@mailwoman/neural 2.0.6 → 2.2.0

package/out/browser.d.ts

@@ -0,0 +1,16 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Browser-safe re-export surface. Excludes `./onnx-runner.js` + `./weights.js` (Node-only — they
+ *   statically reference `onnxruntime-node` + `node:fs`), the dynamic `loadFromWeights` /
+ *   `loadFromFile` paths from those modules guard the corresponding imports with `webpackIgnore` so
+ *   Node callers still get them via the main `@mailwoman/neural` entry without bundling them into a
+ *   browser graph.
+ */
+export * from "./classifier.js";
+export * from "./labels.js";
+export * from "./tokenizer.js";
+export type { InferResult } from "./onnx-runner.js";
+//# sourceMappingURL=browser.d.ts.map

package/out/browser.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"browser.d.ts","sourceRoot":"","sources":["../browser.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAEH,cAAc,iBAAiB,CAAA;AAC/B,cAAc,aAAa,CAAA;AAC3B,cAAc,gBAAgB,CAAA;AAG9B,YAAY,EAAE,WAAW,EAAE,MAAM,kBAAkB,CAAA"}

package/out/browser.js

@@ -0,0 +1,15 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Browser-safe re-export surface. Excludes `./onnx-runner.js` + `./weights.js` (Node-only — they
+ *   statically reference `onnxruntime-node` + `node:fs`), the dynamic `loadFromWeights` /
+ *   `loadFromFile` paths from those modules guard the corresponding imports with `webpackIgnore` so
+ *   Node callers still get them via the main `@mailwoman/neural` entry without bundling them into a
+ *   browser graph.
+ */
+export * from "./classifier.js";
+export * from "./labels.js";
+export * from "./tokenizer.js";
+//# sourceMappingURL=browser.js.map

package/out/browser.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"browser.js","sourceRoot":"","sources":["../browser.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAEH,cAAc,iBAAiB,CAAA;AAC/B,cAAc,aAAa,CAAA;AAC3B,cAAc,gBAAgB,CAAA"}

package/out/classifier.d.ts

@@ -9,19 +9,55 @@
  *
  *   Convenience wrappers `parseJson` / `parseTuples` / `parseXml` project the tree on the way out.
  */
-import { type AddressTree, type ComponentTag, decodeAsXml } from "@mailwoman/core/decoder";
-import { OnnxRunner } from "./onnx-runner.js";
+import { decodeAsXml, type AddressTree, type ComponentTag } from "@mailwoman/core/decoder";
+import type { InferResult } from "./onnx-runner.js";
+import { type QueryShapeLike } from "./query-shape-prior.js";
 import { MailwomanTokenizer } from "./tokenizer.js";
-import { type ResolveWeightsOpts } from "./weights.js";
+import type { ResolveWeightsOpts } from "./weights.js";
+/**
+ * Structural type the classifier needs from a runner. Lets callers swap the Node-side `OnnxRunner`
+ * for a browser-side runner (e.g. `@mailwoman/neural-web`'s `WebOnnxRunner`) without inheritance —
+ * the classifier only ever calls `infer(ids)`.
+ */
+export interface NeuralRunner {
+    infer(tokenIds: number[]): Promise<InferResult>;
+}
 export interface NeuralAddressClassifierConfig {
     tokenizer: MailwomanTokenizer;
-    runner: OnnxRunner;
-    /** Label vocabulary in the order the model emits them. Defaults to Stage 1 (v0.1.0/v0.2.0). */
+    runner: NeuralRunner;
+    /**
+     * Label vocabulary in the order the model emits them. Defaults to Stage 2 (v0.3.0). Stage 2
+     * strictly extends Stage 1 at the same indices, so a v0.2.0 Stage 1 model loaded with this
+     * default still decodes correctly — its emissions only span the first 15 entries.
+     */
     labels?: readonly string[];
+    /**
+     * Decoding strategy:
+     *
+     * - `"viterbi"` (default) — linear-chain CRF Viterbi with the BIO structural mask. Prevents
+     *   orphan-`I-*` sequences. If `transitions` is provided, uses learned scores on top.
+     * - `"argmax"` — per-token argmax. Faster but produces structurally invalid sequences. Use only for
+     *   debugging / comparison.
+     */
+    decode?: "viterbi" | "argmax";
+    /**
+     * Optional learned CRF transition scores. Square matrix of size `labels.length × labels.length`.
+     * Added on top of the structural BIO mask. Future weights releases ship this; today's v3.0.0
+     * weights don't, so the structural mask alone is used.
+     */
+    transitions?: number[][];
+    /** Optional learned start-of-sequence transition scores per label. */
+    startTransitions?: number[];
+    /** Optional learned end-of-sequence transition scores per label. */
+    endTransitions?: number[];
 }
 export declare class NeuralAddressClassifier {
     private readonly cfg;
     private readonly labels;
+    private readonly decodeMode;
+    private readonly transitions;
+    private readonly startTransitions;
+    private readonly endTransitions;
     constructor(cfg: NeuralAddressClassifierConfig);
     /**
      * One-call factory that resolves the weights package (or explicit paths), loads the tokenizer and
@@ -29,12 +65,50 @@ export declare class NeuralAddressClassifier {
      *
      * Resolution order: explicit paths in `opts` → `@mailwoman/neural-weights-<locale>` package →
      * throws a single actionable error.
+     *
+     * **Node-only.** The dynamic imports keep `OnnxRunner` (onnxruntime-node) + `resolveWeights`
+     * (uses Node fs) out of the static dependency graph, so this file can be bundled for the browser
+     * by `@mailwoman/neural-web`. Calling this method in a browser will throw at runtime — use
+     * `loadNeuralClassifierFromUrls` from `@mailwoman/neural-web` instead.
      */
     static loadFromWeights(opts?: ResolveWeightsOpts): Promise<NeuralAddressClassifier>;
-    /** Tokenize → infer → argmax/softmax → decoder tree. */
-    parse(text: string): Promise<AddressTree>;
-    parseJson(text: string): Promise<Partial<Record<ComponentTag, string>>>;
-    parseTuples(text: string): Promise<Array<[ComponentTag, string]>>;
-    parseXml(text: string, opts?: Parameters<typeof decodeAsXml>[1]): Promise<string>;
+    /** Tokenize → infer → Viterbi (or argmax) → decoder tree. */
+    parse(text: string, opts?: ParseOpts): Promise<AddressTree>;
+    /**
+     * Like `parse`, but also returns the raw per-token logits and piece offsets needed for per-span
+     * logit aggregation (Option C joint-reconcile integration).
+     */
+    parseWithLogits(text: string, opts?: ParseOpts): Promise<ParseWithLogitsResult>;
+    parseJson(text: string, opts?: ParseOpts): Promise<Partial<Record<ComponentTag, string>>>;
+    parseTuples(text: string, opts?: ParseOpts): Promise<Array<[ComponentTag, string]>>;
+    parseXml(text: string, opts?: ParseOpts & {
+        xml?: Parameters<typeof decodeAsXml>[1];
+    }): Promise<string>;
+}
+/** Result of `parseWithLogits` — tree + raw material for per-span logit aggregation. */
+export interface ParseWithLogitsResult {
+    tree: AddressTree;
+    logits: number[][];
+    pieces: Array<{
+        start: number;
+        end: number;
+    }>;
+}
+/**
+ * Per-call opts for `parse()`. Threading a precomputed `QueryShape` here turns on the soft-prior
+ * bias path in the Viterbi decoder (Stage 2.4 boundary → Stage 3 encoder integration).
+ */
+export interface ParseOpts {
+    /**
+     * Precomputed `QueryShape` for this input (from `@mailwoman/query-shape`'s `computeQueryShape`).
+     * Known-format hits in the shape produce additive emission biases toward the matching BIO label.
+     * Typed structurally — no runtime dependency on `@mailwoman/query-shape`.
+     */
+    queryShape?: QueryShapeLike;
+    /**
+     * Maximum bias magnitude in log-odds units. Default 1.0 — adds up to ~e^1 ≈ 2.7× odds to the
+     * favored label. Confidence-scaled, so a 0.6-confidence format hit gets +0.6 max bias.
+     */
+    queryShapeBiasScale?: number;
 }
 //# sourceMappingURL=classifier.d.ts.map

