SILVER/NODEDC_1C
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Этап 4 / Волна 10: корректировка settlement-кейса — доменная фиксация синтеза, честное покрытие, удержание фокуса / Этап 4 / Волна 11: бизнес-якоря, доменное заземление и устранение утечки дебага

Browse Source
This commit is contained in:
dctouch
2026-03-28 02:17:19 +03:00
parent 914843a8ba
commit a06e575be4
367 changed files with 432257 additions and 3627 deletions
Download Patch File Download Diff File Expand all files Collapse all files
llm_normalizer/backend/dist/services/answerComposer.js
Vendored
+1598 -201
View File
File diff suppressed because it is too large Load Diff
llm_normalizer/backend/dist/services/assistantDataLayer.js
Vendored
+1703 -79
View File
File diff suppressed because it is too large Load Diff
llm_normalizer/backend/dist/services/assistantService.js
Vendored
+267 -33
View File
@@ -1,15 +1,51 @@
"use strict";
var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
if (k2 === undefined) k2 = k;
var desc = Object.getOwnPropertyDescriptor(m, k);
if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
desc = { enumerable: true, get: function() { return m[k]; } };
}
Object.defineProperty(o, k2, desc);
}) : (function(o, m, k, k2) {
if (k2 === undefined) k2 = k;
o[k2] = m[k];
}));
var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
Object.defineProperty(o, "default", { enumerable: true, value: v });
}) : function(o, v) {
o["default"] = v;
});
var __importStar = (this && this.__importStar) || (function () {
var ownKeys = function(o) {
ownKeys = Object.getOwnPropertyNames || function (o) {
var ar = [];
for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
return ar;
};
return ownKeys(o);
};
return function (mod) {
if (mod && mod.__esModule) return mod;
var result = {};
if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
__setModuleDefault(result, mod);
return result;
};
})();
Object.defineProperty(exports, "__esModule", { value: true });
exports.AssistantService = void 0;
const nanoid_1 = require("nanoid");
const stage1Contracts_1 = require("../types/stage1Contracts");
const config_1 = require("../config");
const log_1 = require("../utils/log");
const answerComposer_1 = require("./answerComposer");
const assistantDataLayer_1 = require("./assistantDataLayer");
const assistantSessionLogger_1 = require("./assistantSessionLogger");
const investigationState_1 = require("./investigationState");
const retrievalResultNormalizer_1 = require("./retrievalResultNormalizer");
exports.evaluateCoverageForTests = evaluateCoverageForTests;
exports.extractSubjectTokensForTests = extractSubjectTokensForTests;
// @ts-nocheck
const nanoid_1 = __importStar(require("nanoid"));
const stage1Contracts_1 = __importStar(require("../types/stage1Contracts"));
const config_1 = __importStar(require("../config"));
const log_1 = __importStar(require("../utils/log"));
const answerComposer_1 = __importStar(require("./answerComposer"));
const assistantDataLayer_1 = __importStar(require("./assistantDataLayer"));
const assistantSessionLogger_1 = __importStar(require("./assistantSessionLogger"));
const investigationState_1 = __importStar(require("./investigationState"));
const retrievalResultNormalizer_1 = __importStar(require("./retrievalResultNormalizer"));
function retrievalSummaryForRoute(route) {
if (route === "store_canonical")
return "Canonical accounting data path selected.";
@@ -56,6 +92,25 @@ function extractExecutionState(normalized) {
};
});
}
function escapeRegex(value) {
return value.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
}
function enrichFragmentTextWithHints(fragment, text) {
const baseText = String(text ?? "").trim();
const accountHints = Array.isArray(fragment.account_hints)
? Array.from(new Set(fragment.account_hints
.map((item) => String(item ?? "").trim())
.filter((item) => item.length > 0)))
: [];
if (accountHints.length === 0) {
return baseText;
}
const hasAccountInText = accountHints.some((account) => new RegExp(`\\b${escapeRegex(account)}\\b`, "i").test(baseText));
if (hasAccountInText) {
return baseText;
}
return `${baseText}, по счету ${accountHints.join(", ")}`;
}
function fragmentTextById(normalized) {
const result = new Map();
for (const item of extractFragments(normalized)) {
@@ -70,7 +125,7 @@ function fragmentTextById(normalized) {
const text = (typeof fragment.raw_fragment_text === "string" && fragment.raw_fragment_text.trim()) ||
(typeof fragment.normalized_fragment_text === "string" && fragment.normalized_fragment_text.trim()) ||
"";
result.set(fragmentId, text);
result.set(fragmentId, enrichFragmentTextWithHints(fragment, text));
}
return result;
}
@@ -95,13 +150,18 @@ function extractDiscardedIntentSegments(normalized) {
}
function collectDateSpans(text) {
const spans = [];
const datePattern = /\b20\d{2}[-/.](?:0[1-9]|1[0-2])(?:[-/.](?:0[1-9]|[12]\d|3[01]))?\b/g;
let match = null;
while ((match = datePattern.exec(text)) !== null) {
spans.push({
start: match.index,
end: match.index + match[0].length
});
const datePatterns = [
/\b20\d{2}[-/.](?:0[1-9]|1[0-2])(?:[-/.](?:0[1-9]|[12]\d|3[01]))?\b/g,
/\b(?:0?[1-9]|[12]\d|3[01])[./-](?:0?[1-9]|1[0-2])[./-](?:\d{2}|\d{4})\b/g
];
for (const datePattern of datePatterns) {
let match = null;
while ((match = datePattern.exec(text)) !== null) {
spans.push({
start: match.index,
end: match.index + match[0].length
});
}
}
return spans;
}
@@ -111,17 +171,80 @@ function intersectsAnySpan(start, end, spans) {
function extractAccountTokens(text) {
const lower = String(text ?? "").toLowerCase();
const explicitAccounts = new Set();
const knownAccountPrefixes = new Set([
"01",
"02",
"07",
"08",
"10",
"13",
"19",
"20",
"21",
"23",
"25",
"26",
"28",
"29",
"41",
"43",
"44",
"45",
"50",
"51",
"52",
"55",
"57",
"58",
"60",
"62",
"66",
"67",
"68",
"69",
"70",
"71",
"73",
"76",
"90",
"91",
"94",
"96",
"97"
]);
const contextualPattern = /(?:\bсчет(?:а|у|ом|ов)?\b|\bсч\.?\b|\baccount(?:s)?\b|\bschet(?:a|u|om|ov)?\b)\s*(?:№|#|:)?\s*(\d{2}(?:\.\d{2})?)/giu;
let contextual = null;
while ((contextual = contextualPattern.exec(lower)) !== null) {
if (contextual[1]) {
explicitAccounts.add(contextual[1]);
const token = String(contextual[1]).trim();
const prefix = token.match(/^(\d{2})/)?.[1];
if (prefix && knownAccountPrefixes.has(prefix)) {
explicitAccounts.add(token);
}
}
}
const pairPattern = /\b(\d{2}\.\d{2})\s*\/\s*(\d{2}\.\d{2})\b/g;
let pairMatch = null;
while ((pairMatch = pairPattern.exec(lower)) !== null) {
const left = String(pairMatch[1] ?? "").trim();
const right = String(pairMatch[2] ?? "").trim();
const leftPrefix = left.match(/^(\d{2})/)?.[1];
const rightPrefix = right.match(/^(\d{2})/)?.[1];
if (leftPrefix && knownAccountPrefixes.has(leftPrefix)) {
explicitAccounts.add(left);
}
if (rightPrefix && knownAccountPrefixes.has(rightPrefix)) {
explicitAccounts.add(right);
}
}
if (explicitAccounts.size > 0) {
return Array.from(explicitAccounts);
}
const spans = collectDateSpans(lower);
const hasAccountingLexeme = /(?:\bсчет(?:а|у|ом|ов)?\b|\bсч\.?\b|\baccount(?:s)?\b|\bschet(?:a|u|om|ov)?\b|оплат|расчет|аванс|долг|settlement|payment|счет|СЃС‡\.?)/iu.test(lower);
if (!hasAccountingLexeme) {
return [];
}
const accountList = [];
const genericPattern = /\b\d{2}(?:\.\d{2})?\b/g;
let generic = null;
@@ -132,6 +255,10 @@ function extractAccountTokens(text) {
if (intersectsAnySpan(start, end, spans)) {
continue;
}
const prefix = value.match(/^(\d{2})/)?.[1];
if (!prefix || !knownAccountPrefixes.has(prefix)) {
continue;
}
accountList.push(value);
}
return Array.from(new Set(accountList));
@@ -406,12 +533,50 @@ function evaluateSubjectTokenMatch(token, corpus, executedRoutes) {
}
return { matched: corpus.includes(token), critical: false };
}
function evidenceCountForRequirement(requirementId, result) {
const evidence = Array.isArray(result.evidence) ? result.evidence : [];
if (evidence.length === 0) {
return 0;
}
const tagged = evidence.filter((item) => {
const claimRef = typeof item?.claim_ref === "string" ? item.claim_ref : "";
return claimRef.toLowerCase() === `requirement:${String(requirementId).toLowerCase()}`;
}).length;
if (tagged > 0) {
return tagged;
}
if (Array.isArray(result.requirement_ids) &&
result.requirement_ids.length === 1 &&
result.requirement_ids[0] === requirementId) {
return evidence.length;
}
return 0;
}
function hasSubstantiveCoverageForRequirement(requirementId, result) {
const evidenceCount = evidenceCountForRequirement(requirementId, result);
if (evidenceCount > 0) {
return true;
}
const problemUnitsCount = Array.isArray(result.problem_units) ? result.problem_units.length : 0;
const candidateEvidenceCount = Array.isArray(result.candidate_evidence) ? result.candidate_evidence.length : 0;
if (problemUnitsCount > 0 || candidateEvidenceCount > 0) {
if (Array.isArray(result.requirement_ids) &&
result.requirement_ids.length === 1 &&
result.requirement_ids[0] === requirementId) {
return true;
}
}
return false;
}
function evaluateCoverage(requirements, retrievalResults) {
const statusByRequirement = new Map();
for (const result of retrievalResults) {
for (const requirementId of result.requirement_ids) {
const list = statusByRequirement.get(requirementId) ?? [];
list.push(result.status);
list.push({
status: result.status,
substantive: hasSubstantiveCoverageForRequirement(requirementId, result)
});
statusByRequirement.set(requirementId, list);
}
}
@@ -419,19 +584,27 @@ function evaluateCoverage(requirements, retrievalResults) {
if (requirement.status === "out_of_scope" || requirement.status === "clarification_needed") {
return requirement;
}
const statuses = statusByRequirement.get(requirement.requirement_id) ?? [];
if (statuses.length === 0) {
const states = statusByRequirement.get(requirement.requirement_id) ?? [];
if (states.length === 0) {
return { ...requirement, status: "uncovered" };
}
if (statuses.includes("ok")) {
