ARCH: связать metadata grounding с document MCP lane

2026-04-21 23:01:12 +03:00
parent 6d9c1568c3
commit 7ef788fa50
18 changed files with 706 additions and 54 deletions
@@ -149,6 +149,7 @@ interface AssistantMcpDiscoveryCoverageAwareQueryExecution {

 export type AssistantMcpDiscoveryPilotScope =
  | "metadata_inspection_v1"
+  | "counterparty_document_evidence_query_documents_v1"
  | "counterparty_lifecycle_query_documents_v1"
  | "counterparty_value_flow_query_movements_v1"
  | "counterparty_supplier_payout_query_movements_v1"
@@ -291,6 +292,18 @@ function isLifecyclePilotEligible(planner: AssistantMcpDiscoveryPlannerContract)
  );
 }

+function isDocumentEvidencePilotEligible(planner: AssistantMcpDiscoveryPlannerContract): boolean {
+  const meaning = planner.discovery_plan.turn_meaning_ref;
+  const domain = String(meaning?.asked_domain_family ?? "").toLowerCase();
+  const action = String(meaning?.asked_action_family ?? "").toLowerCase();
+  const unsupported = String(meaning?.unsupported_but_understood_family ?? "").toLowerCase();
+  const combined = `${domain} ${action} ${unsupported}`;
+  return (
+    planner.proposed_primitives.includes("query_documents") &&
+    (combined.includes("document") || combined.includes("list_documents"))
+  );
+}
+
 function isValueFlowPilotEligible(planner: AssistantMcpDiscoveryPlannerContract): boolean {
  const meaning = planner.discovery_plan.turn_meaning_ref;
  const domain = String(meaning?.asked_domain_family ?? "").toLowerCase();
@@ -641,6 +654,16 @@ function summarizeLifecycleRows(result: AddressMcpQueryExecutorResult): string |
  return `${result.fetched_rows} MCP document rows fetched, ${result.matched_rows} matched lifecycle scope`;
 }

+function summarizeDocumentRows(result: AddressMcpQueryExecutorResult): string | null {
+  if (result.error) {
+    return null;
+  }
+  if (result.fetched_rows <= 0) {
+    return "0 MCP document rows fetched";
+  }
+  return `${result.fetched_rows} MCP document rows fetched, ${result.matched_rows} matched document scope`;
+}
+
 function summarizeValueFlowRows(result: AssistantMcpDiscoveryCoverageAwareQueryResult): string | null {
  if (result.error) {
    return null;
@@ -1310,6 +1333,17 @@ function buildLifecycleConfirmedFacts(result: AddressMcpQueryExecutorResult, cou
  ];
 }

+function buildDocumentConfirmedFacts(result: AddressMcpQueryExecutorResult, counterparty: string | null): string[] {
+  if (result.error || result.matched_rows <= 0) {
+    return [];
+  }
+  return [
+    counterparty
+      ? `1C document rows were found for counterparty ${counterparty}`
+      : "1C document rows were found for the requested scope"
+  ];
+}
+
 function buildValueFlowConfirmedFacts(
  result: AssistantMcpDiscoveryCoverageAwareQueryResult,
  counterparty: string | null,
@@ -1357,6 +1391,13 @@ function buildLifecycleInferredFacts(result: AddressMcpQueryExecutorResult): str
  return ["Business activity duration may be inferred from first and latest confirmed 1C activity rows"];
 }

+function buildDocumentInferredFacts(result: AddressMcpQueryExecutorResult): string[] {
+  if (result.error || result.fetched_rows <= 0) {
+    return [];
+  }
+  return ["Counterparty document evidence is limited to confirmed 1C document rows in the checked scope"];
+}
+
 function buildValueFlowInferredFacts(derived: AssistantMcpDiscoveryDerivedValueFlow | null): string[] {
  if (!derived) {
    return [];
@@ -1400,6 +1441,14 @@ function buildLifecycleUnknownFacts(): string[] {
  return ["Legal registration date is not proven by this MCP discovery pilot"];
 }

+function buildDocumentUnknownFacts(periodScope: string | null): string[] {
+  return [
+    periodScope
+      ? "Full document history outside the checked period is not proven by this MCP discovery pilot"
+      : "Full document history is not proven without an explicit checked period"
+  ];
+}
+
 function buildValueFlowUnknownFacts(
  periodScope: string | null,
  direction: AssistantMcpDiscoveryDerivedValueFlow["value_flow_direction"],
@@ -1465,6 +1514,9 @@ function pilotScopeForPlanner(planner: AssistantMcpDiscoveryPlannerContract): As
  if (isValueFlowPilotEligible(planner)) {
    return valueFlowPilotProfile(planner).scope;
  }
+  if (isDocumentEvidencePilotEligible(planner)) {
+    return "counterparty_document_evidence_query_documents_v1";
+  }
  return "counterparty_lifecycle_query_documents_v1";
 }