package/out/classifier.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"classifier.d.ts","sourceRoot":"","sources":["../classifier.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAEH,OAAO,EACN,KAAK,WAAW,EAChB,KAAK,YAAY,EAKjB,WAAW,EACX,MAAM,yBAAyB,CAAA;AAEhC,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAA;AAC7C,OAAO,EAAE,kBAAkB,EAAE,MAAM,gBAAgB,CAAA;AACnD,OAAO,EAAE,KAAK,kBAAkB,EAAkB,MAAM,cAAc,CAAA;AAEtE,MAAM,WAAW,6BAA6B;IAC7C,SAAS,EAAE,kBAAkB,CAAA;IAC7B,MAAM,EAAE,UAAU,CAAA;IAClB,+FAA+F;IAC/F,MAAM,CAAC,EAAE,SAAS,MAAM,EAAE,CAAA;CAC1B;AAED,qBAAa,uBAAuB;IAGvB,OAAO,CAAC,QAAQ,CAAC,GAAG;IAFhC,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAmB;gBAEb,GAAG,EAAE,6BAA6B;IAI/D;;;;;;OAMG;WACU,eAAe,CAAC,IAAI,GAAE,kBAAuB,GAAG,OAAO,CAAC,uBAAuB,CAAC;IAS7F,wDAAwD;IAClD,KAAK,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,WAAW,CAAC;IAqBzC,SAAS,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC,MAAM,CAAC,YAAY,EAAE,MAAM,CAAC,CAAC,CAAC;IAIvE,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,YAAY,EAAE,MAAM,CAAC,CAAC,CAAC;IAIjE,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,UAAU,CAAC,OAAO,WAAW,CAAC,CAAC,CAAC,CAAC,GAAG,OAAO,CAAC,MAAM,CAAC;CAGvF"}
+{"version":3,"file":"classifier.d.ts","sourceRoot":"","sources":["../classifier.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAEH,OAAO,EAIN,WAAW,EACX,KAAK,WAAW,EAChB,KAAK,YAAY,EAEjB,MAAM,yBAAyB,CAAA;AAEhC,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,kBAAkB,CAAA;AACnD,OAAO,EAA0C,KAAK,cAAc,EAAE,MAAM,wBAAwB,CAAA;AACpG,OAAO,EAAE,kBAAkB,EAAE,MAAM,gBAAgB,CAAA;AAEnD,OAAO,KAAK,EAAE,kBAAkB,EAAmB,MAAM,cAAc,CAAA;AAEvE;;;;GAIG;AACH,MAAM,WAAW,YAAY;IAC5B,KAAK,CAAC,QAAQ,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC,WAAW,CAAC,CAAA;CAC/C;AAED,MAAM,WAAW,6BAA6B;IAC7C,SAAS,EAAE,kBAAkB,CAAA;IAC7B,MAAM,EAAE,YAAY,CAAA;IACpB;;;;OAIG;IACH,MAAM,CAAC,EAAE,SAAS,MAAM,EAAE,CAAA;IAC1B;;;;;;;OAOG;IACH,MAAM,CAAC,EAAE,SAAS,GAAG,QAAQ,CAAA;IAC7B;;;;OAIG;IACH,WAAW,CAAC,EAAE,MAAM,EAAE,EAAE,CAAA;IACxB,sEAAsE;IACtE,gBAAgB,CAAC,EAAE,MAAM,EAAE,CAAA;IAC3B,oEAAoE;IACpE,cAAc,CAAC,EAAE,MAAM,EAAE,CAAA;CACzB;AAED,qBAAa,uBAAuB;IAOvB,OAAO,CAAC,QAAQ,CAAC,GAAG;IANhC,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAmB;IAC1C,OAAO,CAAC,QAAQ,CAAC,UAAU,CAAsB;IACjD,OAAO,CAAC,QAAQ,CAAC,WAAW,CAAY;IACxC,OAAO,CAAC,QAAQ,CAAC,gBAAgB,CAAU;IAC3C,OAAO,CAAC,QAAQ,CAAC,cAAc,CAAU;gBAEZ,GAAG,EAAE,6BAA6B;IAa/D;;;;;;;;;;;OAWG;WACU,eAAe,CAAC,IAAI,GAAE,kBAAuB,GAAG,OAAO,CAAC,uBAAuB,CAAC;IAuB7F,6DAA6D;IACvD,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,SAAS,GAAG,OAAO,CAAC,WAAW,CAAC;IA8CjE;;;OAGG;IACG,eAAe,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,SAAS,GAAG,OAAO,CAAC,qBAAqB,CAAC;IA6C/E,SAAS,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,SAAS,GAAG,OAAO,CAAC,OAAO,CAAC,MAAM,CAAC,YAAY,EAAE,MAAM,CAAC,CAAC,CAAC;IAIzF,WAAW,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,SAAS,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,YAAY,EAAE,MAAM,CAAC,CAAC,CAAC;IAInF,QAAQ,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,SAAS,GAAG;QAAE,GAAG,CAAC,EAAE,UAAU,CAAC,OAAO,WAAW,CAAC,CAAC,CAAC,CAAC,CAAA;KAAE,GAAG,OAAO,CAAC,MAAM,CAAC;CAG7G;AAED,wFAAwF;AACxF,MAAM,WAAW,qBAAqB;IACrC,IAAI,EAAE,WAAW,CAAA;IACjB,MAAM,EAAE,MAAM,EAAE,EAAE,CAAA;IAClB,MAAM,EAAE,KAAK,CAAC;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,GAAG,EAAE,MAAM,CAAA;KAAE,CAAC,CAAA;CAC7C;AAED;;;GAGG;AACH,MAAM,WAAW,SAAS;IACzB;;;;OAIG;IACH,UAAU,CAAC,EAAE,cAAc,CAAA;IAC3B;;;OAGG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAA;CAC5B"}

package/out/classifier.js

@@ -10,16 +10,30 @@
  *   Convenience wrappers `parseJson` / `parseTuples` / `parseXml` project the tree on the way out.
  */
 import { buildAddressTree, decodeAsJson, decodeAsTuples, decodeAsXml, } from "@mailwoman/core/decoder";
-import { STAGE1_BIO_LABELS } from "./labels.js";
-import { OnnxRunner } from "./onnx-runner.js";
+import { STAGE2_BIO_LABELS } from "./labels.js";
+import { addEmissionMatrix, buildEmissionPriors } from "./query-shape-prior.js";
 import { MailwomanTokenizer } from "./tokenizer.js";
-import { resolveWeights } from "./weights.js";
+import { buildBioEndMask, buildBioStartMask, buildBioTransitionMask, softmax, viterbi } from "./viterbi.js";
 export class NeuralAddressClassifier {
     cfg;
     labels;
+    decodeMode;
+    transitions;
+    startTransitions;
+    endTransitions;
     constructor(cfg) {
         this.cfg = cfg;
-        this.labels = cfg.labels ?? STAGE1_BIO_LABELS;
+        this.labels = cfg.labels ?? STAGE2_BIO_LABELS;
+        this.decodeMode = cfg.decode ?? "viterbi";
+        const structural = buildBioTransitionMask(this.labels);
+        if (cfg.transitions) {
+            this.transitions = addMatrices(structural, cfg.transitions);
+        }
+        else {
+            this.transitions = structural;
+        }
+        this.startTransitions = cfg.startTransitions ?? buildBioStartMask(this.labels);
+        this.endTransitions = cfg.endTransitions ?? buildBioEndMask(this.labels);
     }
     /**
      * One-call factory that resolves the weights package (or explicit paths), loads the tokenizer and
@@ -27,42 +41,120 @@ export class NeuralAddressClassifier {
      *
      * Resolution order: explicit paths in `opts` → `@mailwoman/neural-weights-<locale>` package →
      * throws a single actionable error.
+     *
+     * **Node-only.** The dynamic imports keep `OnnxRunner` (onnxruntime-node) + `resolveWeights`
+     * (uses Node fs) out of the static dependency graph, so this file can be bundled for the browser
+     * by `@mailwoman/neural-web`. Calling this method in a browser will throw at runtime — use
+     * `loadNeuralClassifierFromUrls` from `@mailwoman/neural-web` instead.
      */
     static async loadFromWeights(opts = {}) {
-        const { modelPath, tokenizerPath } = resolveWeights(opts);
+        // /* webpackIgnore: true */ tells webpack to leave the dynamic import statement intact —
+        // it becomes a runtime native ESM import that resolves in Node (which has onnxruntime-node
+        // + node:fs) and throws cleanly in a browser if called. Without the directive, webpack
+        // pulls onnx-runner / weights into the browser chunk graph + then chokes on the Node-only
+        // builtins they reference.
+        const [{ OnnxRunner }, { resolveWeights, readLabelsFromModelCard }] = await Promise.all([
+            import(/* webpackIgnore: true */ "./onnx-runner.js"),
+            import(/* webpackIgnore: true */ "./weights.js"),
+        ]);
+        const resolved = resolveWeights(opts);
+        // Read the trained label vocabulary from the bundled model-card.json when present. Falls
+        // through to the constructor default (STAGE2_BIO_LABELS) for legacy bundles that predate
+        // the `labels` field — those are always Stage 2 cards by construction, so the default is
+        // the correct fallback. A future Stage 3 ship will require the card to carry the field.
+        const labels = readLabelsFromModelCard(resolved.modelCardPath);
         const [tokenizer, runner] = await Promise.all([
-            MailwomanTokenizer.loadFromFile(tokenizerPath),
-            OnnxRunner.create(modelPath),
+            MailwomanTokenizer.loadFromFile(resolved.tokenizerPath),
+            OnnxRunner.create(resolved.modelPath),
         ]);
-        return new NeuralAddressClassifier({ tokenizer, runner });
+        return new NeuralAddressClassifier({ tokenizer, runner, labels });
     }
-    /** Tokenize → infer → argmax/softmax → decoder tree. */
-    async parse(text) {
+    /** Tokenize → infer → Viterbi (or argmax) → decoder tree. */
+    async parse(text, opts) {
         if (text.length === 0)
             return { raw: text, roots: [] };
         const { pieces, ids } = this.cfg.tokenizer.encode(text);
         const { logits } = await this.cfg.runner.infer(ids);
+        // QueryShape soft prior: when the caller supplies a QueryShape (typically from
+        // `@mailwoman/query-shape`'s `computeQueryShape`), nudge per-token emissions toward the
+        // labels implied by known-format hits. Bounded magnitude — confident encoder predictions
+        // still win.
+        const emissions = opts?.queryShape
+            ? addEmissionMatrix(logits, buildEmissionPriors(opts.queryShape, pieces, this.labels, {
+                biasScale: opts.queryShapeBiasScale ?? 1.0,
+            }))
+            : logits;
+        const labelIndices = this.decodeMode === "viterbi"
+            ? viterbi({
+                emissions,
+                transitions: this.transitions,
+                startTransitions: this.startTransitions,
+                endTransitions: this.endTransitions,
+            }).path
+            : emissions.map((row) => argmaxSoftmax(row).idx);
         const tokens = pieces.map((p, i) => {
-            const row = logits[i];
-            const { idx, conf } = argmaxSoftmax(row);
+            const idx = labelIndices[i];
+            // Confidence reports the encoder's *raw* probability (no prior baked in) so callers see
+            // the model's own conviction, not the prior-augmented score.
+            const probs = softmax(logits[i]);
             return {
                 piece: p.piece,
                 start: p.start,
                 end: p.end,
                 label: (this.labels[idx] ?? "O"),
-                confidence: conf,
+                confidence: probs[idx],
             };
         });
         return buildAddressTree(text, tokens);
     }
-    async parseJson(text) {
-        return decodeAsJson(await this.parse(text));
+    /**
+     * Like `parse`, but also returns the raw per-token logits and piece offsets needed for per-span
+     * logit aggregation (Option C joint-reconcile integration).
+     */
+    async parseWithLogits(text, opts) {
+        if (text.length === 0) {
+            return { tree: { raw: text, roots: [] }, logits: [], pieces: [] };
+        }
+        const { pieces, ids } = this.cfg.tokenizer.encode(text);
+        const { logits } = await this.cfg.runner.infer(ids);
+        const emissions = opts?.queryShape
+            ? addEmissionMatrix(logits, buildEmissionPriors(opts.queryShape, pieces, this.labels, {
+                biasScale: opts.queryShapeBiasScale ?? 1.0,
+            }))
+            : logits;
+        const labelIndices = this.decodeMode === "viterbi"
+            ? viterbi({
+                emissions,
+                transitions: this.transitions,
+                startTransitions: this.startTransitions,
+                endTransitions: this.endTransitions,
+            }).path
+            : emissions.map((row) => argmaxSoftmax(row).idx);
+        const tokens = pieces.map((p, i) => {
+            const idx = labelIndices[i];
+            const probs = softmax(logits[i]);
+            return {
+                piece: p.piece,
+                start: p.start,
+                end: p.end,
+                label: (this.labels[idx] ?? "O"),
+                confidence: probs[idx],
+            };
+        });
+        return {
+            tree: buildAddressTree(text, tokens),
+            logits,
+            pieces: pieces.map((p) => ({ start: p.start, end: p.end })),
+        };
+    }
+    async parseJson(text, opts) {
+        return decodeAsJson(await this.parse(text, opts));
     }
-    async parseTuples(text) {
-        return decodeAsTuples(await this.parse(text));
+    async parseTuples(text, opts) {
+        return decodeAsTuples(await this.parse(text, opts));
     }
     async parseXml(text, opts) {
-        return decodeAsXml(await this.parse(text), opts);
+        return decodeAsXml(await this.parse(text, opts), opts?.xml);
     }
 }
 function argmaxSoftmax(row) {
@@ -80,4 +172,16 @@ function argmaxSoftmax(row) {
     const conf = 1 / sumExp;
     return { idx: maxIdx, conf };
 }
+/** Element-wise add two square matrices. Used to compose the structural mask + learned transitions. */
+function addMatrices(a, b) {
+    const n = a.length;
+    const out = [];
+    for (let i = 0; i < n; i++) {
+        const row = new Array(n);
+        for (let j = 0; j < n; j++)
+            row[j] = a[i][j] + b[i][j];
+        out.push(row);
+    }
+    return out;
+}
 //# sourceMappingURL=classifier.js.map

