yj{QhdZddlZddlZddlmZmZddlmZmZddl m Z ddl m Z ddlmZmZmZddlmZdd lmZdd lmZdd lmZmZmZdd lmZdd lmZddl m!Z!m"Z"m#Z#ddl$m%Z%m&Z&m'Z'ddl(m)Z)m*Z*ddl+m,Z,m-Z-m.Z.m/Z/ddl0m1Z1ddl2m3Z3ej4e5Z6dedefdZ7GddZ8dS)z Core Dialectic Agent implementation. This agent uses tools to gather context from the memory system and synthesize responses to queries about a peer. N) AsyncIteratorCallable)Anycast)generate)crud)ConfiguredModelSettingsReasoningLevelsettings) tracked_db)prompts)embedding_client)HonchoLLMCallResponseStreamingResponseWithMetadatahoncho_llm_call)LLMTelemetryContext)prometheus_metrics)DialecticCompletedEventEmbeddingCallPurposeemit)accumulate_metriclog_performance_metricslog_token_usage_metrics)DialecticComponents TokenTypes)DIALECTIC_TOOLSDIALECTIC_TOOLS_MINIMALcreate_tool_executor search_memory)format_new_turn_with_timestamp)embedding_call_purposereasoning_levelreturnc:tjj|jS)N)r DIALECTICLEVELS MODEL_CONFIG)r"s :/DATA/AppData/hermes/projects/honcho/src/dialectic/core.py!_get_dialectic_level_model_configr).s   $_ 5 BBcpeZdZdZ d#dededzdededeedzd eedzd edzd efd Zd$dZded edzfdZ ded e e ee ee fge feedzeffdZd efdZ d%dededzdedededededzdedzdededzdeded dfd Zded efd!Zded eefd"ZdS)&DialecticAgenta An agentic dialectic that iteratively gathers context to answer queries. Unlike the standard dialectic which pre-gathers all context before a single LLM call, this agent uses tools to strategically gather only the context needed to answer the specific query. Nlowworkspace_name session_nameobserverobservedobserver_peer_cardobserved_peer_card metric_keyr"c ||_||_||_||_||_||_||_||_dtj ||||dg|_ d|_ d|_ t|_dS)a[ Initialize the dialectic agent. Args: workspace_name: Workspace identifier session_name: Session identifier (may be None for global queries) observer: The peer making the query observed: The peer being queried about observer_peer_card: Biographical information about the observer observed_peer_card: Biographical information about the observed peer metric_key: Optional key for logging metrics (if provided, agent won't log separately) reasoning_level: Level of reasoning to apply systemrolecontentFrN)r.r/r0r1r2r3r4r"r agent_system_promptmessages_session_history_initialized_prefetched_conclusion_countgenerate_nanoid_run_id) selfr.r/r0r1r2r3r4r"s r(__init__zDialecticAgent.__init__=s0$2(4% % 4F4F&0/> !"6h(: z r9)r<r r%SESSION_HISTORY_MAX_TOKENSr/r get_messagesr.r executescalarsallr r9 created_at peer_nameappendchrjoinr;) r@ max_tokensstmtdbresultr;formatted_messagesmsg formattedsession_history_sections r(_initialize_session_historyz*DialecticAgent._initialize_session_historyks  ,  F,0)'B ??$"3? F&.*"           9:: 5 5 5 5 5 5 5b::d++++++++F~~''++--H  5 5 5 5 5 5 5 5 5 5 5 5 5 5-/  5 5:K #)))4444  5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5  !2ww||.//  ! ! !  a###'>>#####s0ADu(E(E"E"E"E"E"E"E  F F F F F F F F F F F F F F F#0#2$"|)###M"/#2$BBB)"""L%%'' L,A,A,C,C t 1>0A0A0C0ClFVFVFXFX0XD - "E ))++ R ]==%=PPQQQ((** P \<<<NNOOO99U## #    NNBqBB C C C44444 sB,F3A& F3&A**F3-A*.BF34B>F33 G%=G  G%cK|d{Vd}|jr|j}nt}d|}tj}t |dd|jd|jpdd|jd|j d |j d t |d |d | |d{V}|rd |d |}t |d|d nd |}|j d|dt|j|j|j|j tjj|jddd{V}||||fS)au Prepare common state for answering a query. Handles session history initialization, metrics setup, observation prefetching, user message construction, and tool executor creation. Args: query: The question to answer about the peer Returns: A tuple of (tool_executor, task_name, run_id, start_time) Ndialectic_chat_contextz workspace: z session: z(global)z observer: z observed: z reasoning_level: blobrYzQuery: z ## Relevant Observations (prefetched) The following observations were found to be semantically relevant to your query. Use these as primary context. You may still use tools to find additional information if needed. prefetched_observationsuserr7r])r.r/r0r1history_token_limitr^ agent_typer_)rXr4r>time perf_counterrr.r/r0r1r"rxr;rMrr r%HISTORY_TOKEN_LIMITr?)r@rYr^ task_name start_timer} user_content tool_executors r(_prepare_queryzDialecticAgent._prepare_querys..000000000! ? 3II$&&F2&22I&((   ;d1;; -;;;!];;"];;%)$8 ;;   )WeV<<<(,(L(LU(S(S"S"S"S"S"S"S " --%--+ --  46Mv    -U,,L fFFGGG'.