[ToolCallRequest], ToolMessage | Command[Any]]rYrr_r`s`` rCraz._chain_tool_call_wrappers..compose_twos) .$ .L .( .rErb)r\r<rVr<rYr<rhwrappersrar@wrappers rC_chain_tool_call_wrappersrs]$  8}{"b\FHSbM*.Wf-. MrEc|syt|dk(r|dS dd}|d}t|ddD] }|||} |S)zCompose async wrappers into middleware stack (first = outermost). Args: wrappers: Async wrappers in middleware order. Returns: Composed async wrapper, or `None` if empty. NrGrc$ dfd }|S)z3Compose two async wrappers where outer wraps inner.cBKdfd }||d{S7w)Nc2K|d{S7wrIr_rs rCrzZ_chain_async_tool_call_wrappers..compose_two..composed..call_inner3s"30000s rr_rs ` rCr]zF_chain_async_tool_call_wrappers..compose_two..composed.s!  1w 33 33s )rLr;rzBCallable[[ToolCallRequest], Awaitable[ToolMessage | Command[Any]]]rYrr_r`s`` rCraz4_chain_async_tool_call_wrappers..compose_twos) 4$ 4W 4( 4rErb)r\Callable[[ToolCallRequest, Callable[[ToolCallRequest], Awaitable[ToolMessage | Command[Any]]]], Awaitable[ToolMessage | Command[Any]]]rVrrYrrhrs rC_chain_async_tool_call_wrappersrs{,  8}{    :b\FHSbM*.Wf-. MrEr_F) system_promptrr state_schemacontext_schema checkpointerstoreinterrupt_beforeinterrupt_afterdebugnamecachec  ABCDEFGHIJKLMNttr tdJ|t|tr|Jn t|J|g}|dHn8t|tt fr|Hnt|t r|Hn t |HdLtHt rt HjLntHtrHLiILr?LjD]0}tj|}|I|jj<2|Dcgc]}t|dgD]}|}}}|Dcgc]P}|jjt jus&|jj"t j"ur|R}}dN|r$|Dcgc]}|j}}t%|N|Dcgc]P}|jj"t j"us&|jjt jur|R}}dF|r$|Dcgc]}|j"}}t'|FdK|Dcgc]}t|t(s|}}|Dcgc]}t|t(r|}}||z}|sNsFrt+|NFndKKr't-Kj.j1|zGn t-|Gt3|Dchc]}|jc}t3|k7r d}t5||Dcgc]P}|jj6t j6us&|jj8t j8ur|R}}|Dcgc]P}|jj:t j:us&|jj<t j<ur|R}}|Dcgc]P}|jj>t j>us&|jj@t j@ur|R}}|Dcgc]P}|jjBt jBus&|jjDt jDur|R}}|Dcgc]P}|jjFt jFus&|jjHt jHur|R}}|Dcgc]P}|jjHt jHus&|jjFt jFur|R} }dM|r$|Dcgc]}|jF}!}tK|!MdE| r$| Dcgc]}|jH}"}tM|"E|Dchc]}|jN}#}||ntP}$|#jS|$tU|#dd}%tU|#dd }&tU|#d d }'tW|%|&|'| }( d,Ifd D d-FHIKLNfd Cd.CD fd Bd/BGHJMfd })d.CD fd Ad/AEGHJfd }*|(jYdt[|)|*dK|(jYdK|D]}|jj6t j6us&|jj8t j8ur|jj6t j6ur |j6nd}+|jj8t j8ur |j8nd},t[|+|,d}-|(jY|jd|-|%|jj:t j:us&|jj<t j<ur|jj:t j:ur |j:nd}.|jj<t j<ur |j<nd}/t[|.|/d}0|(jY|jd|0|%|jj>t j>us&|jj@t j@ur|jj>t j>ur |j>nd}1|jj@t j@ur |j@nd}2t[|1|2d}3|(jY|jd|3|%|jjBt jBus(|jjDt jDus|jjBt jBur |jBnd}4|jjDt jDur |jDnd}5t[|4|5d}6|(jY|jd|6|%|r|djd}7n|r|djd}7nd}7|r|djd}8nd}8|r|djd}9nd}9|r|djd}:nt\}:|(j_t`|7K|8g};tcdKj.j1DsIr|;je|:|(jgdt[tiK|8I|:d|;d|:g}<|s|9dk7r|to|(|=jd|>jd|8|:tq|=d#"<to|(|djd|8|8|:tq|dd#"|rtsjt|D]:\}=}>to|(|=jd|>jd|8|:tq|=d$"<to|(|djdd|8|:tq|dd$"|r|(j_d|djdtwt3|d%z ddD]D}?||?}=||?d%z }>to|(|=jd|>jd|8|:tq|=d!"F|rtwt3|d%z ddD]D}?||?