ARCH: удержать metadata document pivot через year-switch

This commit is contained in:
2026-04-23 14:44:04 +03:00
parent 14465c7fde
commit f7a8844c90
4 changed files with 298 additions and 19 deletions
@@ -475,6 +475,18 @@ function hasMetadataSignal(text) {
function hasMetadataObjectHint(text) {
return /(?:\u043e\u0431\u044a\u0435\u043a\u0442(?:\u044b|\u0430|\u043e\u0432)?|\u0440\u0435\u0433\u0438\u0441\u0442\u0440(?:\u044b)?|\u0434\u043e\u043a\u0443\u043c\u0435\u043d\u0442(?:\u044b)?|\u0441\u043f\u0440\u0430\u0432\u043e\u0447\u043d\u0438\u043a(?:\u0438)?|\u043f\u043e\u043b(?:\u0435|\u044f)|objects?|registers?|documents?|catalogs?|fields?)/iu.test(text);
}
function hasSimpleDocumentLanePivotSignal(text) {
return /(?:^|\s)по\s+документ(?:ам|ы)?(?:\?|$|\s)/iu.test(text);
}
function hasSimpleMovementLanePivotSignal(text) {
return /(?:^|\s)по\s+движени(?:ям|я)?(?:\?|$|\s)/iu.test(text);
}
function hasUtf8DocumentLanePivotSignal(text) {
return /(?:^|\s)по\s+документ(?:ам|ы)?(?:\?|$|\s)/iu.test(text);
}
function hasUtf8MovementLanePivotSignal(text) {
return /(?:^|\s)по\s+движени(?:ям|я)?(?:\?|$|\s)/iu.test(text);
}
function hasDocumentEvidenceFollowupSignal(text) {
return /(?:\u043f\u043e\s+\u0434\u043e\u043a\u0443\u043c\u0435\u043d\u0442(?:\u0430\u043c|\u044b)?|\u0434\u0430\u0432\u0430\u0439\s+\u0434\u043e\u043a\u0443\u043c\u0435\u043d\u0442(?:\u044b)?|\u0438\u0449\u0438\s+\u0434\u043e\u043a\u0443\u043c\u0435\u043d\u0442(?:\u044b)?|\u043f\u043e\u043a\u0430\u0436\u0438\s+\u0434\u043e\u043a\u0443\u043c\u0435\u043d\u0442(?:\u044b)?|(?:\u043f\u043e\u043a\u0430\u0436\u0438|\u043a\u0430\u043a\u0438\u0435|\u0441\u043f\u0438\u0441\u043e\u043a|\u0434\u0430\u0439|\u0438\u0449\u0438)\s+(?:\u0441\u0447(?:[еe]т|\u0435\u0442)[-\u2011 ]?\u0444\u0430\u043a\u0442\u0443\u0440(?:\u044b|\u0430)?|\u043d\u0430\u043a\u043b\u0430\u0434\u043d(?:\u044b\u0435|\u0430\u044f)?|\u0430\u043a\u0442(?:\u044b)?|\u0440\u0435\u0430\u043b\u0438\u0437\u0430\u0446(?:\u0438\u0438|\u0438\u044e)|invoice(?:s)?|bill(?:s)?|waybill(?:s)?)|document(?:s)?\s+(?:then|next)?|(?:then|next)\s+documents?|go\s+to\s+documents?)/iu.test(text);
}
@@ -742,13 +754,23 @@ function buildAssistantMcpDiscoveryTurnInput(input) {
? resolveEntityResolutionAmbiguityChoice(rawEntitySourceText, followupSeed.entityResolutionAmbiguityCandidates)
: null;
const entityResolutionSignal = rawEntityResolutionSignal || Boolean(rawEntityCandidate) || Boolean(entityResolutionClarificationCandidate);
const metadataDocumentHintSignal = hasDocumentEvidenceFollowupSignal(rawText) || hasPronounDocumentEvidenceFollowupSignal(rawText);
const metadataMovementHintSignal = hasMovementEvidenceFollowupSignal(rawText) || hasPronounMovementEvidenceFollowupSignal(rawText);
const rawDomain = toNonEmptyString(assistantTurnMeaning?.asked_domain_family);
const rawAction = toNonEmptyString(assistantTurnMeaning?.asked_action_family);
const rawAggregationAxis = toNonEmptyString(assistantTurnMeaning?.asked_aggregation_axis);
const unsupported = toNonEmptyString(assistantTurnMeaning?.unsupported_but_understood_family);
const explicitIntentCandidate = toNonEmptyString(assistantTurnMeaning?.explicit_intent_candidate);
