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

Этап 4 / Волна 16: добивка остаточных ошибок и устранение шаблонных ответов

This commit is contained in:
dctouch 2026-03-28 16:23:41 +03:00
parent 014ff65188
commit 6123fafd5b
8 changed files with 787 additions and 51 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

42
llm_normalizer/backend/dist/services/answerComposer.js vendored
View File

@ -1078,7 +1078,7 @@ function isProblemUnitAlignedWithNarrativeDomain(unit, domain) {
if (accounts.some((item) => isVatAccountToken(item))) {
return true;
}
return /(vat|ндс|invoice|book_entry|register|книг|счет[\s-]?фактур|сч[её]т[\s-]?фактур)/i.test(corpus);
return /(vat|ндс|invoice|book_entry|register|книг|сч[её]т(?:а|у|ом|е)?[\s-]?фактур(?:а|ы|е|у|ой)?|вычет|налогов(?:ый|ого)?\s+эффект)/i.test(corpus);
}
if (domain === "month_close_costs_20_44") {
const foreignMonthCloseDomain = ["vat_flow", "bank_settlement", "customer_settlement", "fixed_asset"].includes(String(unit.lifecycle_domain ?? ""));
@ -2030,11 +2030,11 @@ function hasStrongNarrativeDomainSignalInText(userMessage, domain) {
const accountTokens = extractAccountNumbersFromNarrativeText(text);
if (domain === "settlements_60_62") {
return (accountTokens.some((item) => isSettlementAccountToken(item)) ||
/(60\.0[12]|62\.0[12]|долг|аванс|зач[её]т|взаимозач|расч[её]т)/i.test(text));
/(60\.0[12]|62\.0[12]|долг|аванс|зач[её]т|взаимозач|расч[её]т|оплат|плат[её]ж|деньг[аи])/i.test(text));
}
if (domain === "vat_document_register_book") {
return (accountTokens.some((item) => isVatAccountToken(item)) ||
/(ндс|vat|счет[-\s]?фактур|сч[её]т[-\s]?фактур|книг[аи]|регистр)/i.test(text));
/(ндс|vat|сч[её]т(?:а|у|ом|е)?[-\s]?фактур(?:а|ы|е|у|ой)?|книг[аи]|регистр|вычет|налогов(?:ый|ого)?\s+эффект)/i.test(text));
}
if (domain === "month_close_costs_20_44") {
return (accountTokens.some((item) => isCloseCostsAccountToken(item)) ||
@ -2401,7 +2401,7 @@ function buildProblemCentricAnswerStructure(input) {
}
function limitationReasonToUserText(code) {
if (code === "snapshot_only")
return "Вывод сделан по snapshot и может не включать самые свежие изменения.";
return "Оценка построена на snapshot-срезе и может не включать самые свежие изменения.";
if (code === "heuristic_inference")
return "Часть вывода построена эвристически и требует проверки в базе.";
if (code === "missing_mechanism")
@ -2430,7 +2430,10 @@ function inferNarrativeDomainFromText(value) {
if (/(долг|аванс|взаимозач|зачет|зачёт|62\.01|62\.02|60\.01|60\.02|не\s+сход)/i.test(text)) {
settlementScore += 2;
}
if (/(ндс|vat|счет[-\s]?фактур|сч[её]т[-\s]?фактур|книг[аи]|регистр)/i.test(text)) {
if (/(расч[её]т|оплат|плат[её]ж|деньг[аи]|закрыти[ея]\s+расч)/i.test(text)) {
settlementScore += 2;
}
if (/(ндс|vat|сч[её]т(?:а|у|ом|е)?[-\s]?фактур(?:а|ы|е|у|ой)?|книг[аи]|регистр|вычет|налогов(?:ый|ого)?\s+эффект)/i.test(text)) {
vatScore += 3;
}
if (/(закрыти[ея]\s+месяц|закрытие\s+счетов|регламентн|косвенн|затрат|распределени|рбп|амортиз|финансовых\s+результат|month\s*close|period\s*close|close\s+operation)/i.test(text)) {
@ -2595,8 +2598,8 @@ function buildEvidenceSectionLines(structure, questionType = "unknown") {
function buildDefaultChecksByDomain(domain) {
if (domain === "settlements_60_62") {
return [
"Проверьте договор, объект расчетов, регистр расчетов и документ закрытия (зачет аванса или взаимозачет).",
"Сверьте связку платеж -> расчетный документ -> проводки по 60/62/76 и подтверждение закрытия хвоста."
"Сверьте договор и объект расчетов, затем подтвердите запись в регистре расчетов и документ зачета.",
"Проверьте связку платеж -> расчетный документ -> проводки по 60/62/76 и факт закрытия хвоста."
];
}
if (domain === "vat_document_register_book") {
@ -2753,11 +2756,12 @@ function buildChecksSectionLines(structure, context) {
lines.push(...domainFallback.slice(0, 2));
lines.push(...actionLines.slice(0, 2));
}
else if (actionLines.length > 0) {
lines.push(...actionLines.slice(0, 2));
}
else {
lines.push(...domainFallback.slice(0, 2));
lines.push(...domainFallback.slice(0, 1));
lines.push(...actionLines.slice(0, 2));
if (lines.length < 2) {
lines.push(...domainFallback.slice(1, 2));
}
}
}
if (hasMissingPeriod) {
@ -2795,7 +2799,7 @@ function humanizeLimitationToken(value) {
if (normalized === "settlement_primary_evidence_not_confirmed")
return "Опора по расчетному контуру не подтверждена: в приоритете были сигналы из смежных доменов.";
if (normalized.includes("snapshot"))
return "Вывод сделан по snapshot и может не включать часть цепочки.";
return "Оценка сделана на snapshot-срезе и может не включать часть цепочки.";
if (normalized.includes("heuristic"))
return "Часть вывода основана на эвристике.";
if (normalized.includes("weak_source_mapping"))
@ -2823,7 +2827,7 @@ function humanizeLimitationToken(value) {
if (/weak mechanism evidence/i.test(raw))
return "Доказательность механизма слабая, нужен ручной контроль.";
if (/evidence is snapshot-only/i.test(raw))
return "Вывод сделан по snapshot и может не включать самые свежие изменения.";
return "Оценка сделана на snapshot-срезе и может не включать самые свежие изменения.";
if (/source-of-record/i.test(raw))
return "Часть цепочки нужно подтвердить в исходной учетной базе.";
if (/[a-z]/i.test(raw) && !/[а-яё]/iu.test(raw))
@ -2881,6 +2885,18 @@ function buildQuestionTypeShortLine(context) {
if (context.questionType === "what_to_check_first") {
return `\u041a\u043e\u0440\u043e\u0442\u043a\u0438\u0439 \u043c\u0430\u0440\u0448\u0440\u0443\u0442 \u043f\u0435\u0440\u0432\u044b\u0445 \u043f\u0440\u043e\u0432\u0435\u0440\u043e\u043a \u0432\u043d\u0443\u0442\u0440\u0438 ${domainName}.`;
}
if (context.questionType === "why_breaks") {
if (context.focusDomain === "settlements_60_62") {
return "Наиболее вероятная причина: переход от оплаты к закрытию расчета подтвержден не полностью.";
}
if (context.focusDomain === "vat_document_register_book") {
return "Наиболее вероятная причина: переход документа НДС к регистру и книге подтвержден частично.";
}
if (context.focusDomain === "month_close_costs_20_44") {
return "Наиболее вероятная причина: цепочка распределения затрат и закрытия месяца подтверждена не полностью.";
}
return "Наиболее вероятный механизм проблемы подтвержден частично и требует первичной проверки.";
}
return null;
}
function buildQuestionTypeBrokenLine(context) {

