@stll/anonymize 0.0.9 → 1.0.1

package/dist/{index.d.ts → index.d.mts}

@@ -1,5 +1,5 @@
-import { Validator } from "@stll/stdnum";
 import { Match, PatternEntry, TextSearch } from "@stll/text-search";
+import { Validator } from "@stll/stdnum";
 import { Tokenizer } from "@huggingface/tokenizers";
 
 //#region src/types.d.ts
@@ -143,9 +143,7 @@ declare const OPERATOR_TYPES: readonly ["replace", "redact"];
 type OperatorType = (typeof OPERATOR_TYPES)[number];
 /** Per-label operator selection. Key is the entity label. */
 type OperatorConfig = {
-  /** Operator per label. Missing labels default to "replace". */
-  operators: Record<string, OperatorType>;
-  /** Custom replacement string for the redact operator. */
+  /** Operator per label. Missing labels default to "replace". */operators: Record<string, OperatorType>; /** Custom replacement string for the redact operator. */
   redactString: string;
 };
 /** Whether an operator produces a reversible redaction entry. */
@@ -168,8 +166,7 @@ type RedactionResult = {
    * Maps placeholder to original text. Only populated for
    * reversible operators (replace). Empty for redact.
    */
-  redactionMap: Map<string, string>;
-  /** Maps placeholder to the operator that produced it. */
+  redactionMap: Map<string, string>; /** Maps placeholder to the operator that produced it. */
   operatorMap: Map<string, OperatorType>;
   entityCount: number;
 };
@@ -209,8 +206,7 @@ declare const DEFAULT_ENTITY_LABELS: readonly ["person", "organization", "phone
 //#region src/detectors/regex.d.ts
 type RegexMeta = {
   label: string;
-  score: number;
-  /** Post-match stdnum validator for confirmation. */
+  score: number; /** Post-match stdnum validator for confirmation. */
   validator?: Validator;
 };
 /** Flat pattern array for text-search. */
@@ -269,10 +265,8 @@ type DenyListData = {
    * appears in multiple dictionaries (e.g., "Denver"
    * is both a person name and a city name).
    */
-  labels: string[][];
-  /** Maps pattern index → original pattern text. */
-  originals: string[];
-  /** Maps pattern index → source types (plural). */
+  labels: string[][]; /** Maps pattern index → original pattern text. */
+  originals: string[]; /** Maps pattern index → source types (plural). */
   sources: PatternSource[][];
 };
 /**
@@ -317,8 +311,7 @@ type PatternSlice = {
   end: number;
 };
 type GazetteerData = {
-  /** Maps local pattern index to entry label. */
-  labels: string[];
+  /** Maps local pattern index to entry label. */labels: string[];
   /**
    * Whether each pattern is fuzzy (distance > 0).
    * Used by the post-processor to assign scores.
@@ -326,9 +319,7 @@ type GazetteerData = {
   isFuzzy: boolean[];
 };
 type UnifiedSearchInstance = {
-  /** Regex + triggers + legal-forms. */
-  tsRegex: TextSearch;
-  /** Deny-list + street-types + gazetteer. */
+  /** Regex + triggers + legal-forms. */tsRegex: TextSearch; /** Deny-list + street-types + gazetteer. */
   tsLiterals: TextSearch;
   slices: {
     regex: PatternSlice;
@@ -369,8 +360,7 @@ type NameCorpusData = {
   firstNames: ReadonlySet<string>;
   surnames: ReadonlySet<string>;
   titleTokens: ReadonlySet<string>;
-  excludedWords: ReadonlySet<string>;
-  /** Raw arrays exposed for deny-list AC integration. */
+  excludedWords: ReadonlySet<string>; /** Raw arrays exposed for deny-list AC integration. */
   firstNamesList: readonly string[];
   surnamesList: readonly string[];
   titlesList: readonly string[];
@@ -397,8 +387,7 @@ type PipelineContext = {
   allowList: ReadonlySet<string> | null;
   allowListPromise: Promise<ReadonlySet<string>> | null;
   personStopwords: ReadonlySet<string> | null;
-  personStopwordsPromise: Promise<ReadonlySet<string>> | null;
-  /** First-name exclusions for stopword filtering. */
+  personStopwordsPromise: Promise<ReadonlySet<string>> | null; /** First-name exclusions for stopword filtering. */
   firstNameExclusions: ReadonlySet<string> | null;
   firstNameExclusionCorpusLen: number;
   genericRoles: ReadonlySet<string> | null;
@@ -425,11 +414,6 @@ type PipelineContext = {
 };
 /** Create a fresh, empty pipeline context. */
 declare const createPipelineContext: () => PipelineContext;
-/**
- * Module-level default context. Used when callers
- * don't provide an explicit context, preserving full
- * backward compatibility with the existing API.
- */
 //#endregion
 //#region src/pipeline.d.ts
 /** Strip leading/trailing whitespace and punctuation. */