const currentTurnDocumentLaneSignal = rawAction === "list_documents";
const currentTurnMovementLaneSignal = rawAction === "list_movements";
const metadataDocumentHintSignal = currentTurnDocumentLaneSignal ||
hasUtf8DocumentLanePivotSignal(rawText) ||
hasSimpleDocumentLanePivotSignal(rawText) ||
hasDocumentEvidenceFollowupSignal(rawText) ||
hasPronounDocumentEvidenceFollowupSignal(rawText);
const metadataMovementHintSignal = currentTurnMovementLaneSignal ||
hasUtf8MovementLanePivotSignal(rawText) ||
hasSimpleMovementLanePivotSignal(rawText) ||
hasMovementEvidenceFollowupSignal(rawText) ||
hasPronounMovementEvidenceFollowupSignal(rawText);
const assistantTurnMeaningDateScope = toNonEmptyString(assistantTurnMeaning?.explicit_date_scope);
const assistantTurnMeaningOrganizationScope = toNonEmptyString(assistantTurnMeaning?.explicit_organization_scope);
const rawOrganizationMentionSignal = hasOrganizationScopeSignalUtf8(rawText);
@@ -894,6 +916,20 @@ function buildAssistantMcpDiscoveryTurnInput(input) {
!rawValueFlowSignal &&
!rawMetadataSignal &&
metadataDocumentHintSignal);
const metadataScopedMovementEvidenceToDocumentFollowupApplicable = Boolean(followupSeed.pilotScope === "counterparty_movement_evidence_query_movements_v1" &&
followupSeed.subjectResolutionOptional &&
!followupSeed.counterparty &&
metadataLaneCarryoverAvailable &&
!rawLifecycleSignal &&
!rawValueFlowSignal &&
metadataDocumentHintSignal);
const metadataScopedDocumentEvidenceToMovementFollowupApplicable = Boolean(followupSeed.pilotScope === "counterparty_document_evidence_query_documents_v1" &&
followupSeed.subjectResolutionOptional &&
!followupSeed.counterparty &&
metadataLaneCarryoverAvailable &&
!rawLifecycleSignal &&
!rawValueFlowSignal &&
metadataMovementHintSignal);
const metadataGroundedLaneContinuationApplicable = Boolean(followupSeed.pilotScope === "metadata_inspection_v1" &&
(followupSeed.metadataRouteFamily === "document_evidence" ||
followupSeed.metadataRouteFamily === "movement_evidence") &&
@@ -952,13 +988,15 @@ function buildAssistantMcpDiscoveryTurnInput(input) {
metadataLaneCarryoverAvailable &&
!rawLifecycleSignal &&
!rawValueFlowSignal &&
!rawMetadataSignal &&
followupSeed.domain === "documents" &&
followupSeed.action === "list_documents") ||
!currentTurnMovementLaneSignal &&
(!rawMetadataSignal || currentTurnDocumentLaneSignal) &&
(currentTurnDocumentLaneSignal ||
(followupSeed.domain === "documents" && followupSeed.action === "list_documents"))) ||
entityResolutionGroundedDocumentFollowupApplicable ||
entityResolutionClarifiedDocumentFollowupApplicable ||
valueFlowGroundedDocumentFollowupApplicable ||
movementEvidenceGroundedDocumentFollowupApplicable ||
metadataScopedMovementEvidenceToDocumentFollowupApplicable ||
(metadataGroundedLaneContinuationApplicable && followupSeed.metadataRouteFamily === "document_evidence") ||
metadataAmbiguityCollapsedDocumentLaneContinuationApplicable;
const metadataGroundedMovementLaneApplicable = metadataGroundedMovementFollowupApplicable ||
@@ -968,13 +1006,15 @@ function buildAssistantMcpDiscoveryTurnInput(input) {
metadataLaneCarryoverAvailable &&