135
llm_normalizer/backend/dist/services/assistantDataLayer.js vendored
View File

@ -245,10 +245,11 @@ const P0_DOMAIN_CARDS = [
symptom_markers: [
/\bvat\b/i,
/\u043d\u0434\u0441/i,
/\u0441\u0447[её]т.?фактур/i,
/\u0441\u0447[её]т(?:а|у|ом|е)?.?фактур/i,
/\u043a\u043d\u0438\u0433[аи]\s+\u043f\u043e\u043a\u0443\u043f/i,
/\u043a\u043d\u0438\u0433[аи]\s+\u043f\u0440\u043e\u0434\u0430\u0436/i,
/\u0432\u044b\u0447\u0435\u0442/i
/\u0432\u044b\u0447\u0435\u0442/i,
/\u043d\u0430\u043b\u043e\u0433\u043e\u0432(?:\u044b\u0439|\u043e\u0433\u043e)?\s+\u044d\u0444\u0444\u0435\u043a\u0442/i
]
},
{
@ -857,6 +858,35 @@ function collectDateLikeSpans(text) {
}
return spans;
}
function collectAmountLikeSpans(text) {
const spans = [];
const patterns = [
/\b\d{1,3}(?:[ \u00A0]\d{3})+(?:[.,]\d{2})?\b/g,
/\b\d+[.,]\d{2}\b/g
];
for (const pattern of patterns) {
let match = null;
while ((match = pattern.exec(text)) !== null) {
spans.push({
start: match.index,
end: match.index + match[0].length
});
}
}
return spans;
}
function collectPercentLikeSpans(text) {
const spans = [];
const pattern = /\b\d{1,3}(?:[.,]\d+)?\s*%/g;
let match = null;
while ((match = pattern.exec(text)) !== null) {
spans.push({
start: match.index,
end: match.index + match[0].length
});
}
return spans;
}
function intersectsSpan(start, end, spans) {
return spans.some((span) => start < span.end && end > span.start);
}
@ -867,7 +897,11 @@ function hasAccountContextAround(text, start, end) {
}
function extractAccountScopeFromText(text) {
const lower = String(text ?? "").toLowerCase();
const dateSpans = collectDateLikeSpans(lower);
const blockedSpans = [
...collectDateLikeSpans(lower),
...collectAmountLikeSpans(lower),
...collectPercentLikeSpans(lower)
];
const accounts = [];
const pushAccount = (raw) => {
const prefix = String(raw ?? "").trim().match(/^(\d{2})/)?.[1];
@ -912,7 +946,7 @@ function extractAccountScopeFromText(text) {
const token = suffixAnchorMatch[0];
const start = suffixAnchorMatch.index;
const end = start + token.length;
if (intersectsSpan(start, end, dateSpans)) {
if (intersectsSpan(start, end, blockedSpans)) {
continue;
}
pushAccount(token);
@ -924,7 +958,7 @@ function extractAccountScopeFromText(text) {
const token = explicitMatch[0];
const start = explicitMatch.index;
const end = start + token.length;
if (intersectsSpan(start, end, dateSpans)) {
if (intersectsSpan(start, end, blockedSpans)) {
continue;
}
const prefix = token.match(/^(\d{2})/)?.[1];
@ -1090,7 +1124,7 @@ function buildSemanticRetrievalProfile(fragmentText) {
pushMany(entityTypes, ["counterparty", "contract", "document", "posting"]);
pushMany(relationPatterns, ["payment_to_settlement", "contract_to_documents"]);
}
if (/РЅРґСЃ|ндс|vat|РєРЅРёРіР° РїРѕРєСѓРїРѕРє|РєРЅРёРіР° продаж|счет.?фактур|книг[аи]\s+покуп|книг[аи]\s+продаж|сч[её]т.?фактур/i.test(lower) ||
if (/РЅРґСЃ|ндс|vat|РєРЅРёРіР° РїРѕРєСѓРїРѕРє|РєРЅРёРіР° продаж|счет.?фактур|книг[аи]\s+покуп|книг[аи]\s+продаж|сч[её]т(?:а|у|ом|е)?[-\s]?фактур(?:а|ы|е|у|ой)?|вычет|налогов(?:ый|ого)?\s+эффект/i.test(lower) ||
hasVatAccountScope) {
pushMany(domainScope, ["vat", "taxes"]);
pushMany(documentTypes, ["invoice", "vat_document"]);
@ -1207,12 +1241,24 @@ function cardResolutionScore(card, fragmentText, profile) {
if (hasExplicitAccountScope && accountMatches.length === 0) {
return 0;
}
const hasHardAnchor = accountMatches.length > 0 || markerHit;
const hasVatSoftAnchor = card.id === "vat_document_register_book" && hasStrongVatDomainSignal(fragmentText, profile);
const hasHardAnchor = accountMatches.length > 0 || markerHit || hasVatSoftAnchor;
if (!hasHardAnchor) {
return 0;
}
return accountMatches.length * 4 + domainMatches.length * 3 + (markerHit ? 2 : 0);
}
function hasStrongVatDomainSignal(fragmentText, profile) {
const text = String(fragmentText ?? "");
const hasVatLexicalAnchor = /(?:ндс|vat|сч[её]т(?:а|у|ом|е)?[-\s]?фактур(?:а|ы|е|у|ой)?|книг[аи]\s+(?:покуп|продаж)|вычет|налогов(?:ый|ого)?\s+эффект)/iu.test(text);
return (hasVatLexicalAnchor ||
profile.account_scope.some((account) => account === "19" || account === "68") ||
profile.domain_scope.some((domain) => domain === "vat" || domain === "taxes") ||
profile.relation_patterns.some((pattern) => ["invoice_to_vat", "register_to_book", "book_entry_generated", "deduction_posted"].includes(pattern)));
}
function hasStrongSettlementAccountSignal(profile) {
return profile.account_scope.some((account) => account === "51" || account === "60" || account === "62" || account === "76");
}
function resolveP0DomainCard(fragmentText, profile) {
const resolved = P0_DOMAIN_CARDS.map((card) => ({
card,
@ -1225,6 +1271,11 @@ function resolveP0DomainCard(fragmentText, profile) {
}
const [first, second] = resolved;
if (second && second.score === first.score) {
const pair = new Set([first.card.id, second.card.id]);
const hasVatSettlementTie = pair.has("vat_document_register_book") && pair.has("settlements_60_62");
if (hasVatSettlementTie && hasStrongVatDomainSignal(fragmentText, profile) && !hasStrongSettlementAccountSignal(profile)) {
return resolved.find((item) => item.card.id === "vat_document_register_book") ?? null;
}
return null;
}
return first;
@ -2823,7 +2874,39 @@ class AssistantDataLayer {
}
executeBatch(fragmentText, data) {
const semanticProfile = buildSemanticRetrievalProfile(fragmentText);
const source = [...data.problemCases, ...data.keyFields, ...data.docs];
const resolvedDomain = resolveP0DomainCard(fragmentText, semanticProfile);
const domainCard = resolvedDomain?.card ?? null;
const fallbackSources = ["problemCases", "keyFields", "docs"];
const sourceScope = domainCard
? uniqueStrings([...domainCard.allowed_evidence_sources.risk, ...domainCard.allowed_evidence_sources.canonical])
: fallbackSources;
const sourcePool = collectSourceRecords(data, sourceScope);
const strictForbidden = Boolean(domainCard);
let sourceGate = domainCard
? applyDomainPuritySourceGate(sourcePool, domainCard, semanticProfile, { strict_forbidden: strictForbidden })
: {
accepted: sourcePool.map((item) => ({
...item,
signals: inferRecordSignals(item.record),
purity: {
allowed: true,
account_match: true,
domain_match: true,
entity_match: true,
edge_match: true,
forbidden_domains: [],
cross_domain_overlap: []
}
})),
rejected_total: 0,
rejected_forbidden: 0
};
let sourceStrictFallbackUsed = false;
if (domainCard && strictForbidden && sourceGate.accepted.length === 0 && sourcePool.length > 0) {
sourceGate = applyDomainPuritySourceGate(sourcePool, domainCard, semanticProfile, { strict_forbidden: false });
sourceStrictFallbackUsed = true;
}
const source = sourceGate.accepted.map((item) => item.record);
const byEntity = new Map();
for (const record of source) {
byEntity.set(record.source_entity, (byEntity.get(record.source_entity) ?? 0) + 1);
@ -2836,29 +2919,55 @@ class AssistantDataLayer {
entity,
records_count: count
}));
const puritySummary = {
enabled: Boolean(domainCard),
domain_card_id: domainCard?.id ?? null,
domain_card_title: domainCard?.title ?? null,
source_scope: sourceScope,
source_pool_records: sourcePool.length,
source_selection_allowed: sourceGate.accepted.length,
source_selection_rejected: sourceGate.rejected_total,
source_selection_rejected_forbidden: sourceGate.rejected_forbidden,
top1_pure: domainCard ? topOnePurityHolds(sourceGate.accepted) : true,
top3_pure: domainCard ? topThreePurityHolds(sourceGate.accepted) : true,
strict_forbidden_mode: strictForbidden,
strict_forbidden_fallback_source: sourceStrictFallbackUsed
};
return {
status: items.length > 0 ? "ok" : "empty",
result_type: "ranking",
items,
summary: {
checked_records: source.length,
checked_records: sourcePool.length,
ranked_entities: items.length,
query_subject: semanticProfile.query_subject,
semantic_profile: semanticProfile,
ranking_basis: semanticProfile.ranking_basis
ranking_basis: semanticProfile.ranking_basis,
domain_purity_guard: puritySummary
},
evidence: items.slice(0, 5).map((item) => ({
entity: item.entity,
records_count: item.records_count
})),
why_included: items.length > 0 ? ["Показаны сущности с максимальным количеством записей."] : [],
selection_reason: ["Ранжирование выполнено по records_count по убыванию."],
why_included: items.length > 0
? [
"Показаны сущности с максимальным количеством записей.",
domainCard ? `P0 domain purity enforced for ${domainCard.id}.` : "P0 domain purity was not enforced."
]
: [],
selection_reason: [
"Ранжирование выполнено по records_count по убыванию.",
domainCard ? `Domain gate source scope: ${sourceScope.join(", ")}.` : "Domain gate source scope not applied."
],
risk_factors: uniqueStrings(["entity_volume_spike", ...semanticProfile.anomaly_patterns]),
business_interpretation: [
"Top entities by volume highlight where lifecycle-focused review should start first."
],
confidence: "medium",
limitations: ["Ранжирование по объему не всегда эквивалентно бизнес-риску."],
limitations: [
"Ранжирование по объему не всегда эквивалентно бизнес-риску.",
domainCard ? "Domain purity guardrail может исключить cross-domain записи на batch-слое." : "Domain purity guardrail не применялся."
],
errors: []
};
}

