marek/MtoRagSystem
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

harden history find tokens and shops earch

Browse Source
This commit is contained in:
team 1
2026-04-26 19:59:31 +02:00
parent ff273ff9a0
commit 308e980b4a
11 changed files with 1230 additions and 75 deletions
Download Patch File Download Diff File Expand all files Collapse all files

504
src/Agent/AgentRunner.php
View File

@@ -121,84 +121,108 @@ final readonly class AgentRunner
$commerceHistoryContext
);
$shopSearchQuery = $optimizedShopQuery !== '' ? $optimizedShopQuery : $prompt;
$shopQueryPreview = $this->shopSearchService->buildSearchQueryPreview(
$shopSearchQuery,
$commerceIntent,
$commerceHistoryContext
$shopSearchQuery = $this->resolveShopSearchQuery(
prompt: $prompt,
optimizedShopQuery: $optimizedShopQuery,
commerceHistoryContext: $commerceHistoryContext,
userId: $userId
);
yield $this->systemMsg(
$this->buildShopSearchMetaMessage(
query: $shopQueryPreview->searchText !== '' ? $shopQueryPreview->searchText : $shopSearchQuery,
commerceIntent: $commerceIntent,
usedOptimizedQuery: $optimizedShopQuery !== '',
originalQuery: $shopSearchQuery
),
'meta'
);
$this->agentLogger->info('Commerce search prepared', [
'userId' => $userId,
'commerceIntent' => $commerceIntent,
'usedOptimizedShopQuery' => $optimizedShopQuery !== '',
'optimizedShopQuery' => $optimizedShopQuery,
'shopSearchQuery' => $shopSearchQuery,
'hasCommerceHistoryContext' => $commerceHistoryContext !== '',
'commerceHistoryContextLength' => mb_strlen($commerceHistoryContext),
]);
yield $this->systemMsg(
sprintf($this->agentRunnerConfig->getFetchSearchDataMessageTemplate(), $commerceIntent),
'think'
);
$primaryShopResults = $this->searchShop(
$shopSearchQuery,
$commerceIntent,
$userId,
$commerceHistoryContext
);
$primaryShopSearchHadSystemFailure = $this->shopSearchService->hadLastSearchSystemFailure();
$primaryShopSearchFailureReason = $this->shopSearchService->getLastSearchFailureReason();
if ($primaryShopSearchHadSystemFailure) {
$this->agentLogger->warning('Shop repair skipped after Store API system failure', [
if ($shopSearchQuery === '') {
$this->agentLogger->info('Commerce search skipped because no concrete shop query could be resolved', [
'userId' => $userId,
'commerceIntent' => $commerceIntent,
'shopSearchQuery' => $shopSearchQuery,
'failureReason' => $primaryShopSearchFailureReason,
'prompt' => $prompt,
'optimizedShopQuery' => $optimizedShopQuery,
'hasCommerceHistoryContext' => $commerceHistoryContext !== '',
'commerceHistoryContextLength' => mb_strlen($commerceHistoryContext),
]);
$shopUnavailableMessage = $this->buildShopUnavailableMessage($primaryShopSearchFailureReason);
yield $this->systemMsg(
$shopUnavailableMessage,
'err'
);
$historyNotices[] = $this->buildHistoryNotice(
'Shopdaten konnten nicht geladen werden',
$primaryShopSearchFailureReason
$this->agentRunnerConfig->getNoConcreteShopQueryMessage(),
'info'
);
$repairPayload = [
'results' => $primaryShopResults,
'attemptedRepair' => false,
'usedRepair' => false,
'repairQueries' => [],
];
return;
} else {
yield $this->systemMsg('Erweiterte Shopsuche wird geprüft…', 'think');
$repairPayload = $this->repairShopResults(
prompt: $prompt,
userId: $userId,
commerceIntent: $commerceIntent,
commerceHistoryContext: $commerceHistoryContext,
primaryQuery: $shopSearchQuery,
primaryShopResults: $primaryShopResults,
knowledgeChunks: $knowledgeChunks
$shopQueryPreview = $this->shopSearchService->buildSearchQueryPreview(
$shopSearchQuery,
$commerceIntent,
$commerceHistoryContext
);
yield $this->systemMsg(
$this->buildShopSearchMetaMessage(
query: $shopQueryPreview->searchText !== '' ? $shopQueryPreview->searchText : $shopSearchQuery,
commerceIntent: $commerceIntent,
usedOptimizedQuery: $optimizedShopQuery !== '',
originalQuery: $shopSearchQuery
),
'meta'
