@stll/anonymize 1.0.4 → 1.0.5

package/LICENSE

@@ -1,6 +1,6 @@
 MIT License
 
-Copyright (c) 2026 Stella
+Copyright (c) 2026 stella labs, s.r.o.
 
 Permission is hereby granted, free of charge, to any
 person obtaining a copy of this software and associated

package/README.md

@@ -1,79 +1,75 @@
 <p align="center">
-  <img src=".github/assets/banner.png" alt="Stella" width="100%" />
+  <img src="../../.github/assets/banner.png" alt="Stella anonymize" width="100%" />
 </p>
 
 # @stll/anonymize
 
-Multi-layer PII detection and anonymization pipeline.
-Regex, NER, deny lists, coreference resolution, and
-legal form detection across 20+ languages.
-
-## Features
-
-- **Regex detector** — IBAN, email, phone, credit card,
-  Czech birth numbers, dates (22 languages), company IDs
-- **Trigger phrases** — Czech, German, English, French,
-  Spanish, Italian, Polish, Hungarian, Romanian, Swedish
-- **Name corpus** — first names, surnames, titles with
-  Czech/Slovak declension handling
-- **Legal form detection** — s.r.o., GmbH, Ltd., S.A.,
-  and 1000+ forms across 20+ countries
-- **GLiNER NER** — zero-shot named entity recognition
-- **Deny-list gazetteer** — workspace-scoped Aho-Corasick
-  + fuzzy matching
-- **Coreference** — tracks "dále jen" / "hereinafter"
-  aliases with Czech declension variants
-- **Confidence boosting** — context-aware score adjustment
-- **False positive filtering** — template placeholders,
-  section numbers, generic roles
-- **Operators** — replace (reversible) and redact
-- **De-anonymization** — reverse replacements with key
+Runtime package for multi-layer PII detection and anonymization.
+
+It combines regex detectors, trigger phrases, deny-list matching, coreference handling, and NER into a single pipeline that works in native Node.js and in browser builds through the WASM entrypoint.
 
 ## Install
 
 ```bash
-npm install @stll/anonymize
-# Optional: install data package for deny lists
-npm install @stll/anonymize-data
+bun add @stll/anonymize
+# Optional data bundle for deny lists and dictionaries
+bun add @stll/anonymize-data
 ```
 
-## Quick Start
+For browser targets, install `@stll/anonymize-wasm` instead. It exposes the same runtime API through WebAssembly and is the supported entrypoint for Vite-based bundles.
+
+## Usage
 
-```typescript
-import { runPipeline } from '@stll/anonymize'
+```ts
+import { runPipeline } from "@stll/anonymize";
 
 const entities = await runPipeline({
   fullText: text,
   config: {
-    labels: ['person', 'organization', 'address',
-             'date', 'iban', 'phone number'],
+    labels: [
+      "person",
+      "organization",
+      "address",
+      "date",
+      "iban",
+      "phone number",
+    ],
     threshold: 0.5,
     enableRegex: true,
     enableTriggerPhrases: true,
     enableLegalForms: true,
     enableNameCorpus: true,
+    enableDenyList: false,
+    enableGazetteer: false,
+    enableNer: false,
+    enableConfidenceBoost: true,
+    enableCoreference: true,
+    workspaceId: "default",
   },
   gazetteerEntries: [],
-})
+});
 ```
 
-## Architecture
+## Browser setup
 
-```
-text → [regex]       → entities₁ ─┐
-text → [triggers]    → entities₂ ─┤
-text → [legal forms] → entities₃ ─┤
-text → [name corpus] → entities₄ ─┼→ merge → coref → boost → filter → result
-text → [gazetteer]   → entities₅ ─┤
-text → [GLiNER NER]  → entities₆ ─┘
+If you use Vite with the WASM build, exclude the bundle from dependency pre-bundling:
+
+```ts
+import stllWasm from "@stll/anonymize-wasm/vite";
+
+export default {
+  plugins: [stllWasm()],
+};
 ```
 
-## Built on
+## Notes
 
-- [@stll/text-search](https://github.com/stella/text-search) — multi-engine search orchestrator
-- [@stll/stdnum](https://github.com/stella/stdnum) — identifier validation (IBAN, IČO, RČ)
-- [@stll/anonymize-data](https://github.com/stella/anonymize) — deny-list dictionaries
+- `labels: []` disables deterministic label filtering; when NER is enabled it falls back to the default label set.
+- `enableNameCorpus` also controls whether first names, surnames, and titles are injected into deny-list matching when `enableDenyList` is enabled.
+- The optional `@stll/anonymize-data` package carries the published dictionary and trigger data used by the deny-list layer.
 
-## License
+## Built on
 
-MIT
+- `@stll/text-search`
+- `@stll/stdnum`
+- `@stll/anonymize-data`

package/dist/index.d.mts

@@ -178,6 +178,19 @@ type PipelineConfig = {
   threshold: number;
   enableTriggerPhrases: boolean;
   enableRegex: boolean;
+  /**
+   * Enables legal-form organization detection.
+   * Required for typed callers; legacy untyped
+   * callers that omit this field are treated as
+   * enabled at runtime for backward compatibility.
+   */
+  enableLegalForms: boolean;
+  /**
+   * Enables first-name/surname/title corpus matching.
+   * When deny-list mode is enabled, this also controls
+   * whether name-corpus entries are injected into the
+   * deny-list search automaton.
+   */
   enableNameCorpus: boolean;
   enableDenyList: boolean;
   denyListCountries?: string[];
@@ -189,6 +202,11 @@ type PipelineConfig = {
   enableCoreference: boolean;
   enableZoneClassification?: boolean;
   enableHotwordRules?: boolean;
+  /**
+   * Requested output labels. An empty array means
+   * "do not filter by label" for deterministic
+   * detectors; NER falls back to DEFAULT_ENTITY_LABELS.
+   */
   labels: string[];
   workspaceId: string;
 };
@@ -243,7 +261,7 @@ declare const CURRENCY_PATTERN_META: Readonly<RegexMeta>;
 declare const processRegexMatches: (allMatches: Match[], sliceStart: number, sliceEnd: number, meta_: readonly RegexMeta[]) => Entity[];
 //#endregion
 //#region src/detectors/deny-list.d.ts
-type DenyListConfig = Pick<PipelineConfig, "enableDenyList" | "denyListCountries" | "denyListRegions" | "denyListExcludeCategories">;
+type DenyListConfig = Pick<PipelineConfig, "enableDenyList" | "enableNameCorpus" | "denyListCountries" | "denyListRegions" | "denyListExcludeCategories">;
 /**
  * Source tag for each pattern in the automaton.
  * "deny-list" = standard deny list entry
@@ -707,7 +725,7 @@ declare const resolveCountries: (regions?: string[], countries?: string[]) => Se
  * Runs as a post-processing step after all detection
  * layers have merged.
  */
-declare const filterFalsePositives: (entities: Entity[], ctx?: PipelineContext) => Entity[];
+declare const filterFalsePositives: (entities: Entity[], ctx?: PipelineContext, fullText?: string) => Entity[];
 //#endregion
 //#region src/filters/confidence-boost.d.ts
 /**