!rawLifecycleSignal &&
!rawValueFlowSignal &&
!rawMetadataSignal &&
followupSeed.domain === "movements" &&
followupSeed.action === "list_movements") ||
!currentTurnDocumentLaneSignal &&
(!rawMetadataSignal || currentTurnMovementLaneSignal) &&
(currentTurnMovementLaneSignal ||
(followupSeed.domain === "movements" && followupSeed.action === "list_movements"))) ||
entityResolutionGroundedMovementFollowupApplicable ||
entityResolutionClarifiedMovementFollowupApplicable ||
valueFlowGroundedMovementFollowupApplicable ||
documentEvidenceGroundedMovementFollowupApplicable ||
metadataScopedDocumentEvidenceToMovementFollowupApplicable ||
(metadataGroundedLaneContinuationApplicable && followupSeed.metadataRouteFamily === "movement_evidence") ||
metadataAmbiguityCollapsedMovementLaneContinuationApplicable;
const effectiveMetadataFollowupSeedApplicable = metadataFollowupSeedApplicable &&
@@ -1379,6 +1419,12 @@ function buildAssistantMcpDiscoveryTurnInput(input) {
if (movementEvidenceGroundedDocumentFollowupApplicable) {
pushReason(reasonCodes, "mcp_discovery_movement_evidence_grounded_document_followup");
}
if (metadataScopedMovementEvidenceToDocumentFollowupApplicable) {
pushReason(reasonCodes, "mcp_discovery_metadata_scoped_movement_to_document_followup");
}
if (metadataScopedDocumentEvidenceToMovementFollowupApplicable) {
pushReason(reasonCodes, "mcp_discovery_metadata_scoped_document_to_movement_followup");
}
if (metadataGroundedLaneContinuationApplicable) {
pushReason(reasonCodes, "mcp_discovery_metadata_grounded_lane_continuation");
}
@@ -678,6 +678,22 @@ function hasMetadataObjectHint(text: string): boolean {
);
}
function hasSimpleDocumentLanePivotSignal(text: string): boolean {
return /(?:^|\s)по\s+документ(?:ам|ы)?(?:\?|$|\s)/iu.test(text);
}
function hasSimpleMovementLanePivotSignal(text: string): boolean {
return /(?:^|\s)по\s+движени(?:ям|я)?(?:\?|$|\s)/iu.test(text);
}
function hasUtf8DocumentLanePivotSignal(text: string): boolean {
return /(?:^|\s)по\s+документ(?:ам|ы)?(?:\?|$|\s)/iu.test(text);
}
function hasUtf8MovementLanePivotSignal(text: string): boolean {
return /(?:^|\s)по\s+движени(?:ям|я)?(?:\?|$|\s)/iu.test(text);
}
function hasDocumentEvidenceFollowupSignal(text: string): boolean {
return /(?:\u043f\u043e\s+\u0434\u043e\u043a\u0443\u043c\u0435\u043d\u0442(?:\u0430\u043c|\u044b)?|\u0434\u0430\u0432\u0430\u0439\s+\u0434\u043e\u043a\u0443\u043c\u0435\u043d\u0442(?:\u044b)?|\u0438\u0449\u0438\s+\u0434\u043e\u043a\u0443\u043c\u0435\u043d\u0442(?:\u044b)?|\u043f\u043e\u043a\u0430\u0436\u0438\s+\u0434\u043e\u043a\u0443\u043c\u0435\u043d\u0442(?:\u044b)?|(?:\u043f\u043e\u043a\u0430\u0436\u0438|\u043a\u0430\u043a\u0438\u0435|\u0441\u043f\u0438\u0441\u043e\u043a|\u0434\u0430\u0439|\u0438\u0449\u0438)\s+(?:\u0441\u0447(?:[еe]т|\u0435\u0442)[-\u2011 ]?\u0444\u0430\u043a\u0442\u0443\u0440(?:\u044b|\u0430)?|\u043d\u0430\u043a\u043b\u0430\u0434\u043d(?:\u044b\u0435|\u0430\u044f)?|\u0430\u043a\u0442(?:\u044b)?|\u0440\u0435\u0430\u043b\u0438\u0437\u0430\u0446(?:\u0438\u0438|\u0438\u044e)|invoice(?:s)?|bill(?:s)?|waybill(?:s)?)|document(?:s)?\s+(?:then|next)?|(?:then|next)\s+documents?|go\s+to\s+documents?)/iu.test(