package/out/classifier.js.map

@@ -1 +1 @@
-{"version":3,"file":"classifier.js","sourceRoot":"","sources":["../classifier.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAEH,OAAO,EAIN,gBAAgB,EAChB,YAAY,EACZ,cAAc,EACd,WAAW,GACX,MAAM,yBAAyB,CAAA;AAChC,OAAO,EAAE,iBAAiB,EAAE,MAAM,aAAa,CAAA;AAC/C,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAA;AAC7C,OAAO,EAAE,kBAAkB,EAAE,MAAM,gBAAgB,CAAA;AACnD,OAAO,EAA2B,cAAc,EAAE,MAAM,cAAc,CAAA;AAStE,MAAM,OAAO,uBAAuB;IAGN;IAFZ,MAAM,CAAmB;IAE1C,YAA6B,GAAkC;QAAlC,QAAG,GAAH,GAAG,CAA+B;QAC9D,IAAI,CAAC,MAAM,GAAG,GAAG,CAAC,MAAM,IAAI,iBAAiB,CAAA;IAC9C,CAAC;IAED;;;;;;OAMG;IACH,MAAM,CAAC,KAAK,CAAC,eAAe,CAAC,OAA2B,EAAE;QACzD,MAAM,EAAE,SAAS,EAAE,aAAa,EAAE,GAAG,cAAc,CAAC,IAAI,CAAC,CAAA;QACzD,MAAM,CAAC,SAAS,EAAE,MAAM,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;YAC7C,kBAAkB,CAAC,YAAY,CAAC,aAAa,CAAC;YAC9C,UAAU,CAAC,MAAM,CAAC,SAAS,CAAC;SAC5B,CAAC,CAAA;QACF,OAAO,IAAI,uBAAuB,CAAC,EAAE,SAAS,EAAE,MAAM,EAAE,CAAC,CAAA;IAC1D,CAAC;IAED,wDAAwD;IACxD,KAAK,CAAC,KAAK,CAAC,IAAY;QACvB,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,GAAG,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,CAAA;QAEtD,MAAM,EAAE,MAAM,EAAE,GAAG,EAAE,GAAG,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;QACvD,MAAM,EAAE,MAAM,EAAE,GAAG,MAAM,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAA;QAEnD,MAAM,MAAM,GAAmB,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;YAClD,MAAM,GAAG,GAAG,MAAM,CAAC,CAAC,CAAE,CAAA;YACtB,MAAM,EAAE,GAAG,EAAE,IAAI,EAAE,GAAG,aAAa,CAAC,GAAG,CAAC,CAAA;YACxC,OAAO;gBACN,KAAK,EAAE,CAAC,CAAC,KAAK;gBACd,KAAK,EAAE,CAAC,CAAC,KAAK;gBACd,GAAG,EAAE,CAAC,CAAC,GAAG;gBACV,KAAK,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,GAAG,CAA0B;gBACzD,UAAU,EAAE,IAAI;aAChB,CAAA;QACF,CAAC,CAAC,CAAA;QAEF,OAAO,gBAAgB,CAAC,IAAI,EAAE,MAAM,CAAC,CAAA;IACtC,CAAC;IAED,KAAK,CAAC,SAAS,CAAC,IAAY;QAC3B,OAAO,YAAY,CAAC,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAA;IAC5C,CAAC;IAED,KAAK,CAAC,WAAW,CAAC,IAAY;QAC7B,OAAO,cAAc,CAAC,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAA;IAC9C,CAAC;IAED,KAAK,CAAC,QAAQ,CAAC,IAAY,EAAE,IAAwC;QACpE,OAAO,WAAW,CAAC,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,IAAI,CAAC,CAAA;IACjD,CAAC;CACD;AAED,SAAS,aAAa,CAAC,GAAa;IACnC,IAAI,MAAM,GAAG,CAAC,CAAA;IACd,IAAI,MAAM,GAAG,GAAG,CAAC,CAAC,CAAE,CAAA;IACpB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,IAAI,GAAG,CAAC,CAAC,CAAE,GAAG,MAAM,EAAE,CAAC;YACtB,MAAM,GAAG,GAAG,CAAC,CAAC,CAAE,CAAA;YAChB,MAAM,GAAG,CAAC,CAAA;QACX,CAAC;IACF,CAAC;IACD,IAAI,MAAM,GAAG,CAAC,CAAA;IACd,KAAK,MAAM,CAAC,IAAI,GAAG;QAAE,MAAM,IAAI,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,MAAM,CAAC,CAAA;IACnD,MAAM,IAAI,GAAG,CAAC,GAAG,MAAM,CAAA;IACvB,OAAO,EAAE,GAAG,EAAE,MAAM,EAAE,IAAI,EAAE,CAAA;AAC7B,CAAC"}
+{"version":3,"file":"classifier.js","sourceRoot":"","sources":["../classifier.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAEH,OAAO,EACN,gBAAgB,EAChB,YAAY,EACZ,cAAc,EACd,WAAW,GAIX,MAAM,yBAAyB,CAAA;AAChC,OAAO,EAAE,iBAAiB,EAAE,MAAM,aAAa,CAAA;AAE/C,OAAO,EAAE,iBAAiB,EAAE,mBAAmB,EAAuB,MAAM,wBAAwB,CAAA;AACpG,OAAO,EAAE,kBAAkB,EAAE,MAAM,gBAAgB,CAAA;AACnD,OAAO,EAAE,eAAe,EAAE,iBAAiB,EAAE,sBAAsB,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,cAAc,CAAA;AA0C3G,MAAM,OAAO,uBAAuB;IAON;IANZ,MAAM,CAAmB;IACzB,UAAU,CAAsB;IAChC,WAAW,CAAY;IACvB,gBAAgB,CAAU;IAC1B,cAAc,CAAU;IAEzC,YAA6B,GAAkC;QAAlC,QAAG,GAAH,GAAG,CAA+B;QAC9D,IAAI,CAAC,MAAM,GAAG,GAAG,CAAC,MAAM,IAAI,iBAAiB,CAAA;QAC7C,IAAI,CAAC,UAAU,GAAG,GAAG,CAAC,MAAM,IAAI,SAAS,CAAA;QACzC,MAAM,UAAU,GAAG,sBAAsB,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;QACtD,IAAI,GAAG,CAAC,WAAW,EAAE,CAAC;YACrB,IAAI,CAAC,WAAW,GAAG,WAAW,CAAC,UAAU,EAAE,GAAG,CAAC,WAAW,CAAC,CAAA;QAC5D,CAAC;aAAM,CAAC;YACP,IAAI,CAAC,WAAW,GAAG,UAAU,CAAA;QAC9B,CAAC;QACD,IAAI,CAAC,gBAAgB,GAAG,GAAG,CAAC,gBAAgB,IAAI,iBAAiB,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;QAC9E,IAAI,CAAC,cAAc,GAAG,GAAG,CAAC,cAAc,IAAI,eAAe,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;IACzE,CAAC;IAED;;;;;;;;;;;OAWG;IACH,MAAM,CAAC,KAAK,CAAC,eAAe,CAAC,OAA2B,EAAE;QACzD,yFAAyF;QACzF,2FAA2F;QAC3F,uFAAuF;QACvF,0FAA0F;QAC1F,2BAA2B;QAC3B,MAAM,CAAC,EAAE,UAAU,EAAE,EAAE,EAAE,cAAc,EAAE,uBAAuB,EAAE,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;YACvF,MAAM,CAAC,yBAAyB,CAAC,kBAAkB,CAAC;YACpD,MAAM,CAAC,yBAAyB,CAAC,cAAc,CAAC;SAChD,CAAC,CAAA;QACF,MAAM,QAAQ,GAAoB,cAAc,CAAC,IAAI,CAAC,CAAA;QACtD,yFAAyF;QACzF,yFAAyF;QACzF,yFAAyF;QACzF,wFAAwF;QACxF,MAAM,MAAM,GAAG,uBAAuB,CAAC,QAAQ,CAAC,aAAa,CAAC,CAAA;QAC9D,MAAM,CAAC,SAAS,EAAE,MAAM,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;YAC7C,kBAAkB,CAAC,YAAY,CAAC,QAAQ,CAAC,aAAa,CAAC;YACvD,UAAU,CAAC,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC;SACrC,CAAC,CAAA;QACF,OAAO,IAAI,uBAAuB,CAAC,EAAE,SAAS,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC,CAAA;IAClE,CAAC;IAED,6DAA6D;IAC7D,KAAK,CAAC,KAAK,CAAC,IAAY,EAAE,IAAgB;QACzC,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,EAAE,GAAG,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,CAAA;QAEtD,MAAM,EAAE,MAAM,EAAE,GAAG,EAAE,GAAG,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;QACvD,MAAM,EAAE,MAAM,EAAE,GAAG,MAAM,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAA;QAEnD,+EAA+E;QAC/E,wFAAwF;QACxF,yFAAyF;QACzF,aAAa;QACb,MAAM,SAAS,GAAG,IAAI,EAAE,UAAU;YACjC,CAAC,CAAC,iBAAiB,CACjB,MAAM,EACN,mBAAmB,CAAC,IAAI,CAAC,UAAU,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,EAAE;gBACzD,SAAS,EAAE,IAAI,CAAC,mBAAmB,IAAI,GAAG;aAC1C,CAAC,CACF;YACF,CAAC,CAAC,MAAM,CAAA;QAET,MAAM,YAAY,GACjB,IAAI,CAAC,UAAU,KAAK,SAAS;YAC5B,CAAC,CAAC,OAAO,CAAC;gBACR,SAAS;gBACT,WAAW,EAAE,IAAI,CAAC,WAAW;gBAC7B,gBAAgB,EAAE,IAAI,CAAC,gBAAgB;gBACvC,cAAc,EAAE,IAAI,CAAC,cAAc;aACnC,CAAC,CAAC,IAAI;YACR,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,aAAa,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAA;QAElD,MAAM,MAAM,GAAmB,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;YAClD,MAAM,GAAG,GAAG,YAAY,CAAC,CAAC,CAAE,CAAA;YAC5B,wFAAwF;YACxF,6DAA6D;YAC7D,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC,CAAA;YACjC,OAAO;gBACN,KAAK,EAAE,CAAC,CAAC,KAAK;gBACd,KAAK,EAAE,CAAC,CAAC,KAAK;gBACd,GAAG,EAAE,CAAC,CAAC,GAAG;gBACV,KAAK,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,GAAG,CAA0B;gBACzD,UAAU,EAAE,KAAK,CAAC,GAAG,CAAE;aACvB,CAAA;QACF,CAAC,CAAC,CAAA;QAEF,OAAO,gBAAgB,CAAC,IAAI,EAAE,MAAM,CAAC,CAAA;IACtC,CAAC;IAED;;;OAGG;IACH,KAAK,CAAC,eAAe,CAAC,IAAY,EAAE,IAAgB;QACnD,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACvB,OAAO,EAAE,IAAI,EAAE,EAAE,GAAG,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,EAAE,EAAE,EAAE,CAAA;QAClE,CAAC;QACD,MAAM,EAAE,MAAM,EAAE,GAAG,EAAE,GAAG,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;QACvD,MAAM,EAAE,MAAM,EAAE,GAAG,MAAM,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAA;QAEnD,MAAM,SAAS,GAAG,IAAI,EAAE,UAAU;YACjC,CAAC,CAAC,iBAAiB,CACjB,MAAM,EACN,mBAAmB,CAAC,IAAI,CAAC,UAAU,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,EAAE;gBACzD,SAAS,EAAE,IAAI,CAAC,mBAAmB,IAAI,GAAG;aAC1C,CAAC,CACF;YACF,CAAC,CAAC,MAAM,CAAA;QAET,MAAM,YAAY,GACjB,IAAI,CAAC,UAAU,KAAK,SAAS;YAC5B,CAAC,CAAC,OAAO,CAAC;gBACR,SAAS;gBACT,WAAW,EAAE,IAAI,CAAC,WAAW;gBAC7B,gBAAgB,EAAE,IAAI,CAAC,gBAAgB;gBACvC,cAAc,EAAE,IAAI,CAAC,cAAc;aACnC,CAAC,CAAC,IAAI;YACR,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,aAAa,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAA;QAElD,MAAM,MAAM,GAAmB,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;YAClD,MAAM,GAAG,GAAG,YAAY,CAAC,CAAC,CAAE,CAAA;YAC5B,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC,CAAA;YACjC,OAAO;gBACN,KAAK,EAAE,CAAC,CAAC,KAAK;gBACd,KAAK,EAAE,CAAC,CAAC,KAAK;gBACd,GAAG,EAAE,CAAC,CAAC,GAAG;gBACV,KAAK,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,GAAG,CAA0B;gBACzD,UAAU,EAAE,KAAK,CAAC,GAAG,CAAE;aACvB,CAAA;QACF,CAAC,CAAC,CAAA;QAEF,OAAO;YACN,IAAI,EAAE,gBAAgB,CAAC,IAAI,EAAE,MAAM,CAAC;YACpC,MAAM;YACN,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC;SAC3D,CAAA;IACF,CAAC;IAED,KAAK,CAAC,SAAS,CAAC,IAAY,EAAE,IAAgB;QAC7C,OAAO,YAAY,CAAC,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAA;IAClD,CAAC;IAED,KAAK,CAAC,WAAW,CAAC,IAAY,EAAE,IAAgB;QAC/C,OAAO,cAAc,CAAC,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAA;IACpD,CAAC;IAED,KAAK,CAAC,QAAQ,CAAC,IAAY,EAAE,IAA8D;QAC1F,OAAO,WAAW,CAAC,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,EAAE,GAAG,CAAC,CAAA;IAC5D,CAAC;CACD;AA2BD,SAAS,aAAa,CAAC,GAAa;IACnC,IAAI,MAAM,GAAG,CAAC,CAAA;IACd,IAAI,MAAM,GAAG,GAAG,CAAC,CAAC,CAAE,CAAA;IACpB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,IAAI,GAAG,CAAC,CAAC,CAAE,GAAG,MAAM,EAAE,CAAC;YACtB,MAAM,GAAG,GAAG,CAAC,CAAC,CAAE,CAAA;YAChB,MAAM,GAAG,CAAC,CAAA;QACX,CAAC;IACF,CAAC;IACD,IAAI,MAAM,GAAG,CAAC,CAAA;IACd,KAAK,MAAM,CAAC,IAAI,GAAG;QAAE,MAAM,IAAI,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,MAAM,CAAC,CAAA;IACnD,MAAM,IAAI,GAAG,CAAC,GAAG,MAAM,CAAA;IACvB,OAAO,EAAE,GAAG,EAAE,MAAM,EAAE,IAAI,EAAE,CAAA;AAC7B,CAAC;AAED,uGAAuG;AACvG,SAAS,WAAW,CAAC,CAAa,EAAE,CAAa;IAChD,MAAM,CAAC,GAAG,CAAC,CAAC,MAAM,CAAA;IAClB,MAAM,GAAG,GAAe,EAAE,CAAA;IAC1B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC5B,MAAM,GAAG,GAAG,IAAI,KAAK,CAAS,CAAC,CAAC,CAAA;QAChC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE;YAAE,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAE,CAAC,CAAC,CAAE,GAAG,CAAC,CAAC,CAAC,CAAE,CAAC,CAAC,CAAE,CAAA;QAC1D,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;IACd,CAAC;IACD,OAAO,GAAG,CAAA;AACX,CAAC"}

package/out/index.d.ts

@@ -7,6 +7,10 @@ export * from "./classifier.js";
 export * from "./labels.js";
 export * from "./onnx-runner.js";
 export * from "./proposal-classifier.js";
+export { addEmissionMatrix, buildEmissionPriors } from "./query-shape-prior.js";
+export type { BuildPriorsOpts, KnownFormatHitLike, QueryShapeLike, TokenLike } from "./query-shape-prior.js";
 export * from "./tokenizer.js";
+export { buildBioEndMask, buildBioStartMask, buildBioTransitionMask, perTokenArgmax, softmax, viterbi, } from "./viterbi.js";
+export type { ViterbiInput, ViterbiResult } from "./viterbi.js";
 export * from "./weights.js";
 //# sourceMappingURL=index.d.ts.map