}=||?d%z }>to|(|=jd|>jd|8|:tq|=d&"Fto|(|djdt\|8|:tq|dd&"d'd(i}@ rd) i@d*<|(jy||| | | | +j{@Scc}}wcc}wcc}wcc}wcc}wcc}wcc}wcc}wcc}wcc}wcc}wcc}wcc}wcc}wcc}wcc}wcc}w)0a7Creates an agent graph that calls tools in a loop until a stopping condition is met. For more details on using `create_agent`, visit the [Agents](https://docs.langchain.com/oss/python/langchain/agents) docs. Args: model: The language model for the agent. Can be a string identifier (e.g., `"openai:gpt-4"`) or a direct chat model instance (e.g., [`ChatOpenAI`][langchain_openai.ChatOpenAI] or other another [LangChain chat model](https://docs.langchain.com/oss/python/integrations/chat)). For a full list of supported model strings, see [`init_chat_model`][langchain.chat_models.init_chat_model(model_provider)]. !!! tip "" See the [Models](https://docs.langchain.com/oss/python/langchain/models) docs for more information. tools: A list of tools, `dict`, or `Callable`. If `None` or an empty list, the agent will consist of a model node without a tool calling loop. !!! tip "" See the [Tools](https://docs.langchain.com/oss/python/langchain/tools) docs for more information. system_prompt: An optional system prompt for the LLM. Can be a `str` (which will be converted to a `SystemMessage`) or a `SystemMessage` instance directly. The system message is added to the beginning of the message list when calling the model. middleware: A sequence of middleware instances to apply to the agent. Middleware can intercept and modify agent behavior at various stages. !!! tip "" See the [Middleware](https://docs.langchain.com/oss/python/langchain/middleware) docs for more information. response_format: An optional configuration for structured responses. Can be a `ToolStrategy`, `ProviderStrategy`, or a Pydantic model class. If provided, the agent will handle structured output during the conversation flow. Raw schemas will be wrapped in an appropriate strategy based on model capabilities. !!! tip "" See the [Structured output](https://docs.langchain.com/oss/python/langchain/structured-output) docs for more information. state_schema: An optional `TypedDict` schema that extends `AgentState`. When provided, this schema is used instead of `AgentState` as the base schema for merging with middleware state schemas. This allows users to add custom state fields without needing to create custom middleware. Generally, it's recommended to use `state_schema` extensions via middleware to keep relevant extensions scoped to corresponding hooks / tools. context_schema: An optional schema for runtime context. checkpointer: An optional checkpoint saver object. Used for persisting the state of the graph (e.g., as chat memory) for a single thread (e.g., a single conversation). store: An optional store object. Used for persisting data across multiple threads (e.g., multiple conversations / users). interrupt_before: An optional list of node names to interrupt before. Useful if you want to add a user confirmation or other interrupt before taking an action. interrupt_after: An optional list of node names to interrupt after. Useful if you want to return directly or run additional processing on an output. debug: Whether to enable verbose logging for graph execution. When