);
$this->agentLogger->info('Commerce search prepared', [
'userId' => $userId,
'commerceIntent' => $commerceIntent,
'usedOptimizedShopQuery' => $optimizedShopQuery !== '',
'optimizedShopQuery' => $optimizedShopQuery,
'shopSearchQuery' => $shopSearchQuery,
'hasCommerceHistoryContext' => $commerceHistoryContext !== '',
'commerceHistoryContextLength' => mb_strlen($commerceHistoryContext),
]);
yield $this->systemMsg(
sprintf($this->agentRunnerConfig->getFetchSearchDataMessageTemplate(), $commerceIntent),
'think'
);
$primaryShopResults = $this->searchShop(
$shopSearchQuery,
$commerceIntent,
$userId,
$commerceHistoryContext
);
$primaryShopSearchHadSystemFailure = $this->shopSearchService->hadLastSearchSystemFailure();
$primaryShopSearchFailureReason = $this->shopSearchService->getLastSearchFailureReason();
if ($primaryShopSearchHadSystemFailure) {
$this->agentLogger->warning('Shop repair skipped after Store API system failure', [
'userId' => $userId,
'commerceIntent' => $commerceIntent,
'shopSearchQuery' => $shopSearchQuery,
'failureReason' => $primaryShopSearchFailureReason,
]);
$shopUnavailableMessage = $this->buildShopUnavailableMessage($primaryShopSearchFailureReason);
yield $this->systemMsg(
$shopUnavailableMessage,
'err'
);
$historyNotices[] = $this->buildHistoryNotice(
'Shopdaten konnten nicht geladen werden',
$primaryShopSearchFailureReason
);
$repairPayload = [
'results' => $primaryShopResults,
'attemptedRepair' => false,
'usedRepair' => false,
'repairQueries' => [],
];
} else {
yield $this->systemMsg('Erweiterte Shopsuche wird geprüft…', 'think');
$repairPayload = $this->repairShopResults(
prompt: $prompt,
userId: $userId,
commerceIntent: $commerceIntent,
commerceHistoryContext: $commerceHistoryContext,
primaryQuery: $shopSearchQuery,
primaryShopResults: $primaryShopResults,
knowledgeChunks: $knowledgeChunks
);
}
}
$shopResults = $repairPayload['results'];
@@ -645,7 +669,7 @@ final readonly class AgentRunner
return '';
}
return $this->sanitizeOptimizedShopQuery($optimizedQuery);
return $this->sanitizeOptimizedShopQuery($optimizedQuery, $prompt, $commerceHistoryContext);
}
/**
@@ -692,6 +716,189 @@ final readonly class AgentRunner
}
}
private function resolveShopSearchQuery(
string $prompt,
string $optimizedShopQuery,
string $commerceHistoryContext,
string $userId
): string {
if ($optimizedShopQuery !== '' && !$this->isMetaOnlyShopQuery($optimizedShopQuery)) {
return $optimizedShopQuery;
}
if (!$this->isMetaOnlyShopQuery($prompt)) {
return $prompt;
}
$contextQuery = $this->extractContextualShopSearchQuery($commerceHistoryContext);
if ($contextQuery !== '' && !$this->isMetaOnlyShopQuery($contextQuery)) {
return $contextQuery;
}
$extendedHistoryBudget = $this->agentRunnerConfig->getShopQueryContextFallbackHistoryBudgetChars();
if ($extendedHistoryBudget > mb_strlen($commerceHistoryContext, 'UTF-8')) {
$extendedHistory = $this->contextService->buildUserContextWithinBudget($userId, $extendedHistoryBudget);
$extendedContextQuery = $this->extractContextualShopSearchQuery($extendedHistory);
if ($extendedContextQuery !== '' && !$this->isMetaOnlyShopQuery($extendedContextQuery)) {
return $extendedContextQuery;
}
}
if ($this->agentRunnerConfig->shouldUseFullHistoryForShopQueryContextFallback()) {
$fullHistory = $this->contextService->buildUserContext($userId, true);
$fullHistoryContextQuery = $this->extractContextualShopSearchQuery($fullHistory);
if ($fullHistoryContextQuery !== '' && !$this->isMetaOnlyShopQuery($fullHistoryContextQuery)) {
return $fullHistoryContextQuery;
}
}
return '';
}
private function extractContextualShopSearchQuery(string $commerceHistoryContext): string
{
if (!$this->agentRunnerConfig->isShopQueryContextFallbackEnabled()) {
return '';
}
$questions = $this->extractRecentUserQuestions(
$commerceHistoryContext,