package/out/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../index.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,cAAc,iBAAiB,CAAA;AAC/B,cAAc,aAAa,CAAA;AAC3B,cAAc,kBAAkB,CAAA;AAChC,cAAc,0BAA0B,CAAA;AACxC,cAAc,gBAAgB,CAAA;AAC9B,cAAc,cAAc,CAAA"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../index.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,cAAc,iBAAiB,CAAA;AAC/B,cAAc,aAAa,CAAA;AAC3B,cAAc,kBAAkB,CAAA;AAChC,cAAc,0BAA0B,CAAA;AACxC,OAAO,EAAE,iBAAiB,EAAE,mBAAmB,EAAE,MAAM,wBAAwB,CAAA;AAC/E,YAAY,EAAE,eAAe,EAAE,kBAAkB,EAAE,cAAc,EAAE,SAAS,EAAE,MAAM,wBAAwB,CAAA;AAC5G,cAAc,gBAAgB,CAAA;AAC9B,OAAO,EACN,eAAe,EACf,iBAAiB,EACjB,sBAAsB,EACtB,cAAc,EACd,OAAO,EACP,OAAO,GACP,MAAM,cAAc,CAAA;AACrB,YAAY,EAAE,YAAY,EAAE,aAAa,EAAE,MAAM,cAAc,CAAA;AAC/D,cAAc,cAAc,CAAA"}

package/out/index.js

@@ -7,6 +7,8 @@ export * from "./classifier.js";
 export * from "./labels.js";
 export * from "./onnx-runner.js";
 export * from "./proposal-classifier.js";
+export { addEmissionMatrix, buildEmissionPriors } from "./query-shape-prior.js";
 export * from "./tokenizer.js";
+export { buildBioEndMask, buildBioStartMask, buildBioTransitionMask, perTokenArgmax, softmax, viterbi, } from "./viterbi.js";
 export * from "./weights.js";
 //# sourceMappingURL=index.js.map

package/out/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","sourceRoot":"","sources":["../index.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,cAAc,iBAAiB,CAAA;AAC/B,cAAc,aAAa,CAAA;AAC3B,cAAc,kBAAkB,CAAA;AAChC,cAAc,0BAA0B,CAAA;AACxC,cAAc,gBAAgB,CAAA;AAC9B,cAAc,cAAc,CAAA"}
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../index.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,cAAc,iBAAiB,CAAA;AAC/B,cAAc,aAAa,CAAA;AAC3B,cAAc,kBAAkB,CAAA;AAChC,cAAc,0BAA0B,CAAA;AACxC,OAAO,EAAE,iBAAiB,EAAE,mBAAmB,EAAE,MAAM,wBAAwB,CAAA;AAE/E,cAAc,gBAAgB,CAAA;AAC9B,OAAO,EACN,eAAe,EACf,iBAAiB,EACjB,sBAAsB,EACtB,cAAc,EACd,OAAO,EACP,OAAO,GACP,MAAM,cAAc,CAAA;AAErB,cAAc,cAAc,CAAA"}

package/out/labels.d.ts

@@ -3,18 +3,39 @@
  * @license AGPL-3.0
  * @author Teffen Ellis, et al.
  *
- *   Mirror of `packages/corpus-python/src/mailwoman_train/labels.py::STAGE1_BIO_LABELS`.
+ *   Mirror of `packages/corpus-python/src/mailwoman_train/labels.py`.
  *
- *   The v0.1.0 / v0.2.0 weight packages were trained with this exact label order. Any drift here
- *   silently corrupts downstream BIO decoding — index 5 must mean `B-locality` on both sides.
+ *   Index ↔ label parity is load-bearing: the model emits logits in one canonical order on both sides
+ *   and any drift here silently corrupts BIO decoding. STAGE2 strictly extends STAGE1 — the first
+ *   15 indices are identical, so reading a v0.2.0 (Stage 1) model with the Stage 2 label vocabulary
+ *   stays correct; the extra entries are unused.
  *
- *   Stage 2+ models will support more labels (street, house_number, venue, …). The plan is to plumb
- *   the label set through `model-card.json` at load time rather than hard-coding it here. Until
- *   then this file is the source of truth on the TS side.
+ *   Runtime loading: as of v0.4.0 the trained label vocabulary is carried in `model-card.json`'s
+ *   `labels` field and read by `loadFromWeights` (see `weights.readLabelsFromModelCard`). These
+ *   constants remain the compile-time fallback for legacy bundles whose cards predate the field —
+ *   safe because such bundles are by construction Stage 1 or Stage 2, and Stage 2 prefix-extends
+ *   Stage 1. A future Stage 3 ship will not be safe under the fallback; the loader treats a missing
+ *   `labels` field as "you are loading a pre-v0.4.0 bundle" rather than "unknown stage".
  */
 import type { BioLabel } from "@mailwoman/core/decoder";
 /** Coarse component tags trained in Phase 2 Stage 1 (v0.1.0 / v0.2.0). */
 export declare const STAGE1_COARSE_TAGS: readonly ["country", "region", "locality", "dependent_locality", "postcode", "subregion", "cedex"];
 /** BIO label vocabulary for Stage 1 — O + (B-/I- per coarse tag). 1 + 14 = 15 labels. */
 export declare const STAGE1_BIO_LABELS: readonly BioLabel[];
+/**
+ * Fine-grained tags added in Phase 2 Stage 2 (v0.3.0). venue covers organization/POI/landmark
+ * names; street + house_number break out the street-address components that Stage 1 collapsed to
+ * `O`.
+ */
+export declare const STAGE2_FINE_TAGS: readonly ["venue", "street", "house_number"];
+/** Stage 2 ships the full coarse + fine set in the order STAGE2_BIO_LABELS is interleaved. */
+export declare const STAGE2_TAGS: readonly ["country", "region", "locality", "dependent_locality", "postcode", "subregion", "cedex", "venue", "street", "house_number"];
+/**
+ * BIO label vocabulary for Stage 2 (v0.3.0) — O + (B-/I- per Stage 2 tag). 1 + 20 = 21 labels.
+ *
+ * Index parity vs Stage 1: STAGE2_BIO_LABELS[i] === STAGE1_BIO_LABELS[i] for i ∈ [0, 15). Anyone
+ * loading a Stage 1 model with this vocabulary still decodes correctly; the tail (15..20) just
+ * never gets argmax'd because Stage 1 only emits 15 logits.
+ */
+export declare const STAGE2_BIO_LABELS: readonly BioLabel[];
 //# sourceMappingURL=labels.d.ts.map

package/out/labels.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"labels.d.ts","sourceRoot":"","sources":["../labels.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAEH,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,yBAAyB,CAAA;AAEvD,0EAA0E;AAC1E,eAAO,MAAM,kBAAkB,oGAQrB,CAAA;AAEV,yFAAyF;AACzF,eAAO,MAAM,iBAAiB,EAAE,SAAS,QAAQ,EAG/C,CAAA"}
+{"version":3,"file":"labels.d.ts","sourceRoot":"","sources":["../labels.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAEH,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,yBAAyB,CAAA;AAEvD,0EAA0E;AAC1E,eAAO,MAAM,kBAAkB,oGAQrB,CAAA;AAEV,yFAAyF;AACzF,eAAO,MAAM,iBAAiB,EAAE,SAAS,QAAQ,EAG/C,CAAA;AAEF;;;;GAIG;AACH,eAAO,MAAM,gBAAgB,8CAA+C,CAAA;AAE5E,8FAA8F;AAC9F,eAAO,MAAM,WAAW,uIAAwD,CAAA;AAEhF;;;;;;GAMG;AACH,eAAO,MAAM,iBAAiB,EAAE,SAAS,QAAQ,EAG/C,CAAA"}

package/out/labels.js

@@ -3,14 +3,19 @@
  * @license AGPL-3.0
  * @author Teffen Ellis, et al.
  *
- *   Mirror of `packages/corpus-python/src/mailwoman_train/labels.py::STAGE1_BIO_LABELS`.
+ *   Mirror of `packages/corpus-python/src/mailwoman_train/labels.py`.
  *
- *   The v0.1.0 / v0.2.0 weight packages were trained with this exact label order. Any drift here
- *   silently corrupts downstream BIO decoding — index 5 must mean `B-locality` on both sides.
+ *   Index ↔ label parity is load-bearing: the model emits logits in one canonical order on both sides
+ *   and any drift here silently corrupts BIO decoding. STAGE2 strictly extends STAGE1 — the first
+ *   15 indices are identical, so reading a v0.2.0 (Stage 1) model with the Stage 2 label vocabulary
+ *   stays correct; the extra entries are unused.
  *
- *   Stage 2+ models will support more labels (street, house_number, venue, …). The plan is to plumb
- *   the label set through `model-card.json` at load time rather than hard-coding it here. Until
- *   then this file is the source of truth on the TS side.
+ *   Runtime loading: as of v0.4.0 the trained label vocabulary is carried in `model-card.json`'s
+ *   `labels` field and read by `loadFromWeights` (see `weights.readLabelsFromModelCard`). These
+ *   constants remain the compile-time fallback for legacy bundles whose cards predate the field —
+ *   safe because such bundles are by construction Stage 1 or Stage 2, and Stage 2 prefix-extends
+ *   Stage 1. A future Stage 3 ship will not be safe under the fallback; the loader treats a missing
+ *   `labels` field as "you are loading a pre-v0.4.0 bundle" rather than "unknown stage".
  */
 /** Coarse component tags trained in Phase 2 Stage 1 (v0.1.0 / v0.2.0). */
 export const STAGE1_COARSE_TAGS = [
@@ -27,4 +32,23 @@ export const STAGE1_BIO_LABELS = Object.freeze([
     "O",
     ...STAGE1_COARSE_TAGS.flatMap((tag) => [`B-${tag}`, `I-${tag}`]),
 ]);
+/**
+ * Fine-grained tags added in Phase 2 Stage 2 (v0.3.0). venue covers organization/POI/landmark
+ * names; street + house_number break out the street-address components that Stage 1 collapsed to
+ * `O`.
+ */
+export const STAGE2_FINE_TAGS = ["venue", "street", "house_number"];
+/** Stage 2 ships the full coarse + fine set in the order STAGE2_BIO_LABELS is interleaved. */
+export const STAGE2_TAGS = [...STAGE1_COARSE_TAGS, ...STAGE2_FINE_TAGS];
+/**
+ * BIO label vocabulary for Stage 2 (v0.3.0) — O + (B-/I- per Stage 2 tag). 1 + 20 = 21 labels.
+ *
+ * Index parity vs Stage 1: STAGE2_BIO_LABELS[i] === STAGE1_BIO_LABELS[i] for i ∈ [0, 15). Anyone
+ * loading a Stage 1 model with this vocabulary still decodes correctly; the tail (15..20) just
+ * never gets argmax'd because Stage 1 only emits 15 logits.
+ */
+export const STAGE2_BIO_LABELS = Object.freeze([
+    "O",
+    ...STAGE2_TAGS.flatMap((tag) => [`B-${tag}`, `I-${tag}`]),
+]);
 //# sourceMappingURL=labels.js.map