text
@@ -1020,16 +1036,25 @@ export function buildAssistantMcpDiscoveryTurnInput(
: null;
const entityResolutionSignal =
rawEntityResolutionSignal || Boolean(rawEntityCandidate) || Boolean(entityResolutionClarificationCandidate);
const metadataDocumentHintSignal =
hasDocumentEvidenceFollowupSignal(rawText) || hasPronounDocumentEvidenceFollowupSignal(rawText);
const metadataMovementHintSignal =
hasMovementEvidenceFollowupSignal(rawText) || hasPronounMovementEvidenceFollowupSignal(rawText);
const rawDomain = toNonEmptyString(assistantTurnMeaning?.asked_domain_family);
const rawAction = toNonEmptyString(assistantTurnMeaning?.asked_action_family);
const rawAggregationAxis = toNonEmptyString(assistantTurnMeaning?.asked_aggregation_axis);
const unsupported = toNonEmptyString(assistantTurnMeaning?.unsupported_but_understood_family);
const explicitIntentCandidate = toNonEmptyString(assistantTurnMeaning?.explicit_intent_candidate);
const currentTurnDocumentLaneSignal = rawAction === "list_documents";
const currentTurnMovementLaneSignal = rawAction === "list_movements";
const metadataDocumentHintSignal =
currentTurnDocumentLaneSignal ||
hasUtf8DocumentLanePivotSignal(rawText) ||
hasSimpleDocumentLanePivotSignal(rawText) ||
hasDocumentEvidenceFollowupSignal(rawText) ||
hasPronounDocumentEvidenceFollowupSignal(rawText);
const metadataMovementHintSignal =
currentTurnMovementLaneSignal ||
hasUtf8MovementLanePivotSignal(rawText) ||
hasSimpleMovementLanePivotSignal(rawText) ||
hasMovementEvidenceFollowupSignal(rawText) ||
hasPronounMovementEvidenceFollowupSignal(rawText);
const assistantTurnMeaningDateScope = toNonEmptyString(assistantTurnMeaning?.explicit_date_scope);
const assistantTurnMeaningOrganizationScope = toNonEmptyString(assistantTurnMeaning?.explicit_organization_scope);
const rawOrganizationMentionSignal = hasOrganizationScopeSignalUtf8(rawText);
@@ -1221,6 +1246,24 @@ export function buildAssistantMcpDiscoveryTurnInput(
!rawMetadataSignal &&
metadataDocumentHintSignal
);
const metadataScopedMovementEvidenceToDocumentFollowupApplicable = Boolean(
followupSeed.pilotScope === "counterparty_movement_evidence_query_movements_v1" &&
followupSeed.subjectResolutionOptional &&
!followupSeed.counterparty &&
metadataLaneCarryoverAvailable &&
!rawLifecycleSignal &&
!rawValueFlowSignal &&
metadataDocumentHintSignal
);
const metadataScopedDocumentEvidenceToMovementFollowupApplicable = Boolean(
followupSeed.pilotScope === "counterparty_document_evidence_query_documents_v1" &&
followupSeed.subjectResolutionOptional &&
!followupSeed.counterparty &&
metadataLaneCarryoverAvailable &&
!rawLifecycleSignal &&
!rawValueFlowSignal &&
metadataMovementHintSignal
);
const metadataGroundedLaneContinuationApplicable = Boolean(
followupSeed.pilotScope === "metadata_inspection_v1" &&
(followupSeed.metadataRouteFamily === "document_evidence" ||
@@ -1290,13 +1333,15 @@ export function buildAssistantMcpDiscoveryTurnInput(
metadataLaneCarryoverAvailable &&
!rawLifecycleSignal &&
!rawValueFlowSignal &&
!rawMetadataSignal &&
followupSeed.domain === "documents" &&
followupSeed.action === "list_documents") ||