13
llm_normalizer/backend/dist/services/questionTypeResolver.js vendored
View File

@ -64,11 +64,24 @@ function countRuleHits(text, rule) {
}
return hits;
}
function hasProofIntent(text) {
return /(?:\bprove\b|\bguess\b|доказан|доказано|доказуем|гипотез|догад|связан\s+ли|зач[её]л(?:ся|ось)\s+ли)/iu.test(text);
}
function hasExplicitChainSplitIntent(text) {
return /(?:какие(?:\s+\S+){0,4}\s+цепочк[аи]|which\s+chains?|complete\s+vs\s+incomplete|что\s+закрыто.*что\s+нет)/iu.test(text);
}
function hasGroundingIntent(text) {
return /(?:на\s+ч(?:е|ё)м[^?!.]{0,40}основан|чем\s+подтвержда|какие\s+основани|what\s+evidence|grounded\s+on|based\s+on)/iu.test(text);
}
function resolveQuestionType(input) {
const text = String(input ?? "").trim();
if (!text) {
return "unknown";
}
// Guard against collapsing proof-intent questions into chain classification.
if (hasProofIntent(text) && !hasExplicitChainSplitIntent(text) && !hasGroundingIntent(text)) {
return "prove_or_guess";
}
let bestType = "unknown";
let bestHits = 0;
let bestPriority = Number.POSITIVE_INFINITY;