package/out/labels.js.map

@@ -1 +1 @@
-{"version":3,"file":"labels.js","sourceRoot":"","sources":["../labels.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAIH,0EAA0E;AAC1E,MAAM,CAAC,MAAM,kBAAkB,GAAG;IACjC,SAAS;IACT,QAAQ;IACR,UAAU;IACV,oBAAoB;IACpB,UAAU;IACV,WAAW;IACX,OAAO;CACE,CAAA;AAEV,yFAAyF;AACzF,MAAM,CAAC,MAAM,iBAAiB,GAAwB,MAAM,CAAC,MAAM,CAAC;IACnE,GAAe;IACf,GAAG,kBAAkB,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,KAAK,GAAG,EAAc,EAAE,KAAK,GAAG,EAAc,CAAC,CAAC;CACxF,CAAC,CAAA"}
+{"version":3,"file":"labels.js","sourceRoot":"","sources":["../labels.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAIH,0EAA0E;AAC1E,MAAM,CAAC,MAAM,kBAAkB,GAAG;IACjC,SAAS;IACT,QAAQ;IACR,UAAU;IACV,oBAAoB;IACpB,UAAU;IACV,WAAW;IACX,OAAO;CACE,CAAA;AAEV,yFAAyF;AACzF,MAAM,CAAC,MAAM,iBAAiB,GAAwB,MAAM,CAAC,MAAM,CAAC;IACnE,GAAe;IACf,GAAG,kBAAkB,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,KAAK,GAAG,EAAc,EAAE,KAAK,GAAG,EAAc,CAAC,CAAC;CACxF,CAAC,CAAA;AAEF;;;;GAIG;AACH,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,OAAO,EAAE,QAAQ,EAAE,cAAc,CAAU,CAAA;AAE5E,8FAA8F;AAC9F,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,GAAG,kBAAkB,EAAE,GAAG,gBAAgB,CAAU,CAAA;AAEhF;;;;;;GAMG;AACH,MAAM,CAAC,MAAM,iBAAiB,GAAwB,MAAM,CAAC,MAAM,CAAC;IACnE,GAAe;IACf,GAAG,WAAW,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,KAAK,GAAG,EAAc,EAAE,KAAK,GAAG,EAAc,CAAC,CAAC;CACjF,CAAC,CAAA"}

package/out/proposal-classifier.d.ts

@@ -24,7 +24,11 @@ export interface NeuralProposalClassifierConfig {
     id: string;
     /** The underlying neural classifier instance. */
     classifier: NeuralAddressClassifier;
-    /** Component tags this classifier may emit. Defaults to Stage 1 coarse tags. */
+    /**
+     * Component tags this classifier may emit. Defaults to the Stage 2 tag set (coarse +
+     * venue/street/house_number). v0.2.0 Stage 1 models never decode to the fine tags anyway, so the
+     * broader default is forwards-compat without back-compat risk.
+     */
     emits?: readonly ComponentTag[];
     /** Locales this classifier is active for. `["*"]` (locale-agnostic) by default. */
     locales?: readonly (string | "*")[];

package/out/proposal-classifier.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"proposal-classifier.d.ts","sourceRoot":"","sources":["../proposal-classifier.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAIH,OAAO,KAAK,EAGX,YAAY,EACZ,kBAAkB,EAElB,MAAM,uBAAuB,CAAA;AAC9B,OAAO,KAAK,EAAE,uBAAuB,EAAE,MAAM,iBAAiB,CAAA;AAG9D,MAAM,WAAW,8BAA8B;IAC9C,wFAAwF;IACxF,EAAE,EAAE,MAAM,CAAA;IACV,iDAAiD;IACjD,UAAU,EAAE,uBAAuB,CAAA;IACnC,gFAAgF;IAChF,KAAK,CAAC,EAAE,SAAS,YAAY,EAAE,CAAA;IAC/B,mFAAmF;IACnF,OAAO,CAAC,EAAE,SAAS,CAAC,MAAM,GAAG,GAAG,CAAC,EAAE,CAAA;IACnC,+DAA+D;IAC/D,OAAO,CAAC,EAAE,MAAM,CAAA;CAChB;AAED,0EAA0E;AAC1E,wBAAgB,8BAA8B,CAAC,GAAG,EAAE,8BAA8B,GAAG,kBAAkB,CA6CtG"}
+{"version":3,"file":"proposal-classifier.d.ts","sourceRoot":"","sources":["../proposal-classifier.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAIH,OAAO,KAAK,EAGX,YAAY,EACZ,kBAAkB,EAElB,MAAM,uBAAuB,CAAA;AAC9B,OAAO,KAAK,EAAE,uBAAuB,EAAE,MAAM,iBAAiB,CAAA;AAG9D,MAAM,WAAW,8BAA8B;IAC9C,wFAAwF;IACxF,EAAE,EAAE,MAAM,CAAA;IACV,iDAAiD;IACjD,UAAU,EAAE,uBAAuB,CAAA;IACnC;;;;OAIG;IACH,KAAK,CAAC,EAAE,SAAS,YAAY,EAAE,CAAA;IAC/B,mFAAmF;IACnF,OAAO,CAAC,EAAE,SAAS,CAAC,MAAM,GAAG,GAAG,CAAC,EAAE,CAAA;IACnC,+DAA+D;IAC/D,OAAO,CAAC,EAAE,MAAM,CAAA;CAChB;AAED,0EAA0E;AAC1E,wBAAgB,8BAA8B,CAAC,GAAG,EAAE,8BAA8B,GAAG,kBAAkB,CA6CtG"}

package/out/proposal-classifier.js

@@ -17,10 +17,10 @@
  *   inference is a future optimization once the policy layer has a way to invoke a classifier "once
  *   per parse" instead of per section.
  */
-import { STAGE1_COARSE_TAGS } from "./labels.js";
+import { STAGE2_TAGS } from "./labels.js";
 /** Build a `ProposalClassifier` backed by a `NeuralAddressClassifier`. */
 export function createNeuralProposalClassifier(cfg) {
-    const emits = cfg.emits ?? STAGE1_COARSE_TAGS;
+    const emits = cfg.emits ?? STAGE2_TAGS;
     const emitsSet = new Set(emits);
     const penalty = cfg.penalty ?? 0;
     async function classify(section, _ctx) {