*]] ( 2 F<"'     i;;r*cJt|jddd|j|jS)aJBuild the LLMTelemetryContext shared by answer() and answer_stream(). Carries the instance's `_run_id` (always set in __init__) + workspace + peer identifiers so LLMCallCompletedEvent and 's AgentIterationEvent can attribute every per-iteration LLM call back to this dialectic invocation. zdialectic.answerr])r. call_purposer_rr^rL)rr.r?r1)r@s r(_telemetry_contextz!DialecticAgent._telemetry_context0s4#.+'"<m     r*Frr^rresponse_content input_tokens output_tokenscache_read_input_tokenscache_creation_input_tokenstool_calls_countthinking_content iterationshit_input_token_capc t|d| d| rt|d| dt||||pd|pdt|d|dtj|z dz} t|d| d |js|t d |t jjrvtj |tj j tjj |j tj |tjj tjj |j t%t'|j|j|j|j|j| |j| | |||pd|pd| d S)a Log metrics common to both streaming and non-streaming responses. Args: task_name: Metrics task identifier run_id: Run identifier (None if using caller-provided metric_key) start_time: Start time from time.perf_counter() response_content: The full response text input_tokens: Input token count (actual from API) output_tokens: Output token count (actual from API) cache_read_input_tokens: Cache read tokens (if any) cache_creation_input_tokens: Cache creation tokens (if any) tool_calls_count: Number of tool calls made thinking_content: Thinking trace content (if any) iterations: Number of iterations in the tool execution loop tool_callscountthinkingr|rresponseitotal_durationmsNdialectic_chat)r token_type componentr")r^r.rLr/r"total_iterationsprefetched_conclusion_countrtotal_duration_msrrcache_read_tokenscache_creation_tokensr)rrrrr4rr METRICSENABLEDrrecord_dialectic_tokensrINPUTrjrTOTALr"OUTPUTrrr?r.r1r/r=)r@rr^rrrrrrrrrr elapsed_mss r(_log_response_metricsz$DialecticAgent._log_response_metricsAs> )\3CWMMM  O i5Ev N N N    # (q ' ,1     )Z1A6JJJ'))J6$> )%5z4HHH >6#5 #$4f = = =   #   6"%+1-39 $ 4       6#%,2-39 $ 4      #|#2-!. $ 4!+,0,M!1",)+"9">Q&A&FQ$7        r*cxK||d{V\}}}}tjj|j}|jdkrt nt }|j|jntjj}tt|jd|||j ||j |j dtjj d| d{V} ||||| j| j| j| j| jt+| j| j| j| j | jS)aw Answer a query about the peer using agentic tool calling. The agent will: 1. Receive the query 2. Use tools to gather relevant context 3. Synthesize a response grounded in the gathered context Args: query: The question to answer about the peer Returns: The synthesized answer string Nr[zDialectic Agentr) model_configpromptrPtools tool_choicermax_tool_iterationsr; track_namemax_input_tokens trace_name telemetry rr^rrrrrrrrrr)rr r%r&r"rrMAX_OUTPUT_TOKENSrr) TOOL_CHOICEMAX_TOOL_ITERATIONSr;MAX_INPUT_TOKENSrrr9rrrrrmtool_calls_maderrr) r@rYrrr^rlevel_settingsrrPrs r(answerzDialecticAgent.answers>B=P=PQV=W=W7W7W7W7W7W7W4 y&*"+243GH #y00 $ #  /;  , ,#5  6E:4;OPP!&2' . B](%/@'--// 6 6 6 0 0 0 0 0 0  ""!%-!."0$,$D(0(L !9::%6* ( < # r*c$K||d{V\}}}}tjj|j}|jdkrt nt }|j|jntjj}tttt|jd|dd||j ||j |jdtjjd|d{V} g} | 23d{V} | jr$| | j| jWV36||||d| | j| j| j| jt3| j| j| j| j dS) a Answer a query about the peer using agentic tool calling, streaming the response. The agent will: 1. Receive the query 2. Use tools to gather relevant context (non-streaming) 3. Stream the synthesized response Args: query: The question to answer about the peer Yields: Chunks of the response text as they are generated Nr[rTzDialectic Agent Streamr)rrrPstreamstream_final_onlyrrrrr;rrrrr)rr r%r&r"rrrrrrr)rrr;rrr9rMrrOrrrrrmrrrr) r@rYrrr^rrrrPraccumulated_contentchunks r( answer_streamzDialecticAgent.answer_streams>B=P=PQV=W=W7W7W7W7W7W7W4 y&*"+243GH #y00 $ #  /;  , ,#5   )!>t?STT%"&*6+$2$F3!)!3!D+1133        (*,# $ $ $ $ $ $ $%} $#**5=999m####$ ""!WW%899!."0$,$D(0(L !9::%6* ( < # s3D&)NNNr-)r#N)F)__name__ __module__ __qualname____doc__strlistr rArXrxtuplerdictrfloatrrrintboolrrrrr*r(r,r,4ss04/3!%*/,.,.,.Dj,. ,.  ,. !I, ,.!I,,.$J,.(,.,.,.,.\*?*?*?*?XO3O3:OOOObF<F< xd38n-s23S#*eK LF<F<F<Fsrcr src.configr r r src.dependenciesr src.dialecticr src.embedding_clientrsrc.llmrrr src.llm.typesr src.telemetryrsrc.telemetry.eventsrrrsrc.telemetry.loggingrrr src.telemetry.prometheus.metricsrrsrc.utils.agent_toolsrrrrsrc.utils.formattingr src.utils.typesr! getLoggerrrpr)r,rr*r(rsi 33333333......HHHHHHHHHH''''''!!!!!!111111 .-----,,,,,,TTTTTTTTTT MLLLLLLL @?????222222  8 $ $C#CCCCC l l l l l l l l l l r*