enabled, prints detailed information about each node execution, state updates, and transitions during agent runtime. Useful for debugging middleware behavior and understanding agent execution flow. name: An optional name for the `CompiledStateGraph`. This name will be automatically used when adding the agent graph to another graph as a subgraph node - particularly useful for building multi-agent systems. cache: An optional `BaseCache` instance to enable caching of graph execution. Returns: A compiled `StateGraph` that can be used for chat interactions. Raises: AssertionError: If duplicate middleware instances are provided. The agent node calls the language model with the messages list (after applying the system prompt). If the resulting [`AIMessage`][langchain.messages.AIMessage] contains `tool_calls`, the graph will then call the tools. The tools node executes the tools and adds the responses to the messages list as [`ToolMessage`][langchain.messages.ToolMessage] objects. The agent node then calls the language model again. The process repeats until no more `tool_calls` are present in the response. The agent then returns the full list of messages. Example: ```python from langchain.agents import create_agent def check_weather(location: str) -> str: '''Return the weather forecast for the specified location.''' return f"It's always sunny in {location}" graph = create_agent( model="anthropic:claude-sonnet-4-5-20250929", tools=[check_weather], system_prompt="You are a helpful assistant", ) inputs = {"messages": [{"role": "user", "content": "what is the weather in sf"}]} for chunk in graph.stream(inputs, stream_mode="updates"): print(chunk) ``` N)contentrr)rwrap_tool_callawrap_tool_callz-Please remove duplicate middleware instances. StateSchema InputSchemainput OutputSchemaoutput)r input_schema output_schemarc t|trH|js7tj|j } |j |}|g|dSd|giSt|trt|tr|jr|jDcgc] }|dvs |}}|rd} t|dkDrY|Dcgc]}|d } }t| |} t| |\} } | s| |Dcgc]}t!| |d|d } }d|g| iS|d } |d}|j |d }|j"xsd |}|t!||d|d g|dSd|giS#t$r6}t|j jdd}t|||}||d}~wwxYwcc}wcc}wcc}w#t$rF}t|d||} t| |\} } | s| |d|t!| |d|d gicYd}~Sd}~wwxYw) aHandle model output including structured responses. Args: output: The AI message output from the model. effective_response_format: The actual strategy used (may differ from initial if auto-detected). )messagesrA__name__rNrrrGid)r tool_call_idrrargszReturning structured response: )rBr) tool_callsr*from_schema_spec schema_specparserr|rr-r.r rir'rrtool_message_content)reffective_response_formatprovider_strategy_bindingrAexcr~validation_errortcstructured_tool_callsr tool_names should_retry error_message tool_messages tool_callstructured_tool_bindingrstructured_output_toolss rC_handle_model_outputz*create_agent.._handle_model_outputs /1A B$$,C,T,T-99-) ^*C*I*I&*Q'*0I\]]) ) 0, ?69-!!$..%"V*@W2W% !%%:> ,-17L!M"V*!MJ!M >z6 RI2Q!#<3/L-('#8 %  $$1)+D!