$this->agentRunnerConfig->getShopQueryContextFallbackQuestionLimit()
);
for ($i = count($questions) - 1; $i >= 0; $i--) {
$question = trim($questions[$i]);
if ($question === '' || $this->isMetaOnlyShopQuery($question)) {
continue;
}
$contextQuery = $this->buildContextFallbackShopQuery($question);
if ($contextQuery !== '' && !$this->isMetaOnlyShopQuery($contextQuery)) {
return $contextQuery;
}
}
return '';
}
private function buildContextFallbackShopQuery(string $question): string
{
$tokens = $this->tokenizeShopQueryCandidate($question);
if ($tokens === []) {
return '';
}
$filterTerms = [];
foreach (array_merge(
$this->agentRunnerConfig->getShopQueryMetaOnlyTerms(),
$this->agentRunnerConfig->getShopQueryContextFallbackFilterTerms()
) as $term) {
foreach ($this->tokenizeShopQueryCandidate($term) as $token) {
$filterTerms[$token] = true;
}
}
$maxTerms = max(1, $this->agentRunnerConfig->getShopQueryContextFallbackMaxTerms());
$out = [];
foreach ($tokens as $token) {
if (isset($filterTerms[$token])) {
continue;
}
if (in_array($token, $out, true)) {
continue;
}
$out[] = $token;
if (count($out) >= $maxTerms) {
break;
}
}
return implode(' ', $out);
}
/**
* @return string[]
*/
private function tokenizeShopQueryCandidate(string $value): array
{
$value = mb_strtolower(trim($value), 'UTF-8');
$value = str_replace(['-', '/', '_'], ' ', $value);
if (preg_match_all('/\d+(?:[,.]\d+)?|[\p{L}\p{N}]+/u', $value, $matches) !== 1) {
return [];
}
return array_values(array_filter(
array_map(static fn(string $token): string => trim($token), $matches[0] ?? []),
static fn(string $token): bool => $token !== ''
));
}
private function isMetaOnlyShopQuery(string $query): bool
{
if (!$this->agentRunnerConfig->isShopQueryMetaGuardEnabled()) {
return false;
}
$tokens = $this->tokenizeMetaGuardText($query);
if ($tokens === []) {
return true;
}
$metaTerms = [];
foreach ($this->agentRunnerConfig->getShopQueryMetaOnlyTerms() as $term) {
foreach ($this->tokenizeMetaGuardText($term) as $token) {
$metaTerms[$token] = true;
}
}
if ($metaTerms === []) {
return false;
}
foreach ($tokens as $token) {
if (!isset($metaTerms[$token])) {
return false;
}
}
return true;
}
/**
* @return string[]
*/
private function tokenizeMetaGuardText(string $value): array
{
$value = mb_strtolower(trim($value), 'UTF-8');
$value = str_replace(['-', '/', '_'], ' ', $value);
$value = preg_replace('/[^\p{L}\p{N}]+/u', ' ', $value) ?? $value;
$value = preg_replace('/\s+/u', ' ', $value) ?? $value;
$value = trim($value);
if ($value === '') {
return [];
}
return array_values(array_filter(
explode(' ', $value),
static fn(string $token): bool => $token !== ''
));
}
private function searchShop(
string $query,
string $commerceIntent,
@@ -743,8 +950,11 @@ final readonly class AgentRunner
};
}
private function sanitizeOptimizedShopQuery(string $query): string
{
private function sanitizeOptimizedShopQuery(
string $query,
string $sourcePrompt = '',
string $commerceHistoryContext = ''
): string {
$query = trim($query);
if ($query === '') {
@@ -755,10 +965,162 @@ final readonly class AgentRunner
$query = preg_replace($this->agentRunnerConfig->getOptimizedShopQueryPrefixPattern(), '', $query) ?? $query;
$query = trim($query, $this->agentRunnerConfig->getOptimizedShopQueryTrimCharacters());
$query = preg_replace('/\s+/u', ' ', $query) ?? $query;
$query = $this->preserveOptimizedShopQueryLanguage($query, $sourcePrompt);
$query = $this->enrichReferentialShopQueryFromHistory($query, $sourcePrompt, $commerceHistoryContext);
$query = preg_replace('/\s+/u', ' ', $query) ?? $query;
return trim($query);
}
private function enrichReferentialShopQueryFromHistory(
string $query,
string $sourcePrompt,
string $commerceHistoryContext
): string {
if (!$this->agentRunnerConfig->isShopQueryContextAnchorEnrichmentEnabled()) {
return $query;
}