package/out/proposal-classifier.js.map

@@ -1 +1 @@
-{"version":3,"file":"proposal-classifier.js","sourceRoot":"","sources":["../proposal-classifier.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAYH,OAAO,EAAE,kBAAkB,EAAE,MAAM,aAAa,CAAA;AAehD,0EAA0E;AAC1E,MAAM,UAAU,8BAA8B,CAAC,GAAmC;IACjF,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,IAAI,kBAAkB,CAAA;IAC7C,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAe,KAAgC,CAAC,CAAA;IACxE,MAAM,OAAO,GAAG,GAAG,CAAC,OAAO,IAAI,CAAC,CAAA;IAEhC,KAAK,UAAU,QAAQ,CAAC,OAAgB,EAAE,IAAuB;QAChE,MAAM,IAAI,GAAG,MAAM,GAAG,CAAC,UAAU,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAA;QACrD,MAAM,SAAS,GAA6B,EAAE,CAAA;QAC9C,MAAM,aAAa,GAAG,OAAO,CAAC,KAAK,CAAA;QAEnC,MAAM,KAAK,GAAG,CAAC,IAAiB,EAAQ,EAAE;YACzC,IAAI,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC5B,wFAAwF;gBACxF,oFAAoF;gBACpF,uFAAuF;gBACvF,iFAAiF;gBACjF,mFAAmF;gBACnF,2EAA2E;gBAC3E,MAAM,IAAI,GAAG;oBACZ,KAAK,EAAE,aAAa,GAAG,IAAI,CAAC,KAAK;oBACjC,GAAG,EAAE,aAAa,GAAG,IAAI,CAAC,GAAG;oBAC7B,IAAI,EAAE,IAAI,CAAC,KAAK;iBACG,CAAA;gBACpB,SAAS,CAAC,IAAI,CAAC;oBACd,IAAI;oBACJ,SAAS,EAAE,IAAI,CAAC,GAAG;oBACnB,UAAU,EAAE,IAAI,CAAC,UAAU;oBAC3B,MAAM,EAAE,QAAQ;oBAChB,SAAS,EAAE,GAAG,CAAC,EAAE;oBACjB,OAAO;iBACP,CAAC,CAAA;YACH,CAAC;YACD,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ;gBAAE,KAAK,CAAC,KAAK,CAAC,CAAA;QAChD,CAAC,CAAA;QAED,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK;YAAE,KAAK,CAAC,IAAI,CAAC,CAAA;QAC1C,OAAO,SAAS,CAAA;IACjB,CAAC;IAED,OAAO;QACN,EAAE,EAAE,GAAG,CAAC,EAAE;QACV,KAAK;QACL,OAAO,EAAE,GAAG,CAAC,OAAO,IAAI,CAAC,GAAG,CAAC;QAC7B,QAAQ;KACR,CAAA;AACF,CAAC"}
+{"version":3,"file":"proposal-classifier.js","sourceRoot":"","sources":["../proposal-classifier.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAYH,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAA;AAmBzC,0EAA0E;AAC1E,MAAM,UAAU,8BAA8B,CAAC,GAAmC;IACjF,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,IAAI,WAAW,CAAA;IACtC,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAe,KAAgC,CAAC,CAAA;IACxE,MAAM,OAAO,GAAG,GAAG,CAAC,OAAO,IAAI,CAAC,CAAA;IAEhC,KAAK,UAAU,QAAQ,CAAC,OAAgB,EAAE,IAAuB;QAChE,MAAM,IAAI,GAAG,MAAM,GAAG,CAAC,UAAU,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,CAAA;QACrD,MAAM,SAAS,GAA6B,EAAE,CAAA;QAC9C,MAAM,aAAa,GAAG,OAAO,CAAC,KAAK,CAAA;QAEnC,MAAM,KAAK,GAAG,CAAC,IAAiB,EAAQ,EAAE;YACzC,IAAI,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC5B,wFAAwF;gBACxF,oFAAoF;gBACpF,uFAAuF;gBACvF,iFAAiF;gBACjF,mFAAmF;gBACnF,2EAA2E;gBAC3E,MAAM,IAAI,GAAG;oBACZ,KAAK,EAAE,aAAa,GAAG,IAAI,CAAC,KAAK;oBACjC,GAAG,EAAE,aAAa,GAAG,IAAI,CAAC,GAAG;oBAC7B,IAAI,EAAE,IAAI,CAAC,KAAK;iBACG,CAAA;gBACpB,SAAS,CAAC,IAAI,CAAC;oBACd,IAAI;oBACJ,SAAS,EAAE,IAAI,CAAC,GAAG;oBACnB,UAAU,EAAE,IAAI,CAAC,UAAU;oBAC3B,MAAM,EAAE,QAAQ;oBAChB,SAAS,EAAE,GAAG,CAAC,EAAE;oBACjB,OAAO;iBACP,CAAC,CAAA;YACH,CAAC;YACD,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ;gBAAE,KAAK,CAAC,KAAK,CAAC,CAAA;QAChD,CAAC,CAAA;QAED,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,KAAK;YAAE,KAAK,CAAC,IAAI,CAAC,CAAA;QAC1C,OAAO,SAAS,CAAA;IACjB,CAAC;IAED,OAAO;QACN,EAAE,EAAE,GAAG,CAAC,EAAE;QACV,KAAK;QACL,OAAO,EAAE,GAAG,CAAC,OAAO,IAAI,CAAC,GAAG,CAAC;QAC7B,QAAQ;KACR,CAAA;AACF,CAAC"}

package/out/query-shape-prior.d.ts

@@ -0,0 +1,62 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Soft-prior emission biases derived from `QueryShape`.
+ *
+ *   When the QueryShape sub-system has identified a known-format span (US ZIP, UK postcode, PO box,
+ *   etc.), this module produces an additive bias matrix that nudges the encoder's per-token
+ *   emissions toward the matching BIO label. The biases compose with the structural BIO mask in the
+ *   Viterbi decoder — confident encoder predictions still win, but uncertain ones get pulled toward
+ *   the format-implied label.
+ *
+ *   Bitter-lesson-safe boundary: we don't override the encoder, just bias it. The encoder remains the
+ *   authority on context-dependent calls (the "Buffalo Wild Wings, Buffalo, NY" disambiguation);
+ *   the QueryShape prior helps on the easy cases (a 5-digit token is _probably_ a postcode).
+ *
+ *   Uses structural typing for the QueryShape input so this module has zero dependencies on
+ *   `@mailwoman/query-shape` — consumers compute the shape with that package, pass it in here.
+ */
+/**
+ * Minimal subset of `QueryShape` this module consumes. Compatible with `@mailwoman/query-shape`'s
+ * exported `QueryShape` type by shape — no import required.
+ */
+export interface QueryShapeLike {
+    knownFormats: ReadonlyArray<KnownFormatHitLike>;
+}
+export interface KnownFormatHitLike {
+    format: string;
+    span: {
+        start: number;
+        end: number;
+    };
+    /** 0..1; ambiguous patterns (e.g. 5-digit US/FR/DE overlap) score lower. */
+    confidence: number;
+}
+/** Minimal subset of `TokenizedPiece` this module consumes. */
+export interface TokenLike {
+    start: number;
+    end: number;
+}
+export interface BuildPriorsOpts {
+    /**
+     * Maximum bias magnitude (in log-odds units). Default 1.0 — adds up to ~e^1 ≈ 2.7× odds to the
+     * favored label. Confidence-scaled, so a 0.6-confidence format hit gets +0.6 max bias.
+     */
+    biasScale?: number;
+}
+/**
+ * Build a `[seqLen][numLabels]` matrix of additive log-bias to be added to encoder emissions before
+ * Viterbi decoding.
+ *
+ * For each (token, format-hit) pair where the token's character span overlaps the hit's span, the
+ * matrix entry for the format's mapped label receives `hit.confidence × biasScale`. Tokens that
+ * don't overlap any hit, or for which no label mapping exists, get 0.
+ *
+ * Returns the all-zeros matrix if `shape.knownFormats` is empty — composes harmlessly.
+ */
+export declare function buildEmissionPriors(shape: QueryShapeLike, tokens: ReadonlyArray<TokenLike>, labels: ReadonlyArray<string>, opts?: BuildPriorsOpts): number[][];
+/** Element-wise add two matrices of equal shape. Returns a new matrix. */
+export declare function addEmissionMatrix(emissions: number[][], priors: number[][]): number[][];
+//# sourceMappingURL=query-shape-prior.d.ts.map

package/out/query-shape-prior.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"query-shape-prior.d.ts","sourceRoot":"","sources":["../query-shape-prior.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAEH;;;GAGG;AACH,MAAM,WAAW,cAAc;IAC9B,YAAY,EAAE,aAAa,CAAC,kBAAkB,CAAC,CAAA;CAC/C;AAED,MAAM,WAAW,kBAAkB;IAClC,MAAM,EAAE,MAAM,CAAA;IACd,IAAI,EAAE;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,GAAG,EAAE,MAAM,CAAA;KAAE,CAAA;IACpC,4EAA4E;IAC5E,UAAU,EAAE,MAAM,CAAA;CAClB;AAED,+DAA+D;AAC/D,MAAM,WAAW,SAAS;IACzB,KAAK,EAAE,MAAM,CAAA;IACb,GAAG,EAAE,MAAM,CAAA;CACX;AAiBD,MAAM,WAAW,eAAe;IAC/B;;;OAGG;IACH,SAAS,CAAC,EAAE,MAAM,CAAA;CAClB;AAED;;;;;;;;;GASG;AACH,wBAAgB,mBAAmB,CAClC,KAAK,EAAE,cAAc,EACrB,MAAM,EAAE,aAAa,CAAC,SAAS,CAAC,EAChC,MAAM,EAAE,aAAa,CAAC,MAAM,CAAC,EAC7B,IAAI,GAAE,eAAoB,GACxB,MAAM,EAAE,EAAE,CA4BZ;AAMD,0EAA0E;AAC1E,wBAAgB,iBAAiB,CAAC,SAAS,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,EAAE,MAAM,EAAE,EAAE,GAAG,MAAM,EAAE,EAAE,CAWvF"}

package/out/query-shape-prior.js

@@ -0,0 +1,93 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Soft-prior emission biases derived from `QueryShape`.
+ *
+ *   When the QueryShape sub-system has identified a known-format span (US ZIP, UK postcode, PO box,
+ *   etc.), this module produces an additive bias matrix that nudges the encoder's per-token
+ *   emissions toward the matching BIO label. The biases compose with the structural BIO mask in the
+ *   Viterbi decoder — confident encoder predictions still win, but uncertain ones get pulled toward
+ *   the format-implied label.
+ *
+ *   Bitter-lesson-safe boundary: we don't override the encoder, just bias it. The encoder remains the
+ *   authority on context-dependent calls (the "Buffalo Wild Wings, Buffalo, NY" disambiguation);
+ *   the QueryShape prior helps on the easy cases (a 5-digit token is _probably_ a postcode).
+ *
+ *   Uses structural typing for the QueryShape input so this module has zero dependencies on
+ *   `@mailwoman/query-shape` — consumers compute the shape with that package, pass it in here.
+ */
+/**
+ * Mapping from `KnownFormat` strings to the BIO label that should be boosted. Multiple formats may
+ * map to the same label (all postcode flavors → `B-postcode`).
+ */
+const FORMAT_TO_LABEL = new Map([
+    ["us_zip", "B-postcode"],
+    ["us_zip4", "B-postcode"],
+    ["fr_postcode", "B-postcode"],
+    ["de_postcode", "B-postcode"],
+    ["uk_postcode", "B-postcode"],
+    ["ca_postcode", "B-postcode"],
+    ["jp_postcode", "B-postcode"],
+    ["po_box", "B-po_box"],
+]);
+/**
+ * Build a `[seqLen][numLabels]` matrix of additive log-bias to be added to encoder emissions before
+ * Viterbi decoding.
+ *
+ * For each (token, format-hit) pair where the token's character span overlaps the hit's span, the
+ * matrix entry for the format's mapped label receives `hit.confidence × biasScale`. Tokens that
+ * don't overlap any hit, or for which no label mapping exists, get 0.
+ *
+ * Returns the all-zeros matrix if `shape.knownFormats` is empty — composes harmlessly.
+ */
+export function buildEmissionPriors(shape, tokens, labels, opts = {}) {
+    const T = tokens.length;
+    const L = labels.length;
+    const biasScale = opts.biasScale ?? 1.0;
+    const matrix = [];
+    for (let t = 0; t < T; t++)
+        matrix.push(new Array(L).fill(0));
+    if (shape.knownFormats.length === 0)
+        return matrix;
+    // Index label → column for fast lookup.
+    const labelToCol = new Map();
+    for (let k = 0; k < labels.length; k++)
+        labelToCol.set(labels[k], k);
+    for (const hit of shape.knownFormats) {
+        const targetLabel = FORMAT_TO_LABEL.get(hit.format);
+        if (!targetLabel)
+            continue;
+        const col = labelToCol.get(targetLabel);
+        if (col === undefined)
+            continue;
+        const bias = hit.confidence * biasScale;
+        for (let t = 0; t < T; t++) {
+            const tok = tokens[t];
+            if (overlaps(tok, hit.span)) {
+                matrix[t][col] = Math.max(matrix[t][col], bias);
+            }
+        }
+    }
+    return matrix;
+}
+function overlaps(a, b) {
+    return a.start < b.end && b.start < a.end;
+}
+/** Element-wise add two matrices of equal shape. Returns a new matrix. */
+export function addEmissionMatrix(emissions, priors) {
+    if (priors.length === 0)
+        return emissions.map((row) => row.slice());
+    const out = [];
+    for (let t = 0; t < emissions.length; t++) {
+        const e = emissions[t];
+        const p = priors[t] ?? new Array(e.length).fill(0);
+        const row = new Array(e.length);
+        for (let k = 0; k < e.length; k++)
+            row[k] = e[k] + (p[k] ?? 0);
+        out.push(row);
+    }
+    return out;
+}
+//# sourceMappingURL=query-shape-prior.js.map