const hasAnySubstantive = states.some((item) => item.substantive);
if (!hasAnySubstantive) {
return { ...requirement, status: "uncovered" };
}
const hasOk = states.some((item) => item.status === "ok");
const hasPartial = states.some((item) => item.status === "partial");
const hasEmpty = states.some((item) => item.status === "empty");
const hasError = states.some((item) => item.status === "error");
const hasWeakOk = states.some((item) => item.status === "ok" && !item.substantive);
const hasSubstantiveOk = states.some((item) => item.status === "ok" && item.substantive);
const hasSubstantivePartial = states.some((item) => item.status === "partial" && item.substantive);
if (hasSubstantiveOk && !hasSubstantivePartial && !hasWeakOk && !hasEmpty && !hasError) {
return { ...requirement, status: "covered" };
}
if (statuses.includes("partial")) {
if (hasSubstantiveOk || hasSubstantivePartial || hasOk || hasPartial) {
return { ...requirement, status: "partially_covered" };
}
if (statuses.includes("empty") && !statuses.includes("error")) {
return { ...requirement, status: "covered" };
}
return { ...requirement, status: "uncovered" };
});
const requirementsCovered = resolvedRequirements.filter((item) => item.status === "covered").length;
@@ -459,6 +632,12 @@ function evaluateCoverage(requirements, retrievalResults) {
}
};
}
function evaluateCoverageForTests(requirements, retrievalResults) {
return evaluateCoverage(requirements, retrievalResults);
}
function extractSubjectTokensForTests(text) {
return extractSubjectTokens(text);
}
function checkGrounding(userMessage, requirements, coverage, retrievalResults) {
const whyIncludedSummary = summarizeUnique(retrievalResults.flatMap((item) => item.why_included));
const selectionReasonSummary = summarizeUnique(retrievalResults.flatMap((item) => item.selection_reason));
@@ -627,6 +806,11 @@ function buildAnswerStructureV11(input) {
};
}
const FOLLOWUP_ROUTE_HINTS = new Set(["store_canonical", "store_feature_risk", "hybrid_store_plus_live", "live_mcp_drilldown", "batch_refresh_then_store"]);
const FOLLOWUP_ACTIVE_DOMAIN_ROUTE_MAP = {
settlements_60_62: "hybrid_store_plus_live",
vat_document_register_book: "hybrid_store_plus_live",
month_close_costs_20_44: "hybrid_store_plus_live"
};
const FOLLOWUP_BUSINESS_CONTEXT_MAX = 320;
const FOLLOWUP_SUBJECT_MAX = 160;
const FOLLOWUP_QUESTION_APPEND_MAX = 260;
@@ -669,6 +853,23 @@ function extractNormalizedPeriodLiteral(text) {
}
return null;
}
function extractFollowupAccountAnchorsLoose(text) {
const lower = String(text ?? "").toLowerCase();
const spans = collectDateSpans(lower);
const anchors = [];
const followupAccountPattern = /\b(?:01|02|08|19|20|21|23|25|26|28|29|44|51|60|62|68|76|97)(?:\.\d{2})?\b/g;
let match = null;
while ((match = followupAccountPattern.exec(lower)) !== null) {
const value = String(match[0] ?? "").trim();
const start = match.index;
const end = start + value.length;
if (intersectsAnySpan(start, end, spans)) {
continue;
}
anchors.push(value);
}
return Array.from(new Set(anchors));
}
function hasStrongFollowupAnchors(userMessage, state) {
const explicitPeriod = extractNormalizedPeriodLiteral(userMessage);
if (explicitPeriod && state.focus.period && explicitPeriod !== state.focus.period) {
@@ -678,12 +879,13 @@ function hasStrongFollowupAnchors(userMessage, state) {
}
}
const explicitAccounts = extractAccountTokens(userMessage);
if (explicitAccounts.length > 0) {
const followupAccounts = explicitAccounts.length > 0 ? explicitAccounts : extractFollowupAccountAnchorsLoose(userMessage);
if (followupAccounts.length > 0) {
const knownAccounts = new Set(state.focus.primary_accounts.map((item) => item.trim()));
if (knownAccounts.size === 0) {
return true;
}
if (explicitAccounts.some((item) => !knownAccounts.has(item))) {
if (followupAccounts.some((item) => !knownAccounts.has(item))) {
return true;
}
}
@@ -692,7 +894,10 @@ function hasStrongFollowupAnchors(userMessage, state) {
function routeFromInvestigationState(state) {
const rawDomain = compactWhitespace(state.focus.domain ?? "");
if (!rawDomain) {
return null;
const mappedFromFollowup = state.followup_context?.active_domain
? FOLLOWUP_ACTIVE_DOMAIN_ROUTE_MAP[compactWhitespace(state.followup_context.active_domain)] ?? null
: null;
return mappedFromFollowup;
}
if (FOLLOWUP_ROUTE_HINTS.has(rawDomain)) {
return rawDomain;
@@ -701,6 +906,15 @@ function routeFromInvestigationState(state) {
if (FOLLOWUP_ROUTE_HINTS.has(candidate)) {
return candidate;
}
if (Object.prototype.hasOwnProperty.call(FOLLOWUP_ACTIVE_DOMAIN_ROUTE_MAP, candidate)) {
return FOLLOWUP_ACTIVE_DOMAIN_ROUTE_MAP[candidate];
}
}
const mappedFromFollowup = state.followup_context?.active_domain
? FOLLOWUP_ACTIVE_DOMAIN_ROUTE_MAP[compactWhitespace(state.followup_context.active_domain)] ?? null
: null;
if (mappedFromFollowup) {
return mappedFromFollowup;
}
return null;
}
@@ -752,6 +966,7 @@ function buildFollowupStateBinding(input) {
const hasExplicitExpectedRoute = Boolean(input.payloadContext?.expected_route);
const expectedRouteFromState = !context?.expected_route ? routeFromInvestigationState(input.investigationState) : null;
const periodHintFromState = !context?.period_hint ? input.investigationState.focus.period : null;
const followupContext = input.investigationState.followup_context;
if (expectedRouteFromState) {
context.expected_route = expectedRouteFromState;
}
@@ -771,6 +986,24 @@ function buildFollowupStateBinding(input) {
if (input.investigationState.focus.primary_accounts.length > 0) {
businessContextPatch.push(`focus_accounts:${input.investigationState.focus.primary_accounts.join(",")}`);
}
if (followupContext?.active_domain) {
businessContextPatch.push(`focus_domain:${followupContext.active_domain}`);
}
if ((followupContext?.active_requirement_ids?.length ?? 0) > 0) {
businessContextPatch.push(`active_requirements:${followupContext.active_requirement_ids.slice(0, 4).join(",")}`);
}
if ((followupContext?.uncovered_requirement_ids?.length ?? 0) > 0) {
businessContextPatch.push(`uncovered_requirements:${followupContext.uncovered_requirement_ids.slice(0, 4).join(",")}`);
}
if (followupContext?.last_problem_unit_id) {
businessContextPatch.push(`last_problem_unit:${followupContext.last_problem_unit_id}`);
}
if ((followupContext?.evidence_summary?.length ?? 0) > 0) {
businessContextPatch.push(`evidence_state:${followupContext.evidence_summary.slice(0, 3).join("|")}`);
}
if ((followupContext?.settlement_next_actions?.length ?? 0) > 0) {
businessContextPatch.push("settlement_focus_retained_v1");
}
if (problemContinuityAvailable) {
if (hasExplicitExpectedRoute) {
problemContinuitySkippedReason = "explicit_expected_route";
@@ -800,9 +1033,6 @@ function buildFollowupStateBinding(input) {
if (periodHintFromState && !hasPeriodLiteral(userMessage)) {
appendParts.push(`Период фокуса: ${periodHintFromState}`);
}
if (problemContinuityApplied && (problemState?.focus_problem_types.length ?? 0) > 0) {
appendParts.push(`Problem focus types: ${(problemState?.focus_problem_types ?? []).slice(0, 3).join(", ")}`);
}
const appendBlock = withCappedLength(compactWhitespace(appendParts.join("; ")), FOLLOWUP_QUESTION_APPEND_MAX);
normalizedQuestion = `${userMessage}\n${appendBlock}`.trim();
}
@@ -922,7 +1152,7 @@ class AssistantService {
reason: null
});
try {
const raw = this.dataLayer.executeRoute(planItem.route, planItem.fragment_text);
const raw = await this.dataLayer.executeRouteRuntime(planItem.route, planItem.fragment_text);
retrievalResultsRaw.push({
fragment_id: planItem.fragment_id,
route: planItem.route,
@@ -960,6 +1190,9 @@ class AssistantService {
}
const coverageEvaluation = evaluateCoverage(requirementExtraction.requirements, retrievalResults);
const groundingCheck = checkGrounding(userMessage, coverageEvaluation.requirements, coverageEvaluation.coverage, retrievalResults);
const focusDomainHint = followupBinding.usage?.applied
? session.investigation_state?.followup_context?.active_domain ?? session.investigation_state?.focus.domain ?? null
: null;
const composition = (0, answerComposer_1.composeAssistantAnswer)({
userMessage,
routeSummary: normalized.route_hint_summary,
@@ -967,6 +1200,7 @@ class AssistantService {
requirements: coverageEvaluation.requirements,
coverageReport: coverageEvaluation.coverage,
groundingCheck,
focusDomainHint,
enableAnswerPolicyV11: config_1.FEATURE_ASSISTANT_ANSWER_POLICY_V11,
enableProblemCentricAnswerV1: config_1.FEATURE_ASSISTANT_PROBLEM_CENTRIC_ANSWER_V1,
enableLifecycleAnswerV1: config_1.FEATURE_ASSISTANT_LIFECYCLE_ANSWER_V1
llm_normalizer/backend/dist/services/assistantSessionLogger.js
Vendored
+43 -34
View File
@@ -158,41 +158,50 @@ class AssistantSessionLogger {
this.rootDir = rootDir;
}
persistSession(session) {
(0, files_1.ensureDir)(this.rootDir);
const filePath = path_1.default.resolve(this.rootDir, `${session.session_id}.json`);
const startedAt = session.items[0]?.created_at ?? session.updated_at;
const userMessages = session.items.filter((item) => item.role === "user").length;
const assistantMessages = session.items.filter((item) => item.role === "assistant").length;
const assistantItems = session.items.filter((item) => item.role === "assistant");
const lastAssistant = assistantItems.length > 0 ? assistantItems[assistantItems.length - 1] : null;
const traceIds = unique(session.items.map((item) => item.trace_id));
const replyTypes = Array.from(new Set(session.items
.map((item) => item.reply_type)
.filter((item) => typeof item === "string" && item.length > 0)));
const turns = buildTurns(session.items);
const record = {
schema_version: "assistant_session_log_v1",
session_id: session.session_id,
started_at: startedAt,
updated_at: session.updated_at,
counters: {
total_messages: session.items.length,
user_messages: userMessages,
assistant_messages: assistantMessages
},
trace_ids: traceIds,
reply_types: replyTypes,