!currentTurnMovementLaneSignal &&
(!rawMetadataSignal || currentTurnDocumentLaneSignal) &&
(currentTurnDocumentLaneSignal ||
(followupSeed.domain === "documents" && followupSeed.action === "list_documents"))) ||
entityResolutionGroundedDocumentFollowupApplicable ||
entityResolutionClarifiedDocumentFollowupApplicable ||
valueFlowGroundedDocumentFollowupApplicable ||
movementEvidenceGroundedDocumentFollowupApplicable ||
metadataScopedMovementEvidenceToDocumentFollowupApplicable ||
(metadataGroundedLaneContinuationApplicable && followupSeed.metadataRouteFamily === "document_evidence") ||
metadataAmbiguityCollapsedDocumentLaneContinuationApplicable;
const metadataGroundedMovementLaneApplicable =
@@ -1307,13 +1352,15 @@ export function buildAssistantMcpDiscoveryTurnInput(
metadataLaneCarryoverAvailable &&
!rawLifecycleSignal &&
!rawValueFlowSignal &&
!rawMetadataSignal &&
followupSeed.domain === "movements" &&
followupSeed.action === "list_movements") ||
!currentTurnDocumentLaneSignal &&
(!rawMetadataSignal || currentTurnMovementLaneSignal) &&
(currentTurnMovementLaneSignal ||
(followupSeed.domain === "movements" && followupSeed.action === "list_movements"))) ||
entityResolutionGroundedMovementFollowupApplicable ||
entityResolutionClarifiedMovementFollowupApplicable ||
valueFlowGroundedMovementFollowupApplicable ||
documentEvidenceGroundedMovementFollowupApplicable ||
metadataScopedDocumentEvidenceToMovementFollowupApplicable ||
(metadataGroundedLaneContinuationApplicable && followupSeed.metadataRouteFamily === "movement_evidence") ||
metadataAmbiguityCollapsedMovementLaneContinuationApplicable;
const effectiveMetadataFollowupSeedApplicable =
@@ -1761,6 +1808,12 @@ export function buildAssistantMcpDiscoveryTurnInput(
if (movementEvidenceGroundedDocumentFollowupApplicable) {
pushReason(reasonCodes, "mcp_discovery_movement_evidence_grounded_document_followup");
}
if (metadataScopedMovementEvidenceToDocumentFollowupApplicable) {
pushReason(reasonCodes, "mcp_discovery_metadata_scoped_movement_to_document_followup");
}
if (metadataScopedDocumentEvidenceToMovementFollowupApplicable) {
pushReason(reasonCodes, "mcp_discovery_metadata_scoped_document_to_movement_followup");
}
if (metadataGroundedLaneContinuationApplicable) {
pushReason(reasonCodes, "mcp_discovery_metadata_grounded_lane_continuation");
}
@@ -2017,4 +2017,90 @@ describe("assistant MCP discovery turn input adapter", () => {
expect(result.turn_meaning_ref?.explicit_entity_candidates).toBeUndefined();
expect(result.data_need_graph?.clarification_gaps).toEqual(["period"]);
});
it("pivots a metadata-scoped subjectless movement retrieval into documents without inventing a counterparty", () => {
const orgName =
"\u041e\u041e\u041e \u0410\u043b\u044c\u0442\u0435\u0440\u043d\u0430\u0442\u0438\u0432\u0430 \u041f\u043b\u044e\u0441";
const result = buildAssistantMcpDiscoveryTurnInput({
userMessage: "\u0430 \u0442\u0435\u043f\u0435\u0440\u044c \u043f\u043e \u0434\u043e\u043a\u0443\u043c\u0435\u043d\u0442\u0430\u043c?",
assistantTurnMeaning: {
asked_domain_family: "counterparty",
asked_action_family: "list_documents",
explicit_intent_candidate: "list_documents_by_counterparty"