@@ -1533,10 +1585,11 @@ export async function executeAssistantMcpDiscoveryPilot(
  }

  const metadataPilotEligible = isMetadataPilotEligible(planner);
+  const documentPilotEligible = isDocumentEvidencePilotEligible(planner);
  const lifecyclePilotEligible = isLifecyclePilotEligible(planner);
  const valueFlowPilotEligible = isValueFlowPilotEligible(planner);

-  if (!metadataPilotEligible && !lifecyclePilotEligible && !valueFlowPilotEligible) {
+  if (!metadataPilotEligible && !documentPilotEligible && !lifecyclePilotEligible && !valueFlowPilotEligible) {
    pushReason(reasonCodes, "pilot_scope_unsupported_for_live_execution");
    for (const step of dryRun.execution_steps) {
      skippedPrimitives.push(step.primitive_id);
@@ -1631,6 +1684,89 @@ export async function executeAssistantMcpDiscoveryPilot(
    };
  }

+  if (documentPilotEligible) {
+    let queryResult: AddressMcpQueryExecutorResult | null = null;
+    const filters = buildLifecycleFilters(planner);
+    const selection = selectAddressRecipe("list_documents_by_counterparty", filters);
+    if (!selection.selected_recipe) {
+      pushReason(reasonCodes, "pilot_document_recipe_not_available");
+      const evidence = buildEmptyEvidence(planner, dryRun, probeResults, "Document-evidence recipe is not available");
+      return {
+        schema_version: ASSISTANT_MCP_DISCOVERY_PILOT_EXECUTOR_SCHEMA_VERSION,
+        policy_owner: "assistantMcpDiscoveryPilotExecutor",
+        pilot_status: "unsupported",
+        pilot_scope: "counterparty_document_evidence_query_documents_v1",
+        dry_run: dryRun,
+        mcp_execution_performed: false,
+        executed_primitives: executedPrimitives,
+        skipped_primitives: skippedPrimitives,
+        probe_results: probeResults,
+        evidence,
+        source_rows_summary: null,
+        derived_metadata_surface: null,
+        derived_activity_period: null,
+        derived_value_flow: null,
+        derived_bidirectional_value_flow: null,
+        query_limitations: ["Document-evidence recipe is not available"],
+        reason_codes: reasonCodes
+      };
+    }
+
+    const recipePlan = buildAddressRecipePlan(selection.selected_recipe, filters);
+    for (const step of dryRun.execution_steps) {
+      if (step.primitive_id !== "query_documents") {
+        skippedPrimitives.push(step.primitive_id);
+        probeResults.push(skippedProbeResult(step, "pilot_only_executes_query_documents"));
+        continue;
+      }
+      queryResult = await runtimeDeps.executeAddressMcpQuery({
+        query: recipePlan.query,
+        limit: recipePlan.limit,
+        account_scope: recipePlan.account_scope
+      });
+      executedPrimitives.push(step.primitive_id);
+      probeResults.push(queryResultToProbeResult(step.primitive_id, queryResult));
+      if (queryResult.error) {
+        pushUnique(queryLimitations, queryResult.error);
+        pushReason(reasonCodes, "pilot_query_documents_mcp_error");
+      } else {
+        pushReason(reasonCodes, "pilot_query_documents_mcp_executed");
+      }
+    }
+
+    const sourceRowsSummary = queryResult ? summarizeDocumentRows(queryResult) : null;
+    const evidence = resolveAssistantMcpDiscoveryEvidence({
+      plan: planner.discovery_plan,
+      probeResults,
+      confirmedFacts: queryResult ? buildDocumentConfirmedFacts(queryResult, counterparty) : [],
+      inferredFacts: queryResult ? buildDocumentInferredFacts(queryResult) : [],
+      unknownFacts: buildDocumentUnknownFacts(dateScope),
+      sourceRowsSummary,
+      queryLimitations,
+      recommendedNextProbe: "explain_evidence_basis"
+    });
+
+    return {
+      schema_version: ASSISTANT_MCP_DISCOVERY_PILOT_EXECUTOR_SCHEMA_VERSION,
+      policy_owner: "assistantMcpDiscoveryPilotExecutor",
+      pilot_status: "executed",
+      pilot_scope: "counterparty_document_evidence_query_documents_v1",
+      dry_run: dryRun,
+      mcp_execution_performed: executedPrimitives.length > 0,
+      executed_primitives: executedPrimitives,
+      skipped_primitives: skippedPrimitives,
+      probe_results: probeResults,
+      evidence,
+      source_rows_summary: sourceRowsSummary,
+      derived_metadata_surface: null,
+      derived_activity_period: null,
+      derived_value_flow: null,
+      derived_bidirectional_value_flow: null,
+      query_limitations: queryLimitations,
+      reason_codes: reasonCodes
+    };
+  }
+
  if (valueFlowPilotEligible) {
    let queryResult: AssistantMcpDiscoveryCoverageAwareQueryResult | null = null;
    const filters = buildValueFlowFilters(planner);