#F%M%'(@-(@AA2!4 %.EiPVFW.X+*A*G*G RXHY*Z'2FFS<=P._get_bound_models ,4N7N#% &/&=&=&B&B&D #"!# ]] 6a&a*qvv=T/T&--aff5  6"'-.E.J.J.L'M$188'9)=9!o% g--| <*7== N,6M6T6T6VWd W W   / 0 /1A B.>>@F( (((,06:A:P:P*   / >0<< *77"997yACC %S/) *$;%@S@SK( ((-8._execute_model_syncus-=W,E))##  ! !..::Hx( FK.f6OP&z2 ,001FG 3  rEc t |dd||} |}n  |}d|ji}|j|j|d<|S)zASync model request handler with sequential middleware processing.rNrrr"rrrstateruntimerArr@rA) r+r,rLresponse state_updatesr( default_toolsrrr"wrap_model_call_handlers rC model_nodez create_agent..model_nodes~)3:&   # **73H/w8KLH$X__5  ' ' 33;3O3OM/ 0rEcK|\}}|j}|jr|jg|}|j|d{} r |_ ||}|d}|j d}t ||S79w)zExecute model asynchronously and return response. This is the core async model execution logic wrapped by `wrap_model_call` handlers. Raises any exceptions that occur during model invocation. NrrAr?)rr"ainvokerrr r$s rC_execute_model_asyncz*create_agent.._execute_model_asyncs-=W,E))##  ! !..::H~~h// FK.f6OP&z2 ,001FG 3  0sAB B :Bc Kt  |dd||}|d{}n|d{}d|ji}|j|j|d<|S7A71w)zBAsync model request handler with sequential middleware processing.rNr*rAr-) r+r,rLr.r/r5awrap_model_call_handlerr0rrr"s rC amodel_nodez!create_agent..amodel_nodes)3:&   $ +1'::H6g?STTH$X__5  ' ' 33;3O3OM/ 0;Us$A-A)A-A+0A-+A-rFtracez .before_agent)rz .before_modelz .after_modelz .after_agentrrbc34K|]}|jywrI) return_direct)rr s rCrzcreate_agent..`sPt""Ps)rmodel_destinationrend_destination)r=rr>r=r> after_model)rdefault_destinationr=r> can_jump_to before_agent before_modelrG after_agentrecursion_limiti' lc_agent_namer)rrrrrrr)rr rzResponseFormat[Any] | NonerYdict[str, Any])rLrrYz5tuple[Runnable[Any, Any], ResponseFormat[Any] | None])rLrrYr )r+zAgentState[Any]r,zRuntime[ContextT]rYrH)>rBrr/rr.r)r&rrr(rr rr|rrrrrrrrrrr riAssertionErrorrC abefore_agentrD abefore_modelr@ aafter_modelrE aafter_agentwrap_model_callawrap_model_callrmrwrraddrradd_noderradd_edgerrradd_conditional_edges_make_tools_to_model_edge_make_model_to_tools_edge_make_model_to_model_edge_add_middleware_edger itertoolspairwiserangecompile with_config)Orrrrrrrrrrrrrrresponse_schemastructured_tool_infomrmiddleware_toolsmiddleware_w_wrap_tool_callrmiddleware_w_awrap_tool_callasync_wrappersbuilt_in_tools regular_toolsavailable_toolsrmiddleware_w_before_agentmiddleware_w_before_modelmiddleware_w_after_modelmiddleware_w_after_agentmiddleware_w_wrap_model_callmiddleware_w_awrap_model_call sync_handlersasync_handlers state_schemas base_stateresolved_state_schemarrgraphr2r8sync_before_agentasync_before_agentbefore_agent_node sync_before async_before before_node sync_after async_after after_nodesync_after_agentasync_after_agentafter_agent_node entry_nodeloop_entry_nodeloop_exit_node exit_nodetools_to_model_destinationsmodel_to_tools_destinationsm1m2idxconfigr5r(r rr7rr0rrr"rrr1rsO` ` @@@@@@@@@@@@@@rC create_agentrCsf%&,0N m] 3*N*=AN }"& Ol4D%E F"1 O\ 2"1#/o"F9=)<8".6M6T6T"U +\ :"9AC6CC [O#4#E#Eo#V FZ #$8$=$=$B$B C [$.Oawq'27NO!OOO# ;; % %_-K-K K ;; & &o.M.M M ##"".IJA$$JJ!:8!D$ ;; & &o.M.M M ;; % %_-K-K K $ $##5QR!