},
followupContext: {
previous_discovery_pilot_scope: "counterparty_movement_evidence_query_movements_v1",
previous_discovery_loop_status: "ready_for_next_hop",
previous_discovery_loop_selected_chain_id: "movement_evidence",
previous_discovery_loop_pending_axes: [],
previous_discovery_loop_asked_domain_family: "movements",
previous_discovery_loop_asked_action_family: "list_movements",
previous_discovery_loop_unsupported_family: "movement_evidence",
previous_discovery_loop_metadata_scope_hint: "\u041d\u0414\u0421",
previous_discovery_loop_subject_resolution_optional: true,
previous_discovery_entity_candidates: ["\u041d\u0414\u0421"],
previous_filters: {
organization: orgName,
period_from: "2020-01-01",
period_to: "2020-12-31"
}
}
});
expect(result.adapter_status).toBe("ready");
expect(result.should_run_discovery).toBe(true);
expect(result.semantic_data_need).toBe("document evidence");
expect(result.turn_meaning_ref).toMatchObject({
asked_domain_family: "documents",
asked_action_family: "list_documents",
metadata_scope_hint: "\u041d\u0414\u0421",
subject_resolution_optional: true,
explicit_organization_scope: orgName,
explicit_date_scope: "2020",
unsupported_but_understood_family: "document_evidence",
stale_replay_forbidden: true
});
expect(result.turn_meaning_ref?.explicit_entity_candidates).toBeUndefined();
expect(result.reason_codes).toContain("mcp_discovery_metadata_scoped_movement_to_document_followup");
});
it("keeps the metadata-scoped document lane after a year-switch follow-up", () => {
const orgName =
"\u041e\u041e\u041e \u0410\u043b\u044c\u0442\u0435\u0440\u043d\u0430\u0442\u0438\u0432\u0430 \u041f\u043b\u044e\u0441";
const result = buildAssistantMcpDiscoveryTurnInput({
userMessage: "\u0430 \u0442\u0435\u043f\u0435\u0440\u044c \u0437\u0430 2021?",
followupContext: {
previous_discovery_pilot_scope: "counterparty_document_evidence_query_documents_v1",
previous_discovery_loop_status: "ready_for_next_hop",
previous_discovery_loop_selected_chain_id: "document_evidence",
previous_discovery_loop_pending_axes: [],
previous_discovery_loop_asked_domain_family: "documents",
previous_discovery_loop_asked_action_family: "list_documents",
previous_discovery_loop_unsupported_family: "document_evidence",
previous_discovery_loop_metadata_scope_hint: "\u041d\u0414\u0421",
previous_discovery_loop_subject_resolution_optional: true,
previous_discovery_entity_candidates: ["\u041d\u0414\u0421"],
previous_filters: {
organization: orgName,
period_from: "2020-01-01",
period_to: "2020-12-31"
}
}
});
expect(result.adapter_status).toBe("ready");
expect(result.should_run_discovery).toBe(true);
expect(result.semantic_data_need).toBe("document evidence");
expect(result.turn_meaning_ref).toMatchObject({
asked_domain_family: "documents",
asked_action_family: "list_documents",
metadata_scope_hint: "\u041d\u0414\u0421",
subject_resolution_optional: true,
explicit_organization_scope: orgName,
explicit_date_scope: "2021",
unsupported_but_understood_family: "document_evidence",
stale_replay_forbidden: true
});
expect(result.turn_meaning_ref?.explicit_entity_candidates).toBeUndefined();
});
});