investigation_state: session.investigation_state,
turns,
conversation: session.items,
last_assistant: {
message_id: lastAssistant?.message_id ?? null,
reply_type: lastAssistant?.reply_type ?? null,
trace_id: lastAssistant?.trace_id ?? null,
created_at: lastAssistant?.created_at ?? null
try {
(0, files_1.ensureDir)(this.rootDir);
const filePath = path_1.default.resolve(this.rootDir, `${session.session_id}.json`);
const startedAt = session.items[0]?.created_at ?? session.updated_at;
const userMessages = session.items.filter((item) => item.role === "user").length;
const assistantMessages = session.items.filter((item) => item.role === "assistant").length;
const assistantItems = session.items.filter((item) => item.role === "assistant");
const lastAssistant = assistantItems.length > 0 ? assistantItems[assistantItems.length - 1] : null;
const traceIds = unique(session.items.map((item) => item.trace_id));
const replyTypes = Array.from(new Set(session.items
.map((item) => item.reply_type)
.filter((item) => typeof item === "string" && item.length > 0)));
const turns = buildTurns(session.items);
const record = {
schema_version: "assistant_session_log_v1",
session_id: session.session_id,
started_at: startedAt,
updated_at: session.updated_at,
counters: {
total_messages: session.items.length,
user_messages: userMessages,
assistant_messages: assistantMessages
},
trace_ids: traceIds,
reply_types: replyTypes,
investigation_state: session.investigation_state,
turns,
conversation: session.items,
last_assistant: {
message_id: lastAssistant?.message_id ?? null,
reply_type: lastAssistant?.reply_type ?? null,
trace_id: lastAssistant?.trace_id ?? null,
created_at: lastAssistant?.created_at ?? null
}
};
(0, files_1.writeJsonFile)(filePath, record);
}
catch (error) {
const code = error?.code;
if (code === "ENOSPC") {
return;
}
};
(0, files_1.writeJsonFile)(filePath, record);
throw error;
}
}
}
exports.AssistantSessionLogger = AssistantSessionLogger;
llm_normalizer/backend/dist/services/evalService.js
Vendored
+156 -38
View File
@@ -8,6 +8,7 @@ const fs_1 = __importDefault(require("fs"));
const path_1 = __importDefault(require("path"));
const nanoid_1 = require("nanoid");
const config_1 = require("../config");
const p0_eval_runner_1 = require("../eval/p0_eval_runner");
const stage1Contracts_1 = require("../types/stage1Contracts");
const stage2EvalContracts_1 = require("../types/stage2EvalContracts");
const http_1 = require("../utils/http");
@@ -218,6 +219,91 @@ const ASSISTANT_STAGE1_COMPARISON_SCHEMA_VERSION = "assistant_stage1_eval_compar
const DEFAULT_ASSISTANT_STAGE2_SUITE_FILE = "assistant_stage2_canonical_v0_1.json";
const ASSISTANT_STAGE2_RUN_SCHEMA_VERSION = "assistant_stage2_eval_run_v0_1";
const ASSISTANT_STAGE2_COMPARISON_SCHEMA_VERSION = "assistant_stage2_eval_comparison_v0_1";
const INMEM_EVAL_REPORT_PREFIX = "inmem_eval_report:";
const INMEM_EVAL_REPORTS = new Map();
function isNoSpaceError(error) {
const code = error?.code;
return code === "ENOSPC";
}
function tryWriteJsonFile(pathname, value) {
try {
(0, files_1.writeJsonFile)(pathname, value);
return true;
}
catch (error) {
if (isNoSpaceError(error)) {
return false;
}
throw error;
}
}
function tryWriteTextFile(pathname, value) {
try {
fs_1.default.writeFileSync(pathname, value, "utf-8");
return true;
}
catch (error) {
if (isNoSpaceError(error)) {
return false;
}
throw error;
}
}
function putInMemoryEvalReport(report) {
const key = `${INMEM_EVAL_REPORT_PREFIX}${(0, nanoid_1.nanoid)(12)}`;
INMEM_EVAL_REPORTS.set(key, report);
return key;
}
function readEvalReportByRef(ref) {
if (ref.startsWith(INMEM_EVAL_REPORT_PREFIX)) {
const report = INMEM_EVAL_REPORTS.get(ref);
if (!report) {
throw new Error(`In-memory eval report not found: ${ref}`);
}
return {
report,
resolved_path: ref
};
}
const resolvedPath = resolveReadablePath(ref);
const report = JSON.parse(fs_1.default.readFileSync(resolvedPath, "utf-8"));
return {
report,
resolved_path: resolvedPath
};
}
function compactAssistantStage1Report(report) {
const results = Array.isArray(report.results) ? report.results : [];
const compactResults = results.map((item) => ({
case_id: item.case_id ?? null,
scenario_tag: item.scenario_tag ?? null,
accountant_usefulness_score: item.accountant_usefulness_score ?? null,
accountant_metrics: typeof item.accountant_metrics === "object" && item.accountant_metrics !== null ? item.accountant_metrics : null
}));
return {
...report,
results: compactResults
};
}
function compactAssistantStage2Report(report) {
const results = Array.isArray(report.results) ? report.results : [];
const compactResults = results.map((item) => {
const metricSubscores = (item.metric_subscores ?? {});
return {
case_id: item.case_id ?? null,
metric_subscores: {
problem_clarity_score: metricSubscores.problem_clarity_score ?? null,
mechanism_coherence_score: metricSubscores.mechanism_coherence_score ?? null,
problem_first_answer_rate: metricSubscores.problem_first_answer_rate ?? null,
entity_leakage_rate: metricSubscores.entity_leakage_rate ?? null
}
};
});
return {
...report,
results: compactResults
};
}
const KNOWN_PROBLEM_UNIT_TYPES = [
"document_conflict",
"broken_chain_segment",
@@ -900,7 +986,7 @@ class EvalService {
results
};
(0, files_1.ensureDir)(config_1.EVAL_CASES_DIR);
(0, files_1.writeJsonFile)(path_1.default.resolve(config_1.EVAL_CASES_DIR, `${runId}.report.json`), report);
tryWriteJsonFile(path_1.default.resolve(config_1.EVAL_CASES_DIR, `${runId}.report.json`), report);
return report;
}
collectAssistantSignals(finalResponse, turnResponses) {
@@ -1237,8 +1323,9 @@ class EvalService {
};
}
buildAssistantComparisonReport(input) {
const baselinePath = resolveReadablePath(input.baselineReportFile);
const baselineReport = JSON.parse(fs_1.default.readFileSync(baselinePath, "utf-8"));
const baselineRef = readEvalReportByRef(input.baselineReportFile);
const baselinePath = baselineRef.resolved_path;
const baselineReport = baselineRef.report;
const currentReport = input.currentReport;
const metricKeys = [
"retrieval_differentiation_rate",
@@ -1330,19 +1417,21 @@ class EvalService {
(0, files_1.ensureDir)(config_1.REPORTS_DIR);
const jsonPath = path_1.default.resolve(config_1.REPORTS_DIR, `${comparisonId}.json`);
const mdPath = path_1.default.resolve(config_1.REPORTS_DIR, `${comparisonId}.md`);
(0, files_1.writeJsonFile)(jsonPath, comparisonReport);
fs_1.default.writeFileSync(mdPath, buildAssistantComparisonMarkdownReport(comparisonReport), "utf-8");
const jsonWritten = tryWriteJsonFile(jsonPath, comparisonReport);
const mdWritten = tryWriteTextFile(mdPath, buildAssistantComparisonMarkdownReport(comparisonReport));
const comparisonRef = jsonWritten ? jsonPath : putInMemoryEvalReport(comparisonReport);
return {
...comparisonReport,
artifacts: {
comparison_report_json_path: jsonPath,
comparison_report_md_path: mdPath
comparison_report_json_path: comparisonRef,
comparison_report_md_path: mdWritten ? mdPath : null
}
};
}
buildAssistantStage2ComparisonReport(input) {
const baselinePath = resolveReadablePath(input.baselineReportFile);
const baselineReport = JSON.parse(fs_1.default.readFileSync(baselinePath, "utf-8"));
const baselineRef = readEvalReportByRef(input.baselineReportFile);
const baselinePath = baselineRef.resolved_path;
const baselineReport = baselineRef.report;
const currentReport = input.currentReport;
const metricKeys = [
"problem_unit_precision",
@@ -1446,13 +1535,14 @@ class EvalService {
(0, files_1.ensureDir)(config_1.REPORTS_DIR);
const jsonPath = path_1.default.resolve(config_1.REPORTS_DIR, `${comparisonId}.json`);
const mdPath = path_1.default.resolve(config_1.REPORTS_DIR, `${comparisonId}.md`);
(0, files_1.writeJsonFile)(jsonPath, comparisonReport);
fs_1.default.writeFileSync(mdPath, buildAssistantStage2ComparisonMarkdownReport(comparisonReport), "utf-8");
const jsonWritten = tryWriteJsonFile(jsonPath, comparisonReport);
const mdWritten = tryWriteTextFile(mdPath, buildAssistantStage2ComparisonMarkdownReport(comparisonReport));
const comparisonRef = jsonWritten ? jsonPath : putInMemoryEvalReport(comparisonReport);
return {
...comparisonReport,
artifacts: {
comparison_report_json_path: jsonPath,
comparison_report_md_path: mdPath
comparison_report_json_path: comparisonRef,
comparison_report_md_path: mdWritten ? mdPath : null
}
};
}
@@ -1473,7 +1563,7 @@ class EvalService {
const limitations = [];
try {
for (const turn of suiteCase.turns) {
const response = await assistantService.handleMessage({
const response = (await assistantService.handleMessage({
session_id: sessionId,
user_message: turn.user_message,
message: turn.user_message,
@@ -1489,7 +1579,7 @@ class EvalService {
domainPrompt: payload.normalizeConfig.domainPrompt,
fewShotExamples: payload.normalizeConfig.fewShotExamples,
useMock: payload.useMock
});
}));
turnResponses.push(response);
requestsTotal += 1;
}
@@ -1773,11 +1863,13 @@ class EvalService {
(0, files_1.ensureDir)(config_1.REPORTS_DIR);
const runJsonPath = path_1.default.resolve(config_1.REPORTS_DIR, `${runId}.json`);
const runMdPath = path_1.default.resolve(config_1.REPORTS_DIR, `${runId}.md`);
(0, files_1.writeJsonFile)(runJsonPath, report);
fs_1.default.writeFileSync(runMdPath, buildAssistantEvalMarkdownReport(report), "utf-8");
const compactReport = compactAssistantStage1Report(report);
const jsonWritten = tryWriteJsonFile(runJsonPath, compactReport);
const mdWritten = tryWriteTextFile(runMdPath, buildAssistantEvalMarkdownReport(compactReport));
const runReportRef = jsonWritten ? runJsonPath : putInMemoryEvalReport(compactReport);
report.artifacts = {
run_report_json_path: runJsonPath,
run_report_md_path: runMdPath
run_report_json_path: runReportRef,
run_report_md_path: mdWritten ? runMdPath : null
};
if (payload.compareWithReportFile) {
report.comparison = this.buildAssistantComparisonReport({
@@ -1806,7 +1898,7 @@ class EvalService {