49
llm_normalizer/backend/src/services/answerComposer.ts
View File

@ -1262,7 +1262,9 @@ function isProblemUnitAlignedWithNarrativeDomain(unit: ProblemUnit, domain: P0Na
if (accounts.some((item) => isVatAccountToken(item))) {
return true;
}
return /(vat|ндс|invoice|book_entry|register|книг|счет[\s-]?фактур|сч[её]т[\s-]?фактур)/i.test(corpus);
return /(vat|ндс|invoice|book_entry|register|книг|сч[её]т(?:а|у|ом|е)?[\s-]?фактур(?:а|ы|е|у|ой)?|вычет|налогов(?:ый|ого)?\s+эффект)/i.test(
corpus
);
}
if (domain === "month_close_costs_20_44") {
@ -2387,13 +2389,15 @@ function hasStrongNarrativeDomainSignalInText(userMessage: string, domain: P0Nar
if (domain === "settlements_60_62") {
return (
accountTokens.some((item) => isSettlementAccountToken(item)) ||
/(60\.0[12]|62\.0[12]|долг|аванс|зач[её]т|взаимозач|расч[её]т)/i.test(text)
/(60\.0[12]|62\.0[12]|долг|аванс|зач[её]т|взаимозач|расч[её]т|оплат|плат[её]ж|деньг[аи])/i.test(text)
);
}
if (domain === "vat_document_register_book") {
return (
accountTokens.some((item) => isVatAccountToken(item)) ||
/(ндс|vat|счет[-\s]?фактур|сч[её]т[-\s]?фактур|книг[аи]|регистр)/i.test(text)
/(ндс|vat|сч[её]т(?:а|у|ом|е)?[-\s]?фактур(?:а|ы|е|у|ой)?|книг[аи]|регистр|вычет|налогов(?:ый|ого)?\s+эффект)/i.test(
text
)
);
}
if (domain === "month_close_costs_20_44") {
@ -2855,7 +2859,7 @@ function buildProblemCentricAnswerStructure(input: {
}
function limitationReasonToUserText(code: EvidenceLimitationReasonCode): string {
if (code === "snapshot_only") return "Вывод сделан по snapshot и может не включать самые свежие изменения.";
if (code === "snapshot_only") return "Оценка построена на snapshot-срезе и может не включать самые свежие изменения.";
if (code === "heuristic_inference") return "Часть вывода построена эвристически и требует проверки в базе.";
if (code === "missing_mechanism") return "Механизм проблемы подтвержден не полностью.";
if (code === "weak_source_mapping") return "Связка между источниками подтверждена частично.";
@ -2884,7 +2888,14 @@ function inferNarrativeDomainFromText(value: string): P0NarrativeDomain {
if (/(долг|аванс|взаимозач|зачет|зачёт|62\.01|62\.02|60\.01|60\.02|не\s+сход)/i.test(text)) {
settlementScore += 2;
}
if (/(ндс|vat|счет[-\s]?фактур|сч[её]т[-\s]?фактур|книг[аи]|регистр)/i.test(text)) {
if (/(расч[её]т|оплат|плат[её]ж|деньг[аи]|закрыти[ея]\s+расч)/i.test(text)) {
settlementScore += 2;
}
if (
/(ндс|vat|сч[её]т(?:а|у|ом|е)?[-\s]?фактур(?:а|ы|е|у|ой)?|книг[аи]|регистр|вычет|налогов(?:ый|ого)?\s+эффект)/i.test(
text
)
) {
vatScore += 3;
}
if (
@ -3080,8 +3091,8 @@ function buildEvidenceSectionLines(
function buildDefaultChecksByDomain(domain: P0NarrativeDomain): string[] {
if (domain === "settlements_60_62") {
return [
"Проверьте договор, объект расчетов, регистр расчетов и документ закрытия (зачет аванса или взаимозачет).",
"Сверьте связку платеж -> расчетный документ -> проводки по 60/62/76 и подтверждение закрытия хвоста."
"Сверьте договор и объект расчетов, затем подтвердите запись в регистре расчетов и документ зачета.",
"Проверьте связку платеж -> расчетный документ -> проводки по 60/62/76 и факт закрытия хвоста."
];
}
if (domain === "vat_document_register_book") {
@ -3248,10 +3259,12 @@ function buildChecksSectionLines(structure: AnswerStructureV11, context?: Answer
if (domain === "settlements_60_62") {
lines.push(...domainFallback.slice(0, 2));
lines.push(...actionLines.slice(0, 2));
} else if (actionLines.length > 0) {
lines.push(...actionLines.slice(0, 2));
} else {
lines.push(...domainFallback.slice(0, 2));
lines.push(...domainFallback.slice(0, 1));
lines.push(...actionLines.slice(0, 2));
if (lines.length < 2) {
lines.push(...domainFallback.slice(1, 2));
}
}
}
if (hasMissingPeriod) {
@ -3283,7 +3296,7 @@ function humanizeLimitationToken(value: string): string | null {
return "Целевой механизм активного домена подтвержден частично; вывод ограничен.";
if (normalized === "settlement_primary_evidence_not_confirmed")
return "Опора по расчетному контуру не подтверждена: в приоритете были сигналы из смежных доменов.";
if (normalized.includes("snapshot")) return "Вывод сделан по snapshot и может не включать часть цепочки.";
if (normalized.includes("snapshot")) return "Оценка сделана на snapshot-срезе и может не включать часть цепочки.";
if (normalized.includes("heuristic")) return "Часть вывода основана на эвристике.";
if (normalized.includes("weak_source_mapping")) return "Связка между источниками подтверждена частично.";
if (normalized.includes("missing_mechanism")) return "Механизм проблемы подтвержден не полностью.";
@ -3297,7 +3310,7 @@ function humanizeLimitationToken(value: string): string | null {
if (/broad query support is limited/i.test(raw)) return "Запрос широкий, поэтому вывод ограничен частичной опорой.";
if (/broad ranking output was tightened/i.test(raw)) return "Часть ранжирования ограничена, чтобы избежать ложной точности.";
if (/weak mechanism evidence/i.test(raw)) return "Доказательность механизма слабая, нужен ручной контроль.";
if (/evidence is snapshot-only/i.test(raw)) return "Вывод сделан по snapshot и может не включать самые свежие изменения.";
if (/evidence is snapshot-only/i.test(raw)) return "Оценка сделана на snapshot-срезе и может не включать самые свежие изменения.";
if (/source-of-record/i.test(raw)) return "Часть цепочки нужно подтвердить в исходной учетной базе.";
if (/[a-z]/i.test(raw) && !/[а-яё]/iu.test(raw)) return null;
@ -3361,6 +3374,18 @@ function buildQuestionTypeShortLine(context: AnswerRenderContext): string | null
if (context.questionType === "what_to_check_first") {
return `\u041a\u043e\u0440\u043e\u0442\u043a\u0438\u0439 \u043c\u0430\u0440\u0448\u0440\u0443\u0442 \u043f\u0435\u0440\u0432\u044b\u0445 \u043f\u0440\u043e\u0432\u0435\u0440\u043e\u043a \u0432\u043d\u0443\u0442\u0440\u0438 ${domainName}.`;
}
if (context.questionType === "why_breaks") {
if (context.focusDomain === "settlements_60_62") {
return "Наиболее вероятная причина: переход от оплаты к закрытию расчета подтвержден не полностью.";
}
if (context.focusDomain === "vat_document_register_book") {
return "Наиболее вероятная причина: переход документа НДС к регистру и книге подтвержден частично.";
}
if (context.focusDomain === "month_close_costs_20_44") {
return "Наиболее вероятная причина: цепочка распределения затрат и закрытия месяца подтверждена не полностью.";
}
return "Наиболее вероятный механизм проблемы подтвержден частично и требует первичной проверки.";
}
return null;
}