@@ -613,10 +597,8 @@ declare const processGazetteerMatches: (allMatches: Match[], sliceStart: number,
 //#region src/detectors/coreference.d.ts
 type DefinedTerm = {
   alias: string;
-  label: string;
-  /** Position of the definition in the source text */
-  definitionStart: number;
-  /** Original entity text the alias refers to */
+  label: string; /** Position of the definition in the source text */
+  definitionStart: number; /** Original entity text the alias refers to */
   sourceText: string;
 };
 declare const extractDefinedTerms: (fullText: string, entities: Entity[], ctx?: PipelineContext) => Promise<DefinedTerm[]>;
@@ -647,7 +629,9 @@ declare const propagateOrgNames: (entities: Entity[], fullText: string) => Entit
 //#endregion
 //#region src/detectors/names.d.ts
 /**
- * Load name corpus data from JSON config files.
+ * Load name corpus data from per-language dictionary
+ * files and legacy config files. Merges all sources.
+ *
  * Safe to call multiple times; only loads once per
  * context. Must be called before detectNameCorpus or
  * the getNameCorpus*() accessors are used.
@@ -673,8 +657,7 @@ declare const detectNameCorpus: (fullText: string, ctx?: PipelineContext) => Ent
 //#endregion
 //#region src/unified-search.d.ts
 type UnifiedResult = {
-  /** All matches from both instances combined. */
-  regexMatches: Match[];
+  /** All matches from both instances combined. */regexMatches: Match[];
   literalMatches: Match[];
 };
 declare const runUnifiedSearch: (instance: UnifiedSearchInstance, fullText: string) => UnifiedResult;
@@ -740,7 +723,6 @@ declare const filterFalsePositives: (entities: Entity[], ctx?: PipelineContext)
  * entities pass through unchanged.
  */
 declare const boostNearMissEntities: (entities: Entity[], threshold: number) => Entity[];
-/** Ensure preposition data is loaded. */
 //#endregion
 //#region src/filters/hotword-rules.d.ts
 type HotwordRule = {
@@ -849,8 +831,6 @@ declare const decodeTokenSpans: (batchSize: number, numWords: number, numEntitie
 //#region src/gliner/processor.d.ts
 /** Tokenize text into words with character offsets. */
 declare const tokenizeText: (text: string) => [words: string[], starts: number[], ends: number[]];
-/** Pad a 2D or 3D array to uniform inner length. */
-
 /** Prepare a complete batch for ONNX inference. */
 declare const prepareBatch: (tokenizer: Tokenizer, texts: string[], entities: string[], maxWidth: number) => {
   inputsIds: number[][];
@@ -936,4 +916,4 @@ declare const levenshtein: (rawA: string, rawB: string) => number;
 declare const normalizeForSearch: (text: string) => string;
 //#endregion
 export { type AnonymisationOperator, CURRENCY_PATTERN_META, type CountryCode, DATE_PATTERN_META, DEFAULT_ENTITY_LABELS, DEFAULT_OPERATOR_CONFIG, DETECTION_SOURCES, DETECTOR_PRIORITY, type DefinitionPattern, type DenyListCategory, type DenyListData, type DetectionSource, type DocumentZone, type Entity, type EntityResult, type GazetteerData, type GazetteerEntry, type HotwordRule, type NameCorpusData, type NerInferenceFn, OPERATOR_REGISTRY, OPERATOR_TYPES, type OperatorConfig, type OperatorType, type PipelineConfig, type PipelineContext, type PipelineOptions, REGEX_META, REGEX_PATTERNS, REGIONS, type RawInferenceResult, type RedactionResult, type RegexMeta, type RegionId, type ReviewDecision, type ReviewedEntity, type TriggerExtension, type TriggerGroupConfig, type TriggerRule, type TriggerStrategy, type TriggerValidation, type UnifiedResult, type UnifiedSearchInstance, ZONE_SCORE_ADJUSTMENTS, type ZoneSpan, applyHotwordRules, applyZoneAdjustments, boostNearMissEntities, buildDenyList, buildGazetteerPatterns, buildLegalFormPatterns, buildPlaceholderMap, buildStreetTypePatterns, buildTriggerPatterns, buildUnifiedSearch, chunkText, classifyZones, computeChunkOffsets, corefKey, createPipelineContext, deanonymise, decodeSpans, decodeTokenSpans, detectNameCorpus, ensureDenyListData, exportRedactionKey, extractDefinedTerms, filterFalsePositives, findCoreferenceSpans, getCurrencyPatterns, getDatePatterns, initHotwordRules, initNameCorpus, initZoneClassifier, levenshtein, mergeAndDedup, mergeChunkEntities, normalizeForSearch, prepareBatch, processAddressSeeds, processDenyListMatches, processGazetteerMatches, processLegalFormMatches, processRegexMatches, processTriggerMatches, propagateOrgNames, redactText, resolveCountries, resolveOperator, runPipeline, runUnifiedSearch, sanitizeEntities, tokenizeText };
-//# sourceMappingURL=index.d.ts.map
+//# sourceMappingURL=index.d.mts.map