++RR"A."Q"&I!&>A*Q*=a>N> %A1Z4-@QAMA'6O 48O !13 Y44;;=>O ^,  J 'qAFF '(C O;=S!!! ;; # #?+G+G G ;; $ $O,I,I I !!! ;; # #?+G+G G ;; $ $O,I,I I !!  ;; " "/*E*E E ;; # #?+G+G G     ;; " "/*E*E E ;; # #?+G+G G   $ ;; & &o.M.M M ;; ' '/O/O O $ $% ;; ' '/O/O O ;; & &o.M.M M %!%##4PQq**Q Q"<]"K $$6ST!,,TT#CN#S 8BC1CMC!-!9zJj!+M=$O"=-IL#M>8LM *!#%   f&f&6Pf& f&PzBzB >zBzBx 66 88 NN7,ZERS w *Q KK $ $O,H,H H{{((0M0MM ;;++?3O3OO ;;,,O4Q4QQ  !11BDV^c d  NN66(-(*;J_   KK $ $O,H,H H{{((0M0MM ;;++?3O3OO ;;,,O4Q4QQ  +; ERK NN66(-(+DY   KK # #?+F+F F{{''/K/KK ;;**/2M2MM  ;;++?3O3OO  **kOJ NNaffX\2JMbN c KK # #?+F+F F{{''/K/KK ;;**/2M2MM  ;;++?3O3OO  00@BS[`a  NN66(,')9H]  _Qh!1!499:-H "1!499:-H  !6q9>>?}M! 4Q7<<=\J  /3889F   NN5*%(7&7# Py/F/F/M/M/OP P& ' . .y 9 ##  )'&5,C$-    ( ((/ &:# n7 ' . . ? ##  )&5,C$-    (  $ % ) ##  )&5$-  i (  7 " ~y1  ,Q/445\B )-%()A!)DmT  !(()BC FB y .')wwi}$="1 ),R@     -b1667}E /-%()B2)FW  !(()BC FB y .')wwi}$="1 ),R@     -b1667}E '-%()B2)FW    w#;B#?#D#D"E\ RS56:ArB C)#.B)#'2B y -')wwi|$<"1 ),R?    56:ArB C)#.B)#'2B y -')wwi|$<"1 ),R?     ,Q/445\B #-%()A!)DmT  08F -t4z ==!)'   k&qP #K $S ?A6 (! !    $%R U Dszw95Aw?x8Ax x=xxx3xxAx"#Ax'>Ax,Ax14Ax6Ax;.yy:y c,|dk(r|S|dk(r|S|dk(ryy)Nrendrr_)jump_tor=r>s rC _resolve_jumprs. '  %' rEctt|dz ddD](}t||ts|}t d||}n|dzdDcgc]}t|t s|}}|fScc}w)NrGrbr )rZrirBr rr)ri last_ai_indexlast_ai_messager_rs rC _fetch_last_ai_and_tool_messagesrs 3x=1$b" - hqk9 -M"; r=rs rCmodel_to_toolsz1_make_model_to_tools_edge..model_to_tools+sii * *7 * "3 /  *J%PZJ[)\&4ABqANNBB )) *a /" "%// w..1V9D[3[    "4 '7"+#   !E )" "! CC  sB>5CCr+rHrYzstr | list[Send] | Noner_)r=rr>rs``` rCrUrU%s! -!-! -!^ rEc  dfd }|S)NcV|jdx}rt|Sd|vrSS)Nrr?rA)rr)r+rr>r=s rCmodel_to_modelz1_make_model_to_model_edge..model_to_modelbsGii * *7 * "3 /  !E )" "! rErr_)r=r>rs`` rCrVrV]s !! !& rEc dfd }|S)Nct|d\}}|jDcgc]}|djvs|}}|rtfd|DrSt fd|DrSScc}w)Nrrc3VK|] }j|dj"yw)rN)rr<)rrrs rCrzD_make_tools_to_model_edge..tools_to_model..s+* ABI # #AfI . < <* s&)c3:K|]}|jvywrI)r)rrrs rCrzD_make_tools_to_model_edge..tools_to_model..sHQqvv00Hs)rrrallr) r+rrrclient_side_tool_callsr>r=rrs rCtools_to_modelz1_make_tools_to_model_edge..tools_to_models)I%PZJ[)\& '11" QvY)BYBY5YA" "  "c* F\* ' # " H-H H" " ! " s A.A.)r+rHrY str | Noner_)rr=rr>rs```` rCrTrTxs!!, rEc|rndfd }g}d|vr|jd|vr|jdd|vr|k7r|j|j|t|d|y|j|y) aAdd an edge to the graph for a middleware node. Args: graph: The graph to add the edge to. name: The name of the middleware node. default_destination: The default destination for the edge. model_destination: The destination for the edge to the model. end_destination: The destination for the edge to the end. can_jump_to: The conditionally jumpable destinations for the edge. cFt|jdxsS)Nrr?)rr)r+rAr>r=s rC jump_edgez'_add_middleware_edge..jump_edges,IIi(&7$3' '  rErrrFr9N)r+rHrYr)rrSrrR)rrrrAr=r>rBr destinationss ``` rCrWrWs* ,, K     0 k !    ( k !d.?&?    1 2 ##D*:9E*RT`a t01rEr)r@zModelResponse | AIMessagerYr )rlzfSequence[Callable[[ModelRequest, Callable[[ModelRequest], ModelResponse]], ModelResponse | AIMessage]]rYzWCallable[[ModelRequest, Callable[[ModelRequest], ModelResponse]], ModelResponse] | None)rlz|Sequence[Callable[[ModelRequest, Callable[[ModelRequest], Awaitable[ModelResponse]]], Awaitable[ModelResponse | AIMessage]]]rYzmCallable[[ModelRequest, Callable[[ModelRequest], Awaitable[ModelResponse]]], Awaitable[ModelResponse]] | NonerI)r}z set[type]r~rrrrYr)rrrYz list[Any])rzAgentMiddleware[Any, Any]rrrYz list[JumpTo])rstr | BaseChatModelrz&list[BaseTool | dict[str, Any]] | NonerYbool)rrrzResponseFormat[Any]rYztuple[bool, str])rzSequence[ToolCallWrapper]rYzToolCallWrapper | None)rzSequence[Callable[[ToolCallRequest, Callable[[ToolCallRequest], Awaitable[ToolMessage | Command[Any]]]], Awaitable[ToolMessage | Command[Any]]]]rYzCallable[[ToolCallRequest, Callable[[ToolCallRequest], Awaitable[ToolMessage | Command[Any]]]], Awaitable[ToolMessage | Command[Any]]] | None)rrrz?Sequence[BaseTool | Callable[..., Any] | dict[str, Any]] | Nonerzstr | SystemMessage | Nonerz.Sequence[AgentMiddleware[StateT_co, ContextT]]rzCResponseFormat[ResponseT] | type[ResponseT] | dict[str, Any] | Nonerz"type[AgentState[ResponseT]] | Nonerztype[ContextT] | NonerzCheckpointer | NonerzBaseStore | Nonerlist[str] | NonerrrrrrrzBaseCache[Any] | NonerYzcCompiledStateGraph[AgentState[ResponseT], ContextT, _InputAgentState, _OutputAgentState[ResponseT]])rz JumpTo | Noner=rr>rrYr)rzlist[AnyMessage]rYz#tuple[AIMessage, list[ToolMessage]])r=rr!dict[str, OutputToolBinding[Any]]r>rrY3Callable[[dict[str, Any]], str | list[Send] | None])r=rr>rrYr) rrr=rrrr>rrYz&Callable[[dict[str, Any]], str | None])rrz[StateGraph[AgentState[ResponseT], ContextT, _InputAgentState, _OutputAgentState[ResponseT]]rrrArr=rr>rrBzlist[JumpTo] | NonerYNone)f__doc__ __future__rrXtypingrrrrrr r *langchain_core.language_models.chat_modelsr langchain_core.messagesr r rrlangchain_core.toolsrlanggraph._internal._runnablerlanggraph.constantsrrlanggraph.graph.staterlanggraph.prebuilt.tool_noderrlanggraph.typesrrtyping_extensionsrrr!langchain.agents.middleware.typesrrrrr r!r"r#r$r%"langchain.agents.structured_outputr&r'r(r)r*r+r,r-r.langchain.chat_modelsr/collections.abcr0r1r2langchain_core.runnablesr3r4langgraph.cache.baser5r6langgraph.runtimer7langgraph.store.baser8r9langgraph.typingr:r;r<rstripr rrDrmrwrr{rrrrrrrrrUrVrTrW__all__r_rErCrsa@"EUU):*,F)>>      2==A.8).,)R#O 0 EG14 *&oo odTT Tn3D  HQU% %'M% %P**(**<-'--`== =DNR@15AC[_7;,0(,")-(,#'@ @ J@. @ ? @ Y @5@*@&@ @'@&@ @ @ !@ !@F        * *( * 55?5 5 9 5p9 6?    , @-2 -2  -2  -2-2-2%-2 -2b rE