147
llm_normalizer/backend/src/services/assistantDataLayer.ts
View File

@ -388,10 +388,11 @@ const P0_DOMAIN_CARDS: P0DomainCard[] = [
symptom_markers: [
/\bvat\b/i,
/\u043d\u0434\u0441/i,
/\u0441\u0447[её]т.?фактур/i,
/\u0441\u0447[её]т(?:а|у|ом|е)?.?фактур/i,
/\u043a\u043d\u0438\u0433[аи]\s+\u043f\u043e\u043a\u0443\u043f/i,
/\u043a\u043d\u0438\u0433[аи]\s+\u043f\u0440\u043e\u0434\u0430\u0436/i,
/\u0432\u044b\u0447\u0435\u0442/i
/\u0432\u044b\u0447\u0435\u0442/i,
/\u043d\u0430\u043b\u043e\u0433\u043e\u0432(?:\u044b\u0439|\u043e\u0433\u043e)?\s+\u044d\u0444\u0444\u0435\u043a\u0442/i
]
},
{
@ -1209,6 +1210,37 @@ function collectDateLikeSpans(text: string): Array<{ start: number; end: number
return spans;
}
function collectAmountLikeSpans(text: string): Array<{ start: number; end: number }> {
const spans: Array<{ start: number; end: number }> = [];
const patterns = [
/\b\d{1,3}(?:[ \u00A0]\d{3})+(?:[.,]\d{2})?\b/g,
/\b\d+[.,]\d{2}\b/g
];
for (const pattern of patterns) {
let match: RegExpExecArray | null = null;
while ((match = pattern.exec(text)) !== null) {
spans.push({
start: match.index,
end: match.index + match[0].length
});
}
}
return spans;
}
function collectPercentLikeSpans(text: string): Array<{ start: number; end: number }> {
const spans: Array<{ start: number; end: number }> = [];
const pattern = /\b\d{1,3}(?:[.,]\d+)?\s*%/g;
let match: RegExpExecArray | null = null;
while ((match = pattern.exec(text)) !== null) {
spans.push({
start: match.index,
end: match.index + match[0].length
});
}
return spans;
}
function intersectsSpan(start: number, end: number, spans: Array<{ start: number; end: number }>): boolean {
return spans.some((span) => start < span.end && end > span.start);
}
@ -1221,7 +1253,11 @@ function hasAccountContextAround(text: string, start: number, end: number): bool
function extractAccountScopeFromText(text: string): string[] {
const lower = String(text ?? "").toLowerCase();
const dateSpans = collectDateLikeSpans(lower);
const blockedSpans = [
...collectDateLikeSpans(lower),
...collectAmountLikeSpans(lower),
...collectPercentLikeSpans(lower)
];
const accounts: string[] = [];
const pushAccount = (raw: string): void => {
const prefix = String(raw ?? "").trim().match(/^(\d{2})/)?.[1];
@ -1271,7 +1307,7 @@ function extractAccountScopeFromText(text: string): string[] {
const token = suffixAnchorMatch[0];
const start = suffixAnchorMatch.index;
const end = start + token.length;
if (intersectsSpan(start, end, dateSpans)) {
if (intersectsSpan(start, end, blockedSpans)) {
continue;
}
pushAccount(token);
@ -1286,7 +1322,7 @@ function extractAccountScopeFromText(text: string): string[] {
const token = explicitMatch[0];
const start = explicitMatch.index;
const end = start + token.length;
if (intersectsSpan(start, end, dateSpans)) {
if (intersectsSpan(start, end, blockedSpans)) {
continue;
}
const prefix = token.match(/^(\d{2})/)?.[1];
@ -1465,7 +1501,7 @@ function buildSemanticRetrievalProfile(fragmentText: string): SemanticRetrievalP
pushMany(relationPatterns, ["payment_to_settlement", "contract_to_documents"]);
}
if (
/РЅРґСЃ|ндс|vat|РєРЅРёРіР° РїРѕРєСѓРїРѕРє|РєРЅРёРіР° продаж|счет.?фактур|книг[аи]\s+покуп|книг[аи]\s+продаж|сч[её]т.?фактур/i.test(
/РЅРґСЃ|ндс|vat|РєРЅРёРіР° РїРѕРєСѓРїРѕРє|РєРЅРёРіР° продаж|счет.?фактур|книг[аи]\s+покуп|книг[аи]\s+продаж|сч[её]т(?:а|у|ом|е)?[-\s]?фактур(?:а|ы|е|у|ой)?|вычет|налогов(?:ый|ого)?\s+эффект/i.test(
lower
) ||
hasVatAccountScope
@ -1599,7 +1635,8 @@ function cardResolutionScore(card: P0DomainCard, fragmentText: string, profile:
return 0;
}
const hasHardAnchor = accountMatches.length > 0 || markerHit;
const hasVatSoftAnchor = card.id === "vat_document_register_book" && hasStrongVatDomainSignal(fragmentText, profile);
const hasHardAnchor = accountMatches.length > 0 || markerHit || hasVatSoftAnchor;
if (!hasHardAnchor) {
return 0;
}
@ -1607,6 +1644,26 @@ function cardResolutionScore(card: P0DomainCard, fragmentText: string, profile:
return accountMatches.length * 4 + domainMatches.length * 3 + (markerHit ? 2 : 0);
}
function hasStrongVatDomainSignal(fragmentText: string, profile: SemanticRetrievalProfile): boolean {
const text = String(fragmentText ?? "");
const hasVatLexicalAnchor =
/(?:ндс|vat|сч[её]т(?:а|у|ом|е)?[-\s]?фактур(?:а|ы|е|у|ой)?|книг[аи]\s+(?:покуп|продаж)|вычет|налогов(?:ый|ого)?\s+эффект)/iu.test(
text
);
return (
hasVatLexicalAnchor ||
profile.account_scope.some((account) => account === "19" || account === "68") ||
profile.domain_scope.some((domain) => domain === "vat" || domain === "taxes") ||
profile.relation_patterns.some((pattern) =>
["invoice_to_vat", "register_to_book", "book_entry_generated", "deduction_posted"].includes(pattern)
)
);
}
function hasStrongSettlementAccountSignal(profile: SemanticRetrievalProfile): boolean {
return profile.account_scope.some((account) => account === "51" || account === "60" || account === "62" || account === "76");
}
function resolveP0DomainCard(fragmentText: string, profile: SemanticRetrievalProfile): P0DomainResolution | null {
const resolved = P0_DOMAIN_CARDS.map((card) => ({
card,
@ -1620,6 +1677,11 @@ function resolveP0DomainCard(fragmentText: string, profile: SemanticRetrievalPro
}
const [first, second] = resolved;
if (second && second.score === first.score) {
const pair = new Set<P0DomainCardId>([first.card.id, second.card.id]);
const hasVatSettlementTie = pair.has("vat_document_register_book") && pair.has("settlements_60_62");
if (hasVatSettlementTie && hasStrongVatDomainSignal(fragmentText, profile) && !hasStrongSettlementAccountSignal(profile)) {
return resolved.find((item) => item.card.id === "vat_document_register_book") ?? null;
}
return null;
}
return first;
@ -3443,7 +3505,39 @@ export class AssistantDataLayer {
private executeBatch(fragmentText: string, data: DatasetBundle): RawRetrievalResult {
const semanticProfile = buildSemanticRetrievalProfile(fragmentText);
const source = [...data.problemCases, ...data.keyFields, ...data.docs];
const resolvedDomain = resolveP0DomainCard(fragmentText, semanticProfile);
const domainCard = resolvedDomain?.card ?? null;
const fallbackSources: DatasetSourceName[] = ["problemCases", "keyFields", "docs"];
const sourceScope = domainCard
? (uniqueStrings([...domainCard.allowed_evidence_sources.risk, ...domainCard.allowed_evidence_sources.canonical]) as DatasetSourceName[])
: fallbackSources;
const sourcePool = collectSourceRecords(data, sourceScope);
const strictForbidden = Boolean(domainCard);
let sourceGate = domainCard
? applyDomainPuritySourceGate(sourcePool, domainCard, semanticProfile, { strict_forbidden: strictForbidden })
: {
accepted: sourcePool.map((item) => ({
...item,
signals: inferRecordSignals(item.record),
purity: {
allowed: true,
account_match: true,
domain_match: true,
entity_match: true,
edge_match: true,
forbidden_domains: [],
cross_domain_overlap: []
}
})),
rejected_total: 0,
rejected_forbidden: 0
};
let sourceStrictFallbackUsed = false;
if (domainCard && strictForbidden && sourceGate.accepted.length === 0 && sourcePool.length > 0) {
sourceGate = applyDomainPuritySourceGate(sourcePool, domainCard, semanticProfile, { strict_forbidden: false });
sourceStrictFallbackUsed = true;
}
const source = sourceGate.accepted.map((item) => item.record);
const byEntity = new Map<string, number>();
for (const record of source) {
byEntity.set(record.source_entity, (byEntity.get(record.source_entity) ?? 0) + 1);
@ -3458,29 +3552,56 @@ export class AssistantDataLayer {
records_count: count
}));
const puritySummary = {
enabled: Boolean(domainCard),
domain_card_id: domainCard?.id ?? null,
domain_card_title: domainCard?.title ?? null,
source_scope: sourceScope,
source_pool_records: sourcePool.length,
source_selection_allowed: sourceGate.accepted.length,
source_selection_rejected: sourceGate.rejected_total,
source_selection_rejected_forbidden: sourceGate.rejected_forbidden,
top1_pure: domainCard ? topOnePurityHolds(sourceGate.accepted) : true,
top3_pure: domainCard ? topThreePurityHolds(sourceGate.accepted) : true,
strict_forbidden_mode: strictForbidden,
strict_forbidden_fallback_source: sourceStrictFallbackUsed
};
return {
status: items.length > 0 ? "ok" : "empty",
result_type: "ranking",
items,
summary: {
checked_records: source.length,
checked_records: sourcePool.length,
ranked_entities: items.length,
query_subject: semanticProfile.query_subject,
semantic_profile: semanticProfile,
ranking_basis: semanticProfile.ranking_basis
ranking_basis: semanticProfile.ranking_basis,
domain_purity_guard: puritySummary
},
evidence: items.slice(0, 5).map((item) => ({
entity: item.entity,
records_count: item.records_count
})),
why_included: items.length > 0 ? ["Показаны сущности с максимальным количеством записей."] : [],
selection_reason: ["Ранжирование выполнено по records_count по убыванию."],
why_included: items.length > 0
? [
"Показаны сущности с максимальным количеством записей.",
domainCard ? `P0 domain purity enforced for ${domainCard.id}.` : "P0 domain purity was not enforced."
]
: [],
selection_reason: [
"Ранжирование выполнено по records_count по убыванию.",
domainCard ? `Domain gate source scope: ${sourceScope.join(", ")}.` : "Domain gate source scope not applied."
],
risk_factors: uniqueStrings(["entity_volume_spike", ...semanticProfile.anomaly_patterns]),
business_interpretation: [
"Top entities by volume highlight where lifecycle-focused review should start first."
],
confidence: "medium",
limitations: ["Ранжирование по объему не всегда эквивалентно бизнес-риску."],
limitations: [
"Ранжирование по объему не всегда эквивалентно бизнес-риску.",
domainCard ? "Domain purity guardrail может исключить cross-domain записи на batch-слое." : "Domain purity guardrail не применялся."
],
errors: []
};
}