if (trim($commerceHistoryContext) === '') {
return $query;
}
$queryTokens = $this->tokenizeShopQueryCandidate($query);
if ($queryTokens === []) {
return $query;
}
$maxTerms = max(1, $this->agentRunnerConfig->getShopQueryContextAnchorEnrichmentMaxQueryTerms());
if (count($queryTokens) > $maxTerms) {
return $query;
}
if (!$this->containsConfiguredShopQueryAnchorTrigger(trim($query . ' ' . $sourcePrompt))) {
return $query;
}
$anchor = $this->normalizeShopQueryAnchor(
$this->extractLatestConfiguredShopQueryContextAnchor($commerceHistoryContext)
);
if ($anchor === '' || $this->queryAlreadyContainsAllAnchorTokens($query, $anchor)) {
return $query;
}
$template = $this->agentRunnerConfig->getShopQueryContextAnchorEnrichmentTemplate();
$enriched = str_replace(['{anchor}', '{query}'], [$anchor, $query], $template);
$enriched = preg_replace('/\s+/u', ' ', $enriched) ?? $enriched;
return trim($enriched) !== '' ? trim($enriched) : $query;
}
private function containsConfiguredShopQueryAnchorTrigger(string $text): bool
{
$tokens = $this->tokenizeShopQueryCandidate($text);
if ($tokens === []) {
return false;
}
$tokenSet = array_fill_keys($tokens, true);
foreach ($this->agentRunnerConfig->getShopQueryContextAnchorEnrichmentTriggerTerms() as $term) {
foreach ($this->tokenizeShopQueryCandidate($term) as $termToken) {
if (isset($tokenSet[$termToken])) {
return true;
}
}
}
return false;
}
private function extractLatestConfiguredShopQueryContextAnchor(string $commerceHistoryContext): string
{
$latest = '';
foreach ($this->agentRunnerConfig->getShopQueryContextAnchorEnrichmentPatterns() as $pattern) {
if (@preg_match_all($pattern, $commerceHistoryContext, $matches, PREG_SET_ORDER) === false) {
continue;
}
foreach ($matches as $match) {
$candidate = trim((string) ($match[0] ?? ''));
if ($candidate !== '') {
$latest = $candidate;
}
}
}
return $latest;
}
private function normalizeShopQueryAnchor(string $anchor): string
{
$anchor = str_replace('®', '', $anchor);
$anchor = mb_strtolower(trim($anchor), 'UTF-8');
$anchor = preg_replace('/[^\p{L}\p{N},.%°+\-\s]+/u', ' ', $anchor) ?? $anchor;
$anchor = preg_replace('/\s+/u', ' ', $anchor) ?? $anchor;
return trim($anchor);
}
private function queryAlreadyContainsAllAnchorTokens(string $query, string $anchor): bool
{
$queryTokens = array_fill_keys($this->tokenizeShopQueryCandidate($query), true);
foreach ($this->tokenizeShopQueryCandidate($anchor) as $token) {
if (!isset($queryTokens[$token])) {
return false;
}
}
return true;
}
private function preserveOptimizedShopQueryLanguage(string $query, string $sourcePrompt): string
{
if (!$this->agentRunnerConfig->isShopQueryLanguagePreservationEnabled()) {
return $query;
}
$language = $this->detectConfiguredShopQueryLanguage($sourcePrompt);
if ($language === null) {
return $query;
}
$replacements = $this->agentRunnerConfig->getShopQueryTranslationReplacements($language);
if ($replacements === []) {
return $query;
}
foreach ($replacements as $source => $target) {
$pattern = '/(?<![\\p{L}\\p{N}])' . preg_replace('/\\s+/u', '\\s+', preg_quote($source, '/')) . '(?![\\p{L}\\p{N}])/iu';
$query = preg_replace($pattern, $target, $query) ?? $query;
}
return $query;
}
private function detectConfiguredShopQueryLanguage(string $sourcePrompt): ?string
{
$normalized = ' ' . strtolower($sourcePrompt) . ' ';
$normalized = preg_replace('/[\\r\\n\\t]+/u', ' ', $normalized) ?? $normalized;
$normalized = preg_replace('/\\s+/u', ' ', $normalized) ?? $normalized;
foreach ($this->agentRunnerConfig->getShopQueryLanguageMarkers() as $language => $markers) {
foreach ($markers as $marker) {
if ($marker !== '' && str_contains($normalized, $marker)) {
return $language;
}
}
}
return null;
}
/**
* @return Generator<int, string, mixed, string>
*/
@@ -993,4 +1355,4 @@ final readonly class AgentRunner
default => $msg,
};
}
}
}