package/out/query-shape-prior.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"query-shape-prior.js","sourceRoot":"","sources":["../query-shape-prior.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAuBH;;;GAGG;AACH,MAAM,eAAe,GAAgC,IAAI,GAAG,CAAC;IAC5D,CAAC,QAAQ,EAAE,YAAY,CAAC;IACxB,CAAC,SAAS,EAAE,YAAY,CAAC;IACzB,CAAC,aAAa,EAAE,YAAY,CAAC;IAC7B,CAAC,aAAa,EAAE,YAAY,CAAC;IAC7B,CAAC,aAAa,EAAE,YAAY,CAAC;IAC7B,CAAC,aAAa,EAAE,YAAY,CAAC;IAC7B,CAAC,aAAa,EAAE,YAAY,CAAC;IAC7B,CAAC,QAAQ,EAAE,UAAU,CAAC;CACtB,CAAC,CAAA;AAUF;;;;;;;;;GASG;AACH,MAAM,UAAU,mBAAmB,CAClC,KAAqB,EACrB,MAAgC,EAChC,MAA6B,EAC7B,OAAwB,EAAE;IAE1B,MAAM,CAAC,GAAG,MAAM,CAAC,MAAM,CAAA;IACvB,MAAM,CAAC,GAAG,MAAM,CAAC,MAAM,CAAA;IACvB,MAAM,SAAS,GAAG,IAAI,CAAC,SAAS,IAAI,GAAG,CAAA;IACvC,MAAM,MAAM,GAAe,EAAE,CAAA;IAC7B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE;QAAE,MAAM,CAAC,IAAI,CAAC,IAAI,KAAK,CAAS,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAA;IAErE,IAAI,KAAK,CAAC,YAAY,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,MAAM,CAAA;IAElD,wCAAwC;IACxC,MAAM,UAAU,GAAG,IAAI,GAAG,EAAkB,CAAA;IAC5C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE;QAAE,UAAU,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAE,EAAE,CAAC,CAAC,CAAA;IAErE,KAAK,MAAM,GAAG,IAAI,KAAK,CAAC,YAAY,EAAE,CAAC;QACtC,MAAM,WAAW,GAAG,eAAe,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;QACnD,IAAI,CAAC,WAAW;YAAE,SAAQ;QAC1B,MAAM,GAAG,GAAG,UAAU,CAAC,GAAG,CAAC,WAAW,CAAC,CAAA;QACvC,IAAI,GAAG,KAAK,SAAS;YAAE,SAAQ;QAC/B,MAAM,IAAI,GAAG,GAAG,CAAC,UAAU,GAAG,SAAS,CAAA;QACvC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;YAC5B,MAAM,GAAG,GAAG,MAAM,CAAC,CAAC,CAAE,CAAA;YACtB,IAAI,QAAQ,CAAC,GAAG,EAAE,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;gBAC7B,MAAM,CAAC,CAAC,CAAE,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC,GAAG,CAAE,EAAE,IAAI,CAAC,CAAA;YACnD,CAAC;QACF,CAAC;IACF,CAAC;IAED,OAAO,MAAM,CAAA;AACd,CAAC;AAED,SAAS,QAAQ,CAAC,CAAiC,EAAE,CAAiC;IACrF,OAAO,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,CAAA;AAC1C,CAAC;AAED,0EAA0E;AAC1E,MAAM,UAAU,iBAAiB,CAAC,SAAqB,EAAE,MAAkB;IAC1E,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,KAAK,EAAE,CAAC,CAAA;IACnE,MAAM,GAAG,GAAe,EAAE,CAAA;IAC1B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC3C,MAAM,CAAC,GAAG,SAAS,CAAC,CAAC,CAAE,CAAA;QACvB,MAAM,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC,IAAI,IAAI,KAAK,CAAS,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QAC1D,MAAM,GAAG,GAAG,IAAI,KAAK,CAAS,CAAC,CAAC,MAAM,CAAC,CAAA;QACvC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,EAAE;YAAE,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAE,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAA;QAC/D,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;IACd,CAAC;IACD,OAAO,GAAG,CAAA;AACX,CAAC"}

package/out/tokenizer.d.ts

@@ -51,7 +51,12 @@ export declare class MailwomanTokenizer {
     private constructor();
     /** Load from a base64-encoded `tokenizer.model`. Use for in-memory / test setups. */
     static loadFromBase64(b64: string): Promise<MailwomanTokenizer>;
-    /** Load from a path to a `tokenizer.model` file on disk. Node-only convenience. */
+    /**
+     * Load from a path to a `tokenizer.model` file on disk. **Node-only** — the dynamic `node:fs`
+     * import keeps this method out of the static dependency graph so the rest of the tokenizer
+     * bundles cleanly for the browser. Calling it in a browser throws at runtime; use
+     * `loadFromBase64` (or the URL-fetching loaders in `@mailwoman/neural-web`) instead.
+     */
     static loadFromFile(modelPath: string): Promise<MailwomanTokenizer>;
     /**
      * Tokenize `text` to pieces + ids + realigned char offsets.

package/out/tokenizer.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"tokenizer.d.ts","sourceRoot":"","sources":["../tokenizer.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8BG;AAKH,4EAA4E;AAC5E,eAAO,MAAM,cAAc,WAAM,CAAA;AAEjC,0EAA0E;AAC1E,MAAM,WAAW,cAAc;IAC9B,wFAAwF;IACxF,KAAK,EAAE,MAAM,CAAA;IACb,mCAAmC;IACnC,EAAE,EAAE,MAAM,CAAA;IACV,yDAAyD;IACzD,KAAK,EAAE,MAAM,CAAA;IACb,uDAAuD;IACvD,GAAG,EAAE,MAAM,CAAA;CACX;AAED,MAAM,WAAW,YAAY;IAC5B,MAAM,EAAE,cAAc,EAAE,CAAA;IACxB,GAAG,EAAE,MAAM,EAAE,CAAA;CACb;AAED,qBAAa,kBAAkB;IACV,OAAO,CAAC,QAAQ,CAAC,SAAS;IAA9C,OAAO;IAEP,qFAAqF;WACxE,cAAc,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC,kBAAkB,CAAC;IAMrE,mFAAmF;WACtE,YAAY,CAAC,SAAS,EAAE,MAAM,GAAG,OAAO,CAAC,kBAAkB,CAAC;IAKzE;;;;;;OAMG;IACH,MAAM,CAAC,IAAI,EAAE,MAAM,GAAG,YAAY;IA2BlC,oFAAoF;IACpF,MAAM,CAAC,GAAG,EAAE,MAAM,EAAE,GAAG,UAAU,GAAG,MAAM;CAI1C"}
+{"version":3,"file":"tokenizer.d.ts","sourceRoot":"","sources":["../tokenizer.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8BG;AAIH,4EAA4E;AAC5E,eAAO,MAAM,cAAc,WAAM,CAAA;AAEjC,0EAA0E;AAC1E,MAAM,WAAW,cAAc;IAC9B,wFAAwF;IACxF,KAAK,EAAE,MAAM,CAAA;IACb,mCAAmC;IACnC,EAAE,EAAE,MAAM,CAAA;IACV,yDAAyD;IACzD,KAAK,EAAE,MAAM,CAAA;IACb,uDAAuD;IACvD,GAAG,EAAE,MAAM,CAAA;CACX;AAED,MAAM,WAAW,YAAY;IAC5B,MAAM,EAAE,cAAc,EAAE,CAAA;IACxB,GAAG,EAAE,MAAM,EAAE,CAAA;CACb;AAED,qBAAa,kBAAkB;IACV,OAAO,CAAC,QAAQ,CAAC,SAAS;IAA9C,OAAO;IAEP,qFAAqF;WACxE,cAAc,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC,kBAAkB,CAAC;IAMrE;;;;;OAKG;WACU,YAAY,CAAC,SAAS,EAAE,MAAM,GAAG,OAAO,CAAC,kBAAkB,CAAC;IAMzE;;;;;;OAMG;IACH,MAAM,CAAC,IAAI,EAAE,MAAM,GAAG,YAAY;IA2BlC,oFAAoF;IACpF,