21
llm_normalizer/backend/src/services/questionTypeResolver.ts
View File

@ -78,12 +78,33 @@ function countRuleHits(text: string, rule: QuestionTypeRule): number {
return hits;
}
function hasProofIntent(text: string): boolean {
return /(?:\bprove\b|\bguess\b|доказан|доказано|доказуем|гипотез|догад|связан\s+ли|зач[её]л(?:ся|ось)\s+ли)/iu.test(text);
}
function hasExplicitChainSplitIntent(text: string): boolean {
return /(?:какие(?:\s+\S+){0,4}\s+цепочк[аи]|which\s+chains?|complete\s+vs\s+incomplete|что\s+закрыто.*что\s+нет)/iu.test(
text
);
}
function hasGroundingIntent(text: string): boolean {
return /(?:на\s+ч(?:е|ё)м[^?!.]{0,40}основан|чем\s+подтвержда|какие\s+основани|what\s+evidence|grounded\s+on|based\s+on)/iu.test(
text
);
}
export function resolveQuestionType(input: string): QuestionTypeClass {
const text = String(input ?? "").trim();
if (!text) {
return "unknown";
}
// Guard against collapsing proof-intent questions into chain classification.
if (hasProofIntent(text) && !hasExplicitChainSplitIntent(text) && !hasGroundingIntent(text)) {
return "prove_or_guess";
}
let bestType: QuestionTypeClass = "unknown";
let bestHits = 0;
let bestPriority = Number.POSITIVE_INFINITY;

431
llm_normalizer/backend/tests/assistantWave16ResidualCleanupRegression.test.ts Normal file
View File