const expectedProblemFirst = suiteCase.expected_hints?.expected_problem_first ?? (suiteCase.broadness_level !== "low" || suiteCase.question_type !== "direct");
try {
for (const turn of suiteCase.turns) {
const response = await assistantService.handleMessage({
const response = (await assistantService.handleMessage({
session_id: sessionId,
user_message: turn.user_message,
message: turn.user_message,
@@ -1822,7 +1914,7 @@ class EvalService {
domainPrompt: payload.normalizeConfig.domainPrompt,
fewShotExamples: payload.normalizeConfig.fewShotExamples,
useMock: payload.useMock
});
}));
turnResponses.push(response);
requestsTotal += 1;
}
@@ -2045,11 +2137,13 @@ class EvalService {
(0, files_1.ensureDir)(config_1.REPORTS_DIR);
const runJsonPath = path_1.default.resolve(config_1.REPORTS_DIR, `${runId}.json`);
const runMdPath = path_1.default.resolve(config_1.REPORTS_DIR, `${runId}.md`);
(0, files_1.writeJsonFile)(runJsonPath, report);
fs_1.default.writeFileSync(runMdPath, buildAssistantStage2EvalMarkdownReport(report), "utf-8");
const compactReport = compactAssistantStage2Report(report);
const jsonWritten = tryWriteJsonFile(runJsonPath, compactReport);
const mdWritten = tryWriteTextFile(runMdPath, buildAssistantStage2EvalMarkdownReport(compactReport));
const runReportRef = jsonWritten ? runJsonPath : putInMemoryEvalReport(compactReport);
report.artifacts = {
run_report_json_path: runJsonPath,
run_report_md_path: runMdPath
run_report_json_path: runReportRef,
run_report_md_path: mdWritten ? runMdPath : null
};
if (payload.compareWithReportFile) {
report.comparison = this.buildAssistantStage2ComparisonReport({
@@ -2059,6 +2153,20 @@ class EvalService {
}
return report;
}
async runAssistantP0(payload) {
if (!config_1.FEATURE_ASSISTANT_STAGE2_EVAL_V1) {
throw new http_1.ApiError("ASSISTANT_P0_EVAL_DISABLED", "Assistant P0 eval target is disabled by FEATURE_ASSISTANT_STAGE2_EVAL_V1.", 409);
}
const runner = new p0_eval_runner_1.P0EvalRunner(this.normalizerService);
return runner.run({
normalizeConfig: payload.normalizeConfig,
caseIds: payload.caseIds,
useMock: payload.useMock,
mode: payload.mode,
caseSetFile: payload.caseSetFile,
compareWithReportFile: payload.compareWithReportFile
});
}
async run(payload) {
const mode = payload.mode ?? "standard";
const evalTarget = payload.evalTarget ?? "normalizer";
@@ -2082,6 +2190,16 @@ class EvalService {
compareWithReportFile: payload.compareWithReportFile
});
}
if (evalTarget === "assistant_p0") {
return this.runAssistantP0({
normalizeConfig: payload.normalizeConfig,
caseIds: payload.caseIds,
useMock: payload.useMock,
mode,
caseSetFile: payload.caseSetFile,
compareWithReportFile: payload.compareWithReportFile
});
}
const promptVersion = String(payload.normalizeConfig.promptVersion ?? "").toLowerCase();
const schemaVersion = String(payload.normalizeConfig.schemaVersion ?? "").toLowerCase();
const isV2 = promptVersion.startsWith("normalizer_v2") || schemaVersion === "v2" || schemaVersion === "v2_0_1" || schemaVersion === "v2_0_2";
@@ -2269,17 +2387,17 @@ class EvalService {
results
};
(0, files_1.ensureDir)(config_1.EVAL_CASES_DIR);
(0, files_1.writeJsonFile)(path_1.default.resolve(config_1.EVAL_CASES_DIR, `${runId}.report.json`), report);
tryWriteJsonFile(path_1.default.resolve(config_1.EVAL_CASES_DIR, `${runId}.report.json`), report);
const shouldWriteV11Artifacts = mode === "single-pass-strict" &&
Boolean(payload.caseSetFile) &&
path_1.default.basename(String(payload.caseSetFile)).toLowerCase() === "normalizer_eval_v1_1_30cases.json";
if (shouldWriteV11Artifacts) {
(0, files_1.ensureDir)(config_1.REPORTS_DIR);
(0, files_1.writeJsonFile)(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_eval_v1_1_run.json"), report);
fs_1.default.writeFileSync(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_eval_v1_1_run.md"), buildMarkdownReport({
tryWriteJsonFile(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_eval_v1_1_run.json"), report);
tryWriteTextFile(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_eval_v1_1_run.md"), buildMarkdownReport({
...report,
report_title: "LLM Normalizer v1.1 Eval Run"
}), "utf-8");
}));
}
const shouldWriteV1121EvalArtifacts = mode === "single-pass-strict" &&
String(payload.normalizeConfig.promptVersion ?? "") === "normalizer_v1_1_2_1" &&
@@ -2287,33 +2405,33 @@ class EvalService {
path_1.default.basename(String(payload.caseSetFile)).toLowerCase() === "normalizer_eval_v1_1_2_1_30cases.json";
if (shouldWriteV1121EvalArtifacts) {
(0, files_1.ensureDir)(config_1.REPORTS_DIR);
(0, files_1.writeJsonFile)(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_v1_1_2_1_eval.json"), report);
fs_1.default.writeFileSync(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_v1_1_2_1_eval.md"), buildMarkdownReport({
tryWriteJsonFile(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_v1_1_2_1_eval.json"), report);
tryWriteTextFile(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_v1_1_2_1_eval.md"), buildMarkdownReport({
...report,
report_title: "LLM Normalizer v1.1.2.1 Eval Run"
}), "utf-8");
}));
}
const shouldWriteV111MicroArtifacts = mode === "single-pass-strict" &&
String(payload.normalizeConfig.promptVersion ?? "") === "normalizer_v1_1_1" &&
isSameCaseSet(payload.caseIds, V111_MICRO_CASE_IDS);
if (shouldWriteV111MicroArtifacts) {
(0, files_1.ensureDir)(config_1.REPORTS_DIR);
(0, files_1.writeJsonFile)(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_v1_1_1_micro_eval.json"), report);
fs_1.default.writeFileSync(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_v1_1_1_micro_eval.md"), buildMarkdownReport({
tryWriteJsonFile(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_v1_1_1_micro_eval.json"), report);
tryWriteTextFile(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_v1_1_1_micro_eval.md"), buildMarkdownReport({
...report,
report_title: "LLM Normalizer v1.1.1 Micro Eval"
}), "utf-8");
}));
}
const shouldWriteV112MicroArtifacts = mode === "single-pass-strict" &&
String(payload.normalizeConfig.promptVersion ?? "") === "normalizer_v1_1_2" &&
isSameCaseSet(payload.caseIds, V112_MICRO_CASE_IDS);
if (shouldWriteV112MicroArtifacts) {
(0, files_1.ensureDir)(config_1.REPORTS_DIR);
(0, files_1.writeJsonFile)(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_v1_1_2_micro_eval.json"), report);
fs_1.default.writeFileSync(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_v1_1_2_micro_eval.md"), buildMarkdownReport({
tryWriteJsonFile(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_v1_1_2_micro_eval.json"), report);
tryWriteTextFile(path_1.default.resolve(config_1.REPORTS_DIR, "normalizer_v1_1_2_micro_eval.md"), buildMarkdownReport({
...report,
report_title: "LLM Normalizer v1.1.2 Micro Eval"
}), "utf-8");
}));
}
return report;
}
llm_normalizer/backend/dist/services/investigationState.js
Vendored
+114 -4
View File
@@ -75,6 +75,78 @@ function collectOpenUncertainties(coverageReport, retrievalResults) {
const limitationNotes = retrievalResults.flatMap((result) => result.limitations).slice(0, 6);
return capStrings([...requirementNotes, ...limitationNotes], stage1Contracts_1.INVESTIGATION_MAX_UNCERTAINTIES);
}
function normalizeAccountPrefix(value) {
const account = String(value ?? "").trim();
if (!account) {
return null;
}
const match = account.match(/^(\d{2})/);
return match?.[1] ?? null;
}
function isSettlementAccount(value) {
const prefix = normalizeAccountPrefix(value);
return prefix === "60" || prefix === "62" || prefix === "51" || prefix === "76";
}
function isVatAccount(value) {
const prefix = normalizeAccountPrefix(value);
return prefix === "19" || prefix === "68";
}
function isCloseCostsAccount(value) {
const prefix = normalizeAccountPrefix(value);
if (!prefix) {
return false;
}
const account = Number(prefix);
return (account >= 20 && account <= 44) || prefix === "97";
}
function inferFollowupActiveDomain(input) {
const corpus = `${input.userMessage} ${input.previous.focus.active_query_subject ?? ""}`.toLowerCase();
const hasSettlementSignal = input.focusAccounts.some((item) => isSettlementAccount(item)) ||
/(60(?:\.\d{2})?|62(?:\.\d{2})?|оплат|расчет|расч[её]т|зачет|зач[её]т|аванс|долг|поставщ|покупат|settlement|payment|supplier|customer)/i.test(corpus);
if (hasSettlementSignal) {
return "settlements_60_62";
}
const hasVatSignal = input.focusAccounts.some((item) => isVatAccount(item)) ||
/(ндс|счет[\s-]?фактур|сч[её]т[\s-]?фактур|книг[аи]|vat|invoice|book|register)/i.test(corpus);
if (hasVatSignal) {
return "vat_document_register_book";
}
const hasCloseSignal = input.focusAccounts.some((item) => isCloseCostsAccount(item)) ||
/(закрыти|закрытие|месяц|затрат|распредел|списан|period\s*close|month\s*close|allocation|residual|cost)/i.test(corpus);
if (hasCloseSignal) {
return "month_close_costs_20_44";
}
const routeDomain = deriveDomain(input.routeSummary);
if (routeDomain && routeDomain !== "no_route") {
return routeDomain;
}
return input.previous.followup_context?.active_domain ?? input.previous.focus.domain ?? null;
}
function collectUncoveredRequirementIds(coverageReport) {
return capStrings([
...coverageReport.requirements_uncovered,
...coverageReport.requirements_partially_covered,
...coverageReport.clarification_needed_for,
...coverageReport.out_of_scope_requirements
], stage1Contracts_1.INVESTIGATION_MAX_REQUIREMENT_LINKS);
}
function collectEvidenceSummary(retrievalResults) {
const lines = retrievalResults.map((result) => {
const requirementRef = result.requirement_ids[0] ?? result.fragment_id;
return `${requirementRef}:${result.status}:${result.route}`;
});
return capStrings(lines, 6);
}
function settlementFocusActions(activeDomain) {
if (activeDomain !== "settlements_60_62") {
return [];
}
return [
"Проверьте договор и объект расчетов по платежу.",
"Сверьте регистр расчетов и привязку платежа к закрывающему документу.",
"Проверьте зачет аванса или взаимозачет по связке 60/62."
];
}
function normalizeEntityBacklinks(values) {
const result = [];
const seen = new Set();
@@ -180,7 +252,27 @@ function cloneInvestigationState(state) {
followup_context: state.followup_context
? {
...state.followup_context,
referenced_requirement_ids: [...state.followup_context.referenced_requirement_ids]
referenced_requirement_ids: [...state.followup_context.referenced_requirement_ids],
...(state.followup_context.active_requirement_ids
? {
active_requirement_ids: [...state.followup_context.active_requirement_ids]
}
: {}),
...(state.followup_context.uncovered_requirement_ids
? {
uncovered_requirement_ids: [...state.followup_context.uncovered_requirement_ids]
}
: {}),
...(state.followup_context.settlement_next_actions
? {
settlement_next_actions: [...state.followup_context.settlement_next_actions]
}
: {}),
...(state.followup_context.evidence_summary
? {
evidence_summary: [...state.followup_context.evidence_summary]
}
: {})
}
: null
};
@@ -222,9 +314,21 @@ function createEmptyInvestigationState(sessionId, timestamp = new Date().toISOSt
function updateInvestigationState(input) {
const previous = input.previous;
const focusFromMessage = capStrings(detectAccounts(input.userMessage), stage1Contracts_1.INVESTIGATION_MAX_PRIMARY_ACCOUNTS);
const mergedFocusAccounts = capStrings([...focusFromMessage, ...previous.focus.primary_accounts], stage1Contracts_1.INVESTIGATION_MAX_PRIMARY_ACCOUNTS);
const requirementIds = capStrings(input.requirements.map((item) => item.requirement_id), stage1Contracts_1.INVESTIGATION_MAX_REQUIREMENT_LINKS);
const mainRequirement = input.requirements[0]?.requirement_text ?? input.userMessage;
const problemUnitState = updateProblemUnitState(previous, input.retrievalResults);
const uncoveredRequirementIds = collectUncoveredRequirementIds(input.coverageReport);
const activeDomain = inferFollowupActiveDomain({
userMessage: input.userMessage,
focusAccounts: mergedFocusAccounts,
routeSummary: input.routeSummary,
previous
});
const focusDomain = activeDomain ?? deriveDomain(input.routeSummary) ?? previous.focus.domain;
const settlementNextActions = settlementFocusActions(activeDomain);
const lastProblemUnitId = problemUnitState?.active_problem_units[0] ?? null;
const evidenceSummary = collectEvidenceSummary(input.retrievalResults);
return {
schema_version: stage1Contracts_1.INVESTIGATION_STATE_SCHEMA_VERSION,
session_id: previous.session_id,
@@ -233,9 +337,9 @@ function updateInvestigationState(input) {
updated_at: input.timestamp,
question_id: input.questionId,
focus: {
domain: deriveDomain(input.routeSummary) ?? previous.focus.domain,
domain: focusDomain,
period: detectPeriod(input.userMessage) ?? previous.focus.period,
primary_accounts: capStrings([...focusFromMessage, ...previous.focus.primary_accounts], stage1Contracts_1.INVESTIGATION_MAX_PRIMARY_ACCOUNTS),
primary_accounts: mergedFocusAccounts,
active_query_subject: mainRequirement.slice(0, 180)
},
narrowing_status: deriveNarrowingStatus(input.routeSummary, input.coverageReport),
@@ -245,7 +349,13 @@ function updateInvestigationState(input) {
followup_context: {
previous_question_id: previous.question_id,
last_user_message: input.userMessage.slice(0, 240),
referenced_requirement_ids: requirementIds
referenced_requirement_ids: requirementIds,
active_domain: activeDomain,
active_requirement_ids: requirementIds,
uncovered_requirement_ids: uncoveredRequirementIds,
last_problem_unit_id: lastProblemUnitId,
settlement_next_actions: settlementNextActions,
evidence_summary: evidenceSummary
},
query_mode_hint: deriveQueryModeHint(input.routeSummary),
...(problemUnitState
llm_normalizer/backend/dist/services/retrievalResultNormalizer.js
Vendored
+60 -12
View File
@@ -4,6 +4,7 @@ exports.normalizeRetrievalResult = normalizeRetrievalResult;
const config_1 = require("../config");
const stage1Contracts_1 = require("../types/stage1Contracts");
const problemUnitAssembler_1 = require("./problemUnitAssembler");
const stage4GraphRuntime_1 = require("./stage4GraphRuntime");
function toObject(value) {
if (!value || typeof value !== "object" || Array.isArray(value)) {
return null;
@@ -75,6 +76,21 @@ function mergeSummaryWithProblemUnitMeta(summary, input) {
problem_unit_lifecycle_defect_distribution: input.lifecycleDefectDistribution
};
}
function mergeSummaryWithGraphMeta(summary, graphSummary) {
return {
...summary,
graph_runtime_enabled: true,
graph_total_units: graphSummary.total_units,
graph_bound_units: graphSummary.bound_units,
graph_nodes_count: graphSummary.node_count,
graph_edges_count: graphSummary.edge_count,
graph_missing_links_count: graphSummary.missing_links_count,
graph_conflicting_links_count: graphSummary.conflicting_links_count,
graph_coverage_grade: graphSummary.graph_coverage_grade,
graph_domain_distribution: graphSummary.domain_distribution,
graph_relation_distribution: graphSummary.relation_distribution
};
}
function normalizeConfidence(value) {
if (value === "high" || value === "medium" || value === "low") {
return value;
@@ -408,23 +424,55 @@ function normalizeRetrievalResult(fragmentId, requirementIds, route, raw) {
selection_reason: baseResult.selection_reason,
business_interpretation: baseResult.business_interpretation
});
const enrichedSummary = mergeSummaryWithProblemUnitMeta(summary, {
candidateEvidenceCount: assembled.candidate_evidence.length,
problemUnitsCount: assembled.problem_units.length,
unitTypes: assembled.problem_unit_summary.unit_types,
duplicateCollapses: assembled.problem_unit_summary.duplicate_collapses,
severityDistribution: assembled.problem_unit_summary.severity_distribution,
confidenceDistribution: assembled.problem_unit_summary.confidence_distribution,
lifecycleEnrichedUnits: assembled.problem_unit_summary.lifecycle_enriched_units ?? 0,
lifecycleDomainDistribution: (assembled.problem_unit_summary.lifecycle_domain_distribution ?? {}),
lifecycleDefectDistribution: (assembled.problem_unit_summary.lifecycle_defect_distribution ?? {})
const graphBuild = config_1.FEATURE_ASSISTANT_GRAPH_RUNTIME_V1
? (0, stage4GraphRuntime_1.buildAccountingGraph)({
route,
candidateEvidence: assembled.candidate_evidence,
problemUnits: assembled.problem_units
})
: null;
const graphBindingByUnitId = new Map(graphBuild?.unit_bindings.map((item) => [item.problem_unit_id, item]) ?? []);
const graphBoundProblemUnits = assembled.problem_units.map((unit) => {
const binding = graphBindingByUnitId.get(unit.problem_unit_id);
if (!binding) {
return unit;
}
return {
...unit,
graph_binding: binding
};
});
const graphBoundSummary = graphBuild
? {
...assembled.problem_unit_summary,
graph_summary: graphBuild.summary
}
: assembled.problem_unit_summary;
let enrichedSummary = mergeSummaryWithProblemUnitMeta(summary, {
candidateEvidenceCount: assembled.candidate_evidence.length,
problemUnitsCount: graphBoundProblemUnits.length,
unitTypes: graphBoundSummary.unit_types,
duplicateCollapses: graphBoundSummary.duplicate_collapses,
severityDistribution: graphBoundSummary.severity_distribution,
confidenceDistribution: graphBoundSummary.confidence_distribution,
lifecycleEnrichedUnits: graphBoundSummary.lifecycle_enriched_units ?? 0,
lifecycleDomainDistribution: (graphBoundSummary.lifecycle_domain_distribution ?? {}),
lifecycleDefectDistribution: (graphBoundSummary.lifecycle_defect_distribution ?? {})
});
if (graphBuild) {
enrichedSummary = mergeSummaryWithGraphMeta(enrichedSummary, graphBuild.summary);
}
return {
...baseResult,
summary: enrichedSummary,
raw_entities: items,
candidate_evidence: assembled.candidate_evidence,
problem_units: assembled.problem_units,
problem_unit_summary: assembled.problem_unit_summary
problem_units: graphBoundProblemUnits,
problem_unit_summary: graphBoundSummary,
...(graphBuild
? {
accounting_graph: graphBuild
}
: {})
};
}
llm_normalizer/backend/dist/services/routeHintAdapter.js
Vendored
+206 -91
View File
@@ -1,5 +1,6 @@
"use strict";
Object.defineProperty(exports, "__esModule", { value: true });
exports.ROUTE_DISCIPLINE_RULE_TABLE = void 0;
exports.simulateDeterministicRouting = simulateDeterministicRouting;
exports.toRouteHintSummary = toRouteHintSummary;
exports.toRouterInput = toRouterInput;
@@ -28,6 +29,188 @@ function toRouteHintSummaryV1(normalized) {
}
};
}
const ACCOUNT_HINT_PATTERN = /(?:\b(?:account|acct|schet|счет|сч)\s*[:#]?\s*(?:[1-9][0-9](?:[./-][0-9]{1,2})?)|\b(?:19|20|21|23|25|26|28|29|44|51|60|62|68)\b)/i;
const PERIOD_PATTERN = /\b20\d{2}(?:[-./](?:0[1-9]|1[0-2]))?\b/i;
const SYMPTOM_MARKER_PATTERN = /(?:\bsymptom\b|\banomaly\b|\bproblem\b|\bissue\b|\btail\b|\bhanging\b|\bblocked\b|\bincomplete\b|remains?\s+open|not\s+(?:confirmed|observed|resolved|closed)|не\s+(?:подтвержден|закрыт|наблюдается)|хвост|сбой|проблем)/i;
const LIFECYCLE_MARKER_PATTERN = /(?:\blifecycle\b|\bchain\b|\btransition\b|\bstep\b|\btrace\b|цепоч|этап|переход|связк|где\s+разрыв)/i;
const CHAIN_BREAK_PATTERN = /(?:\bbreak\b|\bbroken\b|\bgap\b|missing\s+(?:transition|step|link)|chain\s+break|разрыв|обрыв|нет\s+переход|не\s+дошл|не\s+наблюд)/i;
const PERIOD_IMPACT_PATTERN = /(?:period\s*close|month\s*close|month-end|residual|allocation|20[/-]44|закрыти|остатк|распредел|конец\s+месяц)/i;
const CAUSAL_PATTERN = /(?:\bwhy\b|\bbecause\b|\breason\b|explain\s+mechanism|почему|объясни|механизм|причин)/i;
const AMBIGUITY_PATTERN = /(?:\bmaybe\b|\bperhaps\b|not\s+sure|i\s+only\s+know|part\s+may\s+be\s+missing|возможно|может\s+быть|не\s+уверен|не\s+знаю|часть\s+цепочки\s+не\s+подтвержд)/i;
const TRANSLIT_PROBLEM_PATTERN = /(?:raschet|oplata|zakryt|nds|vychet|zatrat|ostatok|cepoch|perehod|pochemu|prichin|period)/i;
const DOMAIN_LEXICAL_ANCHOR_PATTERN = /(?:\b(?:settlement|payment|bank|supplier|customer|vat|nds|invoice|register|book|period\s*close|month\s*close|close\s*operation|allocation|residual|cost|expenses?)\b|оплат|расчет|РЅРґСЃ|СЃС‡[её]С‚.?фактур|РєРЅРёРі[аи]|затрат|закрыт|остатк)/i;
exports.ROUTE_DISCIPLINE_RULE_TABLE = [
{
query_class: "exact_object_trace",
required_route: "live_mcp_drilldown",
allowed_fallback: ["no_route"],
forbidden_fallback: ["store_canonical", "hybrid_store_plus_live", "store_feature_risk", "batch_refresh_then_store"],
description: "Exact object trace queries always run via live drilldown."
},
{
query_class: "ranking_or_period_summary",
required_route: "batch_refresh_then_store",
allowed_fallback: ["no_route"],
forbidden_fallback: ["store_canonical", "hybrid_store_plus_live"],
description: "Ranking and period summary queries require analytical batch path."
},
{
query_class: "symptom_first",
required_route: "hybrid_store_plus_live",
allowed_fallback: ["no_route"],
forbidden_fallback: ["store_canonical"],
description: "Symptom-first intents are deterministically promoted to hybrid path."
},
{
query_class: "lifecycle_first",
required_route: "hybrid_store_plus_live",
allowed_fallback: ["no_route"],
forbidden_fallback: ["store_canonical"],
description: "Lifecycle-first intents are deterministically promoted to hybrid path."
},
{
query_class: "chain_break",
required_route: "hybrid_store_plus_live",
allowed_fallback: ["no_route"],
forbidden_fallback: ["store_canonical"],
description: "Chain-break intents are deterministically promoted to hybrid path."
},
{
query_class: "period_impact",
required_route: "hybrid_store_plus_live",
allowed_fallback: ["no_route"],
forbidden_fallback: ["store_canonical"],
description: "Period-impact problem intents are deterministically promoted to hybrid path."
},
{
query_class: "causal_query",
required_route: "hybrid_store_plus_live",
allowed_fallback: ["no_route"],
forbidden_fallback: ["store_canonical"],
description: "Causal/mechanism intents are deterministically promoted to hybrid path."
},
{
query_class: "mixed_ambiguity",
required_route: "hybrid_store_plus_live",
allowed_fallback: ["no_route"],
forbidden_fallback: ["store_canonical"],
description: "Mixed ambiguity keeps hybrid as primary route, with explicit no-route fallback."
},
{
query_class: "rule_check_without_symptom",
required_route: "store_feature_risk",
allowed_fallback: ["no_route"],
forbidden_fallback: ["store_canonical"],
description: "Rule checks without symptom/lifecycle signals run via risk profile path."
},
{
query_class: "canonical_fact_lookup",
required_route: "store_canonical",
allowed_fallback: ["no_route"],
forbidden_fallback: ["hybrid_store_plus_live"],
description: "Only plain factual lookups are allowed to stay on canonical path."
}
];
const ROUTE_DISCIPLINE_RULE_MAP = new Map(exports.ROUTE_DISCIPLINE_RULE_TABLE.map((item) => [item.query_class, item]));
function mergedFragmentText(fragment) {
return `${fragment.raw_fragment_text ?? ""} ${fragment.normalized_fragment_text ?? ""}`.toLowerCase();
}
function hasLifecycleDomainHint(fragment, lowerText) {
const accountHints = Array.isArray(fragment.account_hints) ? fragment.account_hints.map((item) => String(item)) : [];
if (accountHints.some((item) => /^(97|01|02|08|19|20|21|23|25|26|28|29|44|68(?:\.\d+)?|51|60|62)$/.test(item))) {
return true;
}
return (fragment.candidate_labels.includes("anomaly_probe") ||
fragment.candidate_labels.includes("period_close_risk") ||
PERIOD_IMPACT_PATTERN.test(lowerText));
}
function hasSymptomSignal(fragment, lowerText) {
return (fragment.flags.asks_for_anomaly_scan ||
fragment.candidate_labels.includes("anomaly_probe") ||
fragment.candidate_labels.includes("period_close_risk") ||
SYMPTOM_MARKER_PATTERN.test(lowerText) ||
TRANSLIT_PROBLEM_PATTERN.test(lowerText));
}
function hasLifecycleSignal(fragment, lowerText) {
return (fragment.flags.asks_for_chain_explanation ||
fragment.flags.mentions_period_close_context ||
LIFECYCLE_MARKER_PATTERN.test(lowerText) ||
hasLifecycleDomainHint(fragment, lowerText));
}
function hasChainBreakSignal(lowerText) {
return CHAIN_BREAK_PATTERN.test(lowerText);
}
function hasPeriodImpactSignal(lowerText) {
return PERIOD_IMPACT_PATTERN.test(lowerText);
}
function hasCausalSignal(lowerText) {
return CAUSAL_PATTERN.test(lowerText);
}
function hasAmbiguitySignal(fragment, lowerText) {
return (AMBIGUITY_PATTERN.test(lowerText) ||
fragment.confidence === "low" ||
fragment.domain_relevance === "unclear" ||
fragment.business_scope === "unclear");
}
function hasAccountOrPeriodAnchor(fragment, lowerText) {
return fragment.account_hints.length > 0 || ACCOUNT_HINT_PATTERN.test(lowerText) || PERIOD_PATTERN.test(lowerText);
}
function resolveRouteClass(fragment) {
const lowerText = mergedFragmentText(fragment);
const symptomSignal = hasSymptomSignal(fragment, lowerText);
const lifecycleSignal = hasLifecycleSignal(fragment, lowerText);
const chainBreakSignal = hasChainBreakSignal(lowerText);
const periodImpactSignal = hasPeriodImpactSignal(lowerText);
const causalSignal = hasCausalSignal(lowerText);
const ambiguitySignal = hasAmbiguitySignal(fragment, lowerText);
const accountOrPeriodAnchor = hasAccountOrPeriodAnchor(fragment, lowerText);
if (fragment.flags.asks_for_exact_object_trace) {
return ROUTE_DISCIPLINE_RULE_MAP.get("exact_object_trace");
}
if (fragment.flags.asks_for_ranking_or_top || fragment.flags.asks_for_period_summary) {
return ROUTE_DISCIPLINE_RULE_MAP.get("ranking_or_period_summary");
}
if (ambiguitySignal && (symptomSignal || lifecycleSignal || chainBreakSignal || periodImpactSignal || causalSignal)) {
return ROUTE_DISCIPLINE_RULE_MAP.get("mixed_ambiguity");
}
if (chainBreakSignal) {
return ROUTE_DISCIPLINE_RULE_MAP.get("chain_break");
}
if (periodImpactSignal && accountOrPeriodAnchor) {
return ROUTE_DISCIPLINE_RULE_MAP.get("period_impact");
}
if (lifecycleSignal) {
return ROUTE_DISCIPLINE_RULE_MAP.get("lifecycle_first");
}
if (symptomSignal) {
return ROUTE_DISCIPLINE_RULE_MAP.get("symptom_first");
}
if (causalSignal && accountOrPeriodAnchor) {
return ROUTE_DISCIPLINE_RULE_MAP.get("causal_query");
}
if (fragment.flags.asks_for_rule_check) {
return ROUTE_DISCIPLINE_RULE_MAP.get("rule_check_without_symptom");
}
return ROUTE_DISCIPLINE_RULE_MAP.get("canonical_fact_lookup");
}
function shouldPromoteFromNoRoute(fragment, rule) {
if (rule.required_route === "store_canonical") {
return false;
}
if (explicitNoRouteReason(fragment) === "out_of_scope") {
return false;
}
const lowerText = mergedFragmentText(fragment);
const hasProblemSignal = hasSymptomSignal(fragment, lowerText) ||
hasLifecycleSignal(fragment, lowerText) ||
hasChainBreakSignal(lowerText) ||
hasPeriodImpactSignal(lowerText) ||
hasCausalSignal(lowerText);
const hasAnchor = hasAccountOrPeriodAnchor(fragment, lowerText) ||
fragment.candidate_labels.includes("cross_entity") ||
DOMAIN_LEXICAL_ANCHOR_PATTERN.test(lowerText);
return hasProblemSignal && hasAnchor;
}
function reasonForNoRoute(noRouteReason) {
if (noRouteReason === "out_of_scope") {
return "Fragment is out-of-scope for company-specific accounting contour.";
@@ -77,32 +260,30 @@ function decideRouteForFragment(fragment) {
const readiness = executionReadiness(fragment);
const clarification = clarificationReason(fragment);
const soft = softAssumptions(fragment);
if (status === "no_route") {
return buildNoRouteDecision(fragment, noRouteReason);
}
if (readiness === "needs_clarification" || readiness === "no_route") {
return buildNoRouteDecision(fragment, noRouteReason ?? "insufficient_specificity");
}
if (fragment.domain_relevance !== "in_scope") {
const routeRule = resolveRouteClass(fragment);
if (fragment.domain_relevance === "out_of_scope") {
return buildNoRouteDecision(fragment, "out_of_scope");
}
if (fragment.flags.asks_for_exact_object_trace) {
return {
fragment_id: fragment.fragment_id,
domain_relevance: fragment.domain_relevance,
business_scope: fragment.business_scope,
candidate_labels: fragment.candidate_labels,
decision_flags: fragment.flags,
execution_readiness: readiness,
clarification_reason: clarification,
soft_assumption_used: soft,
route_status: "routed",
no_route_reason: null,
route: "live_mcp_drilldown",
reason: "Exact object trace requested."
};
if (status === "no_route" || readiness === "needs_clarification" || readiness === "no_route") {
if (shouldPromoteFromNoRoute(fragment, routeRule)) {
return {
fragment_id: fragment.fragment_id,
domain_relevance: fragment.domain_relevance,
business_scope: fragment.business_scope,
candidate_labels: fragment.candidate_labels,
decision_flags: fragment.flags,
execution_readiness: readiness,
clarification_reason: clarification,
soft_assumption_used: soft,
route_status: "routed",
no_route_reason: null,
route: routeRule.required_route,
reason: `${routeRule.description} Query class: ${routeRule.query_class}. Promoted from no-route by anchor/symptom guardrail.`
};
}
return buildNoRouteDecision(fragment, noRouteReason);
}
if (fragment.flags.asks_for_ranking_or_top || fragment.flags.asks_for_period_summary) {
if (status === "routed" || status === null) {
return {
fragment_id: fragment.fragment_id,
domain_relevance: fragment.domain_relevance,
@@ -114,74 +295,8 @@ function decideRouteForFragment(fragment) {
soft_assumption_used: soft,
route_status: "routed",
no_route_reason: null,
route: "batch_refresh_then_store",
reason: "Ranking/summary semantics require batch analytical route."
};
}
if (fragment.flags.has_multi_entity_scope && fragment.flags.asks_for_chain_explanation) {
return {
fragment_id: fragment.fragment_id,
domain_relevance: fragment.domain_relevance,
business_scope: fragment.business_scope,
candidate_labels: fragment.candidate_labels,
decision_flags: fragment.flags,
execution_readiness: readiness,
clarification_reason: clarification,
soft_assumption_used: soft,
route_status: "routed",
no_route_reason: null,
route: "hybrid_store_plus_live",
reason: "Multi-entity causal chain requested."
};
}
if (fragment.flags.asks_for_rule_check && !fragment.flags.asks_for_chain_explanation) {
return {
fragment_id: fragment.fragment_id,
domain_relevance: fragment.domain_relevance,
business_scope: fragment.business_scope,
candidate_labels: fragment.candidate_labels,
decision_flags: fragment.flags,
execution_readiness: readiness,
clarification_reason: clarification,
soft_assumption_used: soft,
route_status: "routed",
no_route_reason: null,
route: "store_feature_risk",
reason: "Rule-control check without causal decomposition."
};
}
if (fragment.flags.asks_for_anomaly_scan &&
!fragment.flags.asks_for_ranking_or_top &&
!(fragment.flags.has_multi_entity_scope && fragment.flags.asks_for_chain_explanation)) {
return {
fragment_id: fragment.fragment_id,
domain_relevance: fragment.domain_relevance,
business_scope: fragment.business_scope,
candidate_labels: fragment.candidate_labels,
decision_flags: fragment.flags,
execution_readiness: readiness,
clarification_reason: clarification,
soft_assumption_used: soft,
route_status: "routed",
no_route_reason: null,
route: "store_feature_risk",
reason: "Anomaly scan without heavy ranking or causal chain."
};
}
if (status === "routed") {
return {
fragment_id: fragment.fragment_id,
domain_relevance: fragment.domain_relevance,
business_scope: fragment.business_scope,
candidate_labels: fragment.candidate_labels,
decision_flags: fragment.flags,
execution_readiness: readiness,
clarification_reason: clarification,
soft_assumption_used: soft,
route_status: "routed",
no_route_reason: null,
route: "store_canonical",
reason: "Routed fragment without deep analytical or causal signals."
route: routeRule.required_route,
reason: `${routeRule.description} Query class: ${routeRule.query_class}. Allowed fallback: ${routeRule.allowed_fallback.join(", ")}. Forbidden fallback: ${routeRule.forbidden_fallback.join(", ")}.`
};
}
return buildNoRouteDecision(fragment, "missing_mapping");
llm_normalizer/backend/dist/services/stage4GraphRuntime.js
Vendored
+574
View File
@@ -0,0 +1,574 @@
"use strict";
Object.defineProperty(exports, "__esModule", { value: true });
exports.buildAccountingGraph = buildAccountingGraph;
const stage4Graph_1 = require("../types/stage4Graph");
const GRAPH_CONFIDENCE_ORDER = {
low: 1,
medium: 2,
high: 3
};
const DOMAIN_PATH_HINTS = {
bank_settlement: ["payment_to_settlement", "wrong_closing_document_type"],
customer_settlement: ["invoice_to_payment", "payment_to_closure"],
deferred_expense: ["deferred_expense_to_writeoff", "writeoff_sequence"],
fixed_asset: ["asset_card_to_depreciation", "card_document_register_alignment"],
vat_flow: ["invoice_to_vat_register", "cross_branch_alignment"],
period_close: ["period_close_dependency_chain", "closure_blocker_transition"]
};
function uniqueStrings(values, limit = 16) {
return Array.from(new Set(values.map((item) => String(item ?? "").trim()).filter(Boolean))).slice(0, limit);
}
function compactToken(value) {
const normalized = value.toLowerCase().replace(/[^a-z0-9]+/g, "-").replace(/^-+|-+$/g, "");
return normalized.length > 0 ? normalized.slice(0, 48) : "x";
}
function stableNodeId(type, domain, stableKey) {
return `gnd-${compactToken(type)}-${compactToken(domain)}-${compactToken(stableKey)}`;
}
function stableEdgeId(relation, fromNode, toNode) {
return `ged-${compactToken(relation)}-${compactToken(fromNode)}-${compactToken(toNode)}`;
}
function mergeConfidence(left, right) {
return GRAPH_CONFIDENCE_ORDER[right] > GRAPH_CONFIDENCE_ORDER[left] ? right : left;
}
function mergeProvenance(left, right, routeFallback) {
return {
route: String(right?.route ?? left.route ?? routeFallback),
candidate_ids: uniqueStrings([...(left.candidate_ids ?? []), ...(right?.candidate_ids ?? [])], 24),
evidence_ids: uniqueStrings([...(left.evidence_ids ?? []), ...(right?.evidence_ids ?? [])], 24)
};
}
class GraphAccumulator {
route;
nodesById = new Map();
edgesById = new Map();
constructor(route) {
this.route = route;
}
upsertNode(input) {
const node_id = stableNodeId(input.node_type, input.domain, input.stable_key);
const existing = this.nodesById.get(node_id);
if (!existing) {
const created = {
node_id,
node_type: input.node_type,
domain: input.domain,
label: input.label,
confidence: input.confidence,
attributes: input.attributes ?? {},
provenance: {
route: String(input.provenance?.route ?? this.route),
candidate_ids: uniqueStrings(input.provenance?.candidate_ids ?? [], 24),
evidence_ids: uniqueStrings(input.provenance?.evidence_ids ?? [], 24)
}
};
this.nodesById.set(node_id, created);
return created;
}
existing.confidence = mergeConfidence(existing.confidence, input.confidence);
existing.provenance = mergeProvenance(existing.provenance, input.provenance, this.route);
if (Object.keys(input.attributes ?? {}).length > 0) {
existing.attributes = {
...existing.attributes,
...input.attributes
};
}
return existing;
}
upsertEdge(input) {
const edge_id = stableEdgeId(input.relation_type, input.from_node_id, input.to_node_id);
const existing = this.edgesById.get(edge_id);
if (!existing) {
const created = {
edge_id,
relation_type: input.relation_type,
from_node_id: input.from_node_id,
to_node_id: input.to_node_id,
domain: input.domain,
confidence: input.confidence,
flags: uniqueStrings(input.flags ?? [], 8),
provenance: {
route: String(input.provenance?.route ?? this.route),
candidate_ids: uniqueStrings(input.provenance?.candidate_ids ?? [], 24),
evidence_ids: uniqueStrings(input.provenance?.evidence_ids ?? [], 24)
}
};
this.edgesById.set(edge_id, created);
return created;
}
existing.confidence = mergeConfidence(existing.confidence, input.confidence);
existing.flags = uniqueStrings([...(existing.flags ?? []), ...(input.flags ?? [])], 8);
existing.provenance = mergeProvenance(existing.provenance, input.provenance, this.route);
return existing;
}
export() {
return {
nodes: Array.from(this.nodesById.values()).sort((left, right) => left.node_id.localeCompare(right.node_id)),
edges: Array.from(this.edgesById.values()).sort((left, right) => left.edge_id.localeCompare(right.edge_id))
};
}
}
function inferDomainFromUnit(unit) {
if (unit.lifecycle_domain) {
return unit.lifecycle_domain;
}
const accountText = unit.affected_accounts.join(" ").toLowerCase();
if (/\b97\b/.test(accountText) || unit.problem_unit_type === "lifecycle_anomaly_node") {
return "deferred_expense";
}
if (/\b(01|02|08)\b/.test(accountText)) {
return "fixed_asset";
}
if (/\b(19|68)\b/.test(accountText) || unit.problem_unit_type === "cross_branch_inconsistency_cluster") {
return "vat_flow";
}
if (unit.problem_unit_type === "period_risk_cluster" || unit.period_impact?.impact_class === "close_risk") {
return "period_close";
}
if (/\b62\b/.test(accountText)) {
return "customer_settlement";
}
if (/\b(51|60|76)\b/.test(accountText) || unit.problem_unit_type === "unresolved_settlement_cluster") {
return "bank_settlement";
}
return "unknown";
}
function relationPathHints(domain, unit) {
const path = [`domain:${domain}`];
if (unit.current_lifecycle_state && unit.expected_lifecycle_state) {
path.push(`state:${unit.current_lifecycle_state}->${unit.expected_lifecycle_state}`);
}
else if (unit.current_lifecycle_state) {
path.push(`state:${unit.current_lifecycle_state}`);
}
if (domain !== "unknown") {
path.push(...(DOMAIN_PATH_HINTS[domain] ?? []));
}
if (unit.missing_transition) {
path.push(`missing:${unit.missing_transition}`);
}
if (unit.invalid_transition) {
path.push(`conflict:${unit.invalid_transition}`);
}
return uniqueStrings(path, 10);
}
function graphConfidenceFromUnit(unit) {
return unit.lifecycle_confidence?.grade ?? unit.confidence.grade;
}
function coverageGrade(boundUnits, totalUnits) {
if (totalUnits <= 0) {
return "low";
}
const ratio = boundUnits / totalUnits;
if (ratio >= 0.8)
return "high";
if (ratio >= 0.4)
return "medium";
return "low";
}
function buildSummary(input) {
const domain_distribution = {};
const relation_distribution = {};
for (const binding of input.bindings) {
const domainMarker = binding.relation_path.find((item) => item.startsWith("domain:")) ?? "domain:unknown";
const domainKey = domainMarker.replace(/^domain:/, "");
domain_distribution[domainKey] = (domain_distribution[domainKey] ?? 0) + 1;
}
for (const edge of input.edges) {
relation_distribution[edge.relation_type] = (relation_distribution[edge.relation_type] ?? 0) + 1;
}
const missing_links_count = input.bindings.reduce((acc, item) => acc + item.missing_links.length, 0);
const conflicting_links_count = input.bindings.reduce((acc, item) => acc + item.conflicting_links.length, 0);
const bound_units = input.bindings.length;
return {
total_units: input.totalUnits,
bound_units,
node_count: input.nodes.length,
edge_count: input.edges.length,
missing_links_count,
conflicting_links_count,
graph_coverage_grade: coverageGrade(bound_units, input.totalUnits),
domain_distribution,
relation_distribution
};
}
function buildAccountingGraph(input) {
const accumulator = new GraphAccumulator(input.route);
const candidateById = new Map(input.candidateEvidence.map((item) => [item.candidate_id, item]));
const bindings = [];
const issues = [];
if (input.problemUnits.length === 0) {
issues.push("no_problem_units_for_graph_build");
}
for (const unit of input.problemUnits) {
const domain = inferDomainFromUnit(unit);
const confidence = graphConfidenceFromUnit(unit);
const candidateIds = uniqueStrings(unit.evidence_pack, 12).filter((item) => candidateById.has(item));
const evidenceIds = uniqueStrings(unit.evidence_pack, 12);
const domainNode = accumulator.upsertNode({
node_type: "domain",
domain,
stable_key: `domain:${domain}`,
label: domain,
confidence,
attributes: {
domain
},
provenance: {
route: input.route,
candidate_ids: candidateIds,
evidence_ids: evidenceIds
}
});
const problemNode = accumulator.upsertNode({
node_type: "problem_unit",
domain,
stable_key: `problem:${unit.problem_unit_id}`,
label: unit.title || unit.problem_unit_id,
confidence,
attributes: {
problem_unit_id: unit.problem_unit_id,
problem_unit_type: unit.problem_unit_type,
business_defect_class: unit.business_defect_class,
lifecycle_defect_type: unit.lifecycle_defect_type ?? null
},
provenance: {
route: input.route,
candidate_ids: candidateIds,
evidence_ids: evidenceIds
}
});
accumulator.upsertEdge({
relation_type: "belongs_to_domain",
from_node_id: problemNode.node_id,
to_node_id: domainNode.node_id,
domain,
confidence,
flags: ["actual_link"],
provenance: {
route: input.route,
candidate_ids: candidateIds,
evidence_ids: evidenceIds
}
});
for (const account of uniqueStrings(unit.affected_accounts, 4)) {
const accountNode = accumulator.upsertNode({
node_type: "account",
domain,
stable_key: `account:${account}`,
label: account,
confidence,
attributes: {
account
},
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
accumulator.upsertEdge({
relation_type: "affects_account",
from_node_id: problemNode.node_id,
to_node_id: accountNode.node_id,
domain,
confidence,
flags: ["actual_link"],
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
}
for (const document of uniqueStrings(unit.affected_documents, 3)) {
const documentNode = accumulator.upsertNode({
node_type: "document",
domain,
stable_key: `document:${document}`,
label: document,
confidence,
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
accumulator.upsertEdge({
relation_type: "affects_document",
from_node_id: problemNode.node_id,
to_node_id: documentNode.node_id,
domain,
confidence,
flags: ["actual_link"],
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
}
for (const counterparty of uniqueStrings(unit.affected_counterparties, 3)) {
const counterpartyNode = accumulator.upsertNode({
node_type: "counterparty",
domain,
stable_key: `counterparty:${counterparty}`,
label: counterparty,
confidence,
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
accumulator.upsertEdge({
relation_type: "affects_counterparty",
from_node_id: problemNode.node_id,
to_node_id: counterpartyNode.node_id,
domain,
confidence,
flags: ["actual_link"],
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
}
if (unit.current_lifecycle_state) {
const currentNode = accumulator.upsertNode({
node_type: "lifecycle_state",
domain,
stable_key: `current_state:${unit.current_lifecycle_state}`,
label: unit.current_lifecycle_state,
confidence,
attributes: {
state_role: "current"
},
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
accumulator.upsertEdge({
relation_type: "current_state",
from_node_id: problemNode.node_id,
to_node_id: currentNode.node_id,
domain,
confidence,
flags: ["actual_link"],
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
}
if (unit.expected_lifecycle_state) {
const expectedNode = accumulator.upsertNode({
node_type: "lifecycle_state",
domain,
stable_key: `expected_state:${unit.expected_lifecycle_state}`,
label: unit.expected_lifecycle_state,
confidence,
attributes: {
state_role: "expected"
},
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
accumulator.upsertEdge({
relation_type: "expected_state",
from_node_id: problemNode.node_id,
to_node_id: expectedNode.node_id,
domain,
confidence,
flags: ["expected_link"],
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
}
if (unit.missing_transition) {
const missingNode = accumulator.upsertNode({
node_type: "transition",
domain,
stable_key: `missing_transition:${unit.missing_transition}`,
label: unit.missing_transition,
confidence,
attributes: {
transition_role: "missing"
},
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
accumulator.upsertEdge({
relation_type: "missing_transition",
from_node_id: problemNode.node_id,
to_node_id: missingNode.node_id,
domain,
confidence,
flags: ["missing_link"],
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
}
if (unit.invalid_transition) {
const invalidNode = accumulator.upsertNode({
node_type: "transition",
domain,
stable_key: `invalid_transition:${unit.invalid_transition}`,
label: unit.invalid_transition,
confidence,
attributes: {
transition_role: "invalid"
},
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
accumulator.upsertEdge({
relation_type: "invalid_transition",
from_node_id: problemNode.node_id,
to_node_id: invalidNode.node_id,
domain,
confidence,
flags: ["conflict_link"],
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
}
if (unit.lifecycle_defect_type) {
const defectNode = accumulator.upsertNode({
node_type: "defect",
domain,
stable_key: `defect:${unit.lifecycle_defect_type}`,
label: unit.lifecycle_defect_type,
confidence,
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
accumulator.upsertEdge({
relation_type: "has_defect",
from_node_id: problemNode.node_id,
to_node_id: defectNode.node_id,
domain,
confidence,
flags: unit.lifecycle_defect_type === "cross_branch_state_conflict" ? ["conflict_link"] : ["actual_link"],
provenance: {
route: input.route,
evidence_ids: evidenceIds
}
});
}
for (const candidateId of candidateIds.slice(0, 6)) {
const candidate = candidateById.get(candidateId);
const evidenceLabel = candidate
? `${candidate.source_ref.entity}:${candidate.source_ref.id}`
: `candidate:${candidateId}`;
const evidenceNode = accumulator.upsertNode({
node_type: "evidence",
domain,
stable_key: `evidence:${candidateId}`,
label: evidenceLabel,
confidence,
attributes: {
candidate_id: candidateId
},
provenance: {
route: input.route,
candidate_ids: [candidateId],
evidence_ids: [candidateId]
}
});
accumulator.upsertEdge({
relation_type: "supported_by_evidence",
from_node_id: problemNode.node_id,
to_node_id: evidenceNode.node_id,
domain,
confidence,
flags: ["actual_link"],
provenance: {
route: input.route,
candidate_ids: [candidateId],
evidence_ids: [candidateId]
}
});
if (candidate && candidate.relation_pattern_hits.length > 0) {
for (const relationHint of uniqueStrings(candidate.relation_pattern_hits, 2)) {
const hintNode = accumulator.upsertNode({
node_type: "transition",
domain,
stable_key: `hint:${relationHint}`,
label: relationHint,
confidence,
attributes: {
transition_role: "hint"
},
provenance: {
route: input.route,
candidate_ids: [candidateId],
evidence_ids: [candidateId]
}
});
accumulator.upsertEdge({
relation_type: "supports_path",
from_node_id: evidenceNode.node_id,
to_node_id: hintNode.node_id,
domain,
confidence,
flags: ["actual_link"],
provenance: {
route: input.route,
candidate_ids: [candidateId],
evidence_ids: [candidateId]
}
});
}
}
}
const missing_links = uniqueStrings([
unit.missing_transition,
...(unit.lifecycle_resolution?.missing_transitions ?? [])
]);
const conflicting_links = uniqueStrings([
unit.invalid_transition,
...(unit.lifecycle_resolution?.invalid_transitions ?? [])
]);
bindings.push({
problem_unit_id: unit.problem_unit_id,
graph_node_id: problemNode.node_id,
relation_path: relationPathHints(domain, unit),
missing_links,
conflicting_links,
provenance_evidence_ids: evidenceIds,
graph_confidence: confidence
});
}
const exported = accumulator.export();
const summary = buildSummary({
nodes: exported.nodes,
edges: exported.edges,
bindings,
totalUnits: input.problemUnits.length
});
if (summary.bound_units < summary.total_units) {
issues.push("some_problem_units_not_bound_to_graph");
}
if (summary.node_count === 0 || summary.edge_count === 0) {
issues.push("graph_runtime_empty");
}
return {
schema_version: stage4Graph_1.ACCOUNTING_GRAPH_SCHEMA_VERSION,
nodes: exported.nodes,
edges: exported.edges,
unit_bindings: bindings,
summary,
issues: uniqueStrings(issues, 8)
};
}
llm_normalizer/backend/dist/services/traceLogger.js
Vendored
+34 -7
View File
@@ -19,10 +19,22 @@ function redactSecrets(payload) {
delete output.apiKey;
return output;
}
function isNoSpaceError(error) {
const code = error?.code;
return code === "ENOSPC";
}
function saveTrace(record) {
(0, files_1.ensureDir)(config_1.TRACES_DIR);
const target = path_1.default.resolve(config_1.TRACES_DIR, `${record.trace_id}.json`);
(0, files_1.writeJsonFile)(target, record);
try {
(0, files_1.ensureDir)(config_1.TRACES_DIR);
const target = path_1.default.resolve(config_1.TRACES_DIR, `${record.trace_id}.json`);
(0, files_1.writeJsonFile)(target, record);
}
catch (error) {
if (isNoSpaceError(error)) {
return;
}
throw error;
}
}
function listTraces(limit = 100) {
(0, files_1.ensureDir)(config_1.TRACES_DIR);
@@ -60,8 +72,16 @@ function getTrace(traceId) {
return JSON.parse(raw);
}
function savePreset(preset) {
(0, files_1.ensureDir)(config_1.PRESETS_DIR);
(0, files_1.writeJsonFile)(path_1.default.resolve(config_1.PRESETS_DIR, `${preset.id}.json`), preset);
try {
(0, files_1.ensureDir)(config_1.PRESETS_DIR);
(0, files_1.writeJsonFile)(path_1.default.resolve(config_1.PRESETS_DIR, `${preset.id}.json`), preset);
}
catch (error) {
if (isNoSpaceError(error)) {
return;
}
throw error;
}
}
function listPresets() {
(0, files_1.ensureDir)(config_1.PRESETS_DIR);
@@ -75,9 +95,16 @@ function listPresets() {
.sort((a, b) => b.updatedAt.localeCompare(a.updatedAt));
}
function saveEvalCase(casePayload) {
(0, files_1.ensureDir)(config_1.EVAL_CASES_DIR);
const id = String(casePayload.case_id ?? `NQ-${Date.now()}`);
(0, files_1.writeJsonFile)(path_1.default.resolve(config_1.EVAL_CASES_DIR, `${id}.json`), casePayload);
try {
(0, files_1.ensureDir)(config_1.EVAL_CASES_DIR);
(0, files_1.writeJsonFile)(path_1.default.resolve(config_1.EVAL_CASES_DIR, `${id}.json`), casePayload);
}
catch (error) {
if (!isNoSpaceError(error)) {
throw error;
}
}
return id;
}
function redactRequestPayload(payload) {