@ -0,0 +1,431 @@
import fs from "fs";
import os from "os";
import path from "path";
import { afterEach, describe, expect, it } from "vitest";
import { AssistantDataLayer } from "../src/services/assistantDataLayer";
import { composeAssistantAnswer } from "../src/services/answerComposer";
import { resolveQuestionType } from "../src/services/questionTypeResolver";
import type { AnswerGroundingCheck, RequirementCoverageReport, UnifiedRetrievalResult } from "../src/types/assistant";
import type { ProblemUnit } from "../src/types/stage2ProblemUnits";
const TEMP_DIRS: string[] = [];
function cleanupTempDirs(): void {
for (const dir of TEMP_DIRS.splice(0)) {
fs.rmSync(dir, { recursive: true, force: true });
}
}
function createSnapshotRoot(records: Array<Record<string, unknown>>): string {
const root = fs.mkdtempSync(path.join(os.tmpdir(), "assistant-wave16-"));
TEMP_DIRS.push(root);
const payload = JSON.stringify({ records }, null, 2);
const write = (name: string, content = payload): void => {
fs.writeFileSync(path.resolve(root, name), content, "utf-8");
};
write("09_samples_key_fields_Recorder_Ref_Supplier_Buyer_Responsible.json");
write("03_snapshot_fragment_problem_cases.json");
write("07_samples_DocumentJournals.json");
write("08_samples_NDS_registers.json");
write("04_samples_SpisanieSRaschetnogoScheta.json");
write("05_samples_RealizaciyaTovarovUslug.json");
write("06_samples_PostuplenieTovarovUslug.json");
return root;
}
function buildMixedDomainRecord(id: string, description: string, account?: string): Record<string, unknown> {
return {
source_entity: "Document",
source_id: id,
display_name: description,
unknown_link_count: 0,
attributes: {
Description: description,
Account: account ?? "",
Period: "2020-07-15T00:00:00",
Amount: 276873.6
},
links: [
{
relation: "document_has_counterparty",
target_entity: "Counterparty",
target_id: `CP-${id}`,
source_field: "Counterparty"
}
]
};
}
function buildRouteSummary() {
return {
mode: "deterministic_v2" as const,
message_in_scope: true,
scope_confidence: "high" as const,
planner: {
total_fragments: 1,
in_scope_fragments: 1,
out_of_scope_fragments: 0,
discarded_fragments: 0,
contains_multiple_tasks: false
},
decisions: [],
fallback: {
type: "none" as const,
message: null
}
};
}
function buildCoverage(input?: Partial<RequirementCoverageReport>): RequirementCoverageReport {
return {
requirements_total: 1,
requirements_covered: 1,
requirements_uncovered: [],
requirements_partially_covered: [],
clarification_needed_for: [],
out_of_scope_requirements: [],
...input
};
}
function buildGrounding(input?: Partial<AnswerGroundingCheck>): AnswerGroundingCheck {
return {
status: "partial",
route_subject_match: true,
missing_requirements: [],
reasons: [],
why_included_summary: ["wave16-test"],
selection_reason_summary: ["wave16-test"],
...input
};
}
function buildProblemUnit(input: {
id: string;
type: ProblemUnit["problem_unit_type"];
account: string;
defect: string;
lifecycleDomain: ProblemUnit["lifecycle_domain"];
}): ProblemUnit {
return {
schema_version: "problem_unit_v0_1",
problem_unit_id: input.id,
problem_unit_type: input.type,
title: "Wave16 problem unit",
mechanism_summary: `Mechanism candidate: ${input.defect}.`,
business_defect_class: input.defect,
severity: {
score: 0.73,
grade: "high"
},
confidence: {
score: 0.62,
grade: "medium"
},
lifecycle_domain: input.lifecycleDomain,
affected_entities: ["Document:DOC-1"],
affected_documents: ["Document:DOC-1"],
affected_postings: ["Posting:POST-1"],
affected_accounts: [input.account],
affected_counterparties: ["Counterparty:CP-1"],
affected_contracts: ["Contract:CTR-1"],
failed_expected_edge: input.defect,
period_impact: {
is_period_sensitive: true,
impact_class: "close_risk"
},
evidence_pack: ["cand-1"],
entity_backlinks: [{ entity: "Document", id: "DOC-1" }],
snapshot_limitations: []
};
}
function buildRetrieval(input: {
status: UnifiedRetrievalResult["status"];
accountScope?: string[];
domainScope?: string[];
relationPatterns?: string[];
units?: ProblemUnit[];
}): UnifiedRetrievalResult {
const accountScope = input.accountScope ?? [];
const domainScope = input.domainScope ?? ["settlements"];
const relationPatterns = input.relationPatterns ?? ["payment_to_settlement"];
const units = input.units ?? [];
return {
fragment_id: "F1",
requirement_ids: ["R1"],
route: "hybrid_store_plus_live",
status: input.status,
result_type: "chain",
items:
input.status === "empty"
? []
: [
{
source_entity: "Document",
source_id: "DOC-1",
display_name: "Счет №4 от 07.07.20",
account_context: accountScope,
graph_domain_scope: domainScope,
relation_pattern_hits: relationPatterns,
period: "2020-07",
amount: "276 873,60"
}
],
summary: {
semantic_profile: {
account_scope: accountScope,
domain_scope: domainScope,
relation_patterns: relationPatterns,
period_scope: {
from: "2020-07-01",
to: "2020-07-31",
granularity: "month"
}
},
domain_purity_guard: {
domain_card_id: domainScope.includes("vat") ? "vat_document_register_book" : "settlements_60_62"
},
broad_query_detected: false,
broad_result_flag: false,
minimum_evidence_failed: false,
narrowing_strength: "strong",
degraded_to: null
},
evidence:
input.status === "empty"
? []
: [
{
evidence_id: "ev-R1",
claim_ref: "requirement:R1",
source_type: "retrieval_item",
source_ref: {
schema_version: "evidence_source_ref_v1",
namespace: "snapshot_2020_07",
entity: "document",
id: "DOC-1",
period: "2020-07",
canonical_ref: "evidence_source_ref_v1|snapshot_2020_07|document|doc-1|2020-07"
},
pointer: {
fragment_id: "F1",
route: "hybrid_store_plus_live",
source: {
namespace: "snapshot_2020_07",
entity: "document",
id: "DOC-1",
period: "2020-07"
},
locator: {
field_path: "risk_score",
item_index: 0
}
},
evidence_kind: "mechanism_link",
mechanism_note: relationPatterns[0],
confidence: "medium",
limitation: null,
payload: {
contract: "договор № 01/19-ПТ"
}
}
],
candidate_evidence: [],
problem_units: units,
problem_unit_summary:
units.length > 0
? {
schema_version: "problem_unit_summary_v0_1",
units_total: units.length,
duplicate_collapses: 0,
unit_types: units.map((unit) => unit.problem_unit_type),
type_distribution: {
[units[0]?.problem_unit_type ?? "broken_chain_segment"]: units.length
},
severity_distribution: {
low: 0,
medium: 0,
high: units.length
},
confidence_distribution: {
low: 0,
medium: units.length,
high: 0
},
primary_unit_type: units[0]?.problem_unit_type ?? null
}
: null,
why_included: ["wave16-test"],
selection_reason: ["wave16-test"],
risk_factors: ["wave16"],
business_interpretation: ["wave16"],
confidence: input.status === "ok" ? "medium" : "low",
limitations: [],
errors: []
};
}
afterEach(() => {
cleanupTempDirs();
});
describe("wave16 residual fail cleanup", () => {
it("q10-style VAT query must not derive account scope from percent/amount fragments", () => {
const dataLayer = new AssistantDataLayer(
createSnapshotRoot([
buildMixedDomainRecord("SET-1", "Оплата покупателя по договору", "62.01"),
buildMixedDomainRecord("VAT-1", "Счет-фактура и НДС в книге продаж", "68.02")
])
);
const result = dataLayer.executeRoute(
"hybrid_store_plus_live",
"По оплате от 13 июля на 276 873,60 указан НДС 20% = 46 145,60. Докажи отражение НДС."
);
expect(result.status).toBe("ok");
const summary = result.summary as Record<string, unknown>;
const semanticProfile = (summary.semantic_profile ?? {}) as Record<string, unknown>;
const accountScope = Array.isArray(semanticProfile.account_scope) ? semanticProfile.account_scope : [];
expect(accountScope).not.toContain("20");
expect(accountScope).not.toContain("60");
const domainGuard = (summary.domain_purity_guard ?? {}) as Record<string, unknown>;
expect(domainGuard.domain_card_id).toBe("vat_document_register_book");
});
it("q13-style broad VAT query on batch route must stay VAT-domain", () => {
const dataLayer = new AssistantDataLayer(
createSnapshotRoot([
buildMixedDomainRecord("SET-2", "Платеж и расчеты с покупателем", "62.02"),
buildMixedDomainRecord("VAT-2", "Полученный счет-фактура и налоговый эффект", "19.03")
])
);
const result = dataLayer.executeRoute(
"batch_refresh_then_store",
"Есть ли в июльском срезе покупки, где есть товар/услуга, но не видно счета-фактуры или налогового эффекта?"
);
const summary = result.summary as Record<string, unknown>;
const domainGuard = (summary.domain_purity_guard ?? {}) as Record<string, unknown>;
expect(domainGuard.domain_card_id).toBe("vat_document_register_book");
});
it("q06 first-check should be settlement-specific for prove_or_guess without explicit account tokens", () => {
const settlementUnit = buildProblemUnit({
id: "pu-settlement-wave16",
type: "broken_chain_segment",
account: "62.02",
defect: "failed_edge:payment_to_settlement",
lifecycleDomain: "customer_settlement"
});
const output = composeAssistantAnswer({
userMessage:
"Есть ли в июльском срезе ситуация, где деньги уже пришли, но закрытие расчётов не подтверждено тем документом?",
routeSummary: buildRouteSummary(),
retrievalResults: [buildRetrieval({ status: "ok", units: [settlementUnit], domainScope: ["settlements"] })],
requirements: [
{
requirement_id: "R1",
source_fragment_id: "F1",
requirement_text: "Wave16 requirement",
subject_tokens: [],
status: "covered",
route: "hybrid_store_plus_live"
}
],
coverageReport: buildCoverage(),
groundingCheck: buildGrounding(),
focusDomainHint: null,
questionTypeHint: "prove_or_guess",
enableAnswerPolicyV11: true,
enableProblemCentricAnswerV1: true,
enableLifecycleAnswerV1: true
});
expect(output.assistant_reply).toMatch(/регистр расчет|60\/62\/76|договор|объект расчет/i);
});
it("q12-like question should resolve prove_or_guess instead of chain classification", () => {
const resolved = resolveQuestionType(
"Связан ли полученный 31 июля счёт-фактура с услугой так, чтобы вычет был корректен, и это доказано или только предположение?"
);
expect(resolved).toBe("prove_or_guess");
});
it("why_breaks short line should avoid generic collapse and keep domain-specific mechanism", () => {
const settlementUnit = buildProblemUnit({
id: "pu-settlement-wave16-2",
type: "broken_chain_segment",
account: "62.01",
defect: "failed_edge:payment_to_settlement",
lifecycleDomain: "customer_settlement"
});
const output = composeAssistantAnswer({
userMessage: "Почему по расчетам долг остался после оплаты?",
routeSummary: buildRouteSummary(),
retrievalResults: [buildRetrieval({ status: "ok", units: [settlementUnit], domainScope: ["settlements"] })],
requirements: [
{
requirement_id: "R1",
source_fragment_id: "F1",
requirement_text: "Wave16 requirement",
subject_tokens: [],
status: "covered",
route: "hybrid_store_plus_live"
}
],
coverageReport: buildCoverage({
requirements_covered: 0,
requirements_partially_covered: ["R1"]
}),
groundingCheck: buildGrounding({
status: "partial",
reasons: ["Mechanism is unresolved for part of the evidence."]
}),
focusDomainHint: "settlements_60_62",
questionTypeHint: "why_breaks",
enableAnswerPolicyV11: true,
enableProblemCentricAnswerV1: true,
enableLifecycleAnswerV1: true
});
expect(output.assistant_reply).toMatch(/наиболее вероятн/i);
expect(output.assistant_reply).not.toContain("Коротко: Проблема с закрытием расчета подтверждается частично.");
});
it("VAT why_breaks first-check must include VAT-specific checks", () => {
const vatUnit = buildProblemUnit({
id: "pu-vat-wave16",
type: "cross_branch_inconsistency_cluster",
account: "68.02",
defect: "invoice_to_vat",
lifecycleDomain: "vat_flow"
});
const output = composeAssistantAnswer({
userMessage: "Почему по НДС не видно ожидаемого налогового эффекта?",
routeSummary: buildRouteSummary(),
retrievalResults: [buildRetrieval({ status: "ok", units: [vatUnit], domainScope: ["vat", "taxes"] })],
requirements: [
{
requirement_id: "R1",
source_fragment_id: "F1",
requirement_text: "Wave16 requirement",
subject_tokens: [],
status: "covered",
route: "hybrid_store_plus_live"
}
],
coverageReport: buildCoverage(),
groundingCheck: buildGrounding(),
focusDomainHint: "vat_document_register_book",
questionTypeHint: "why_breaks",
enableAnswerPolicyV11: true,
enableProblemCentricAnswerV1: true,
enableLifecycleAnswerV1: true
});
expect(output.assistant_reply).toMatch(/счет-?фактур|регистр НДС|19\/68|книг/i);
});
});

BIN
llm_normalizer/docs/runs/2026-03-28_Stage_04_Wave_16_Residual_FAIL_Cleanup_Generic_Answer_Squeeze.zip Normal file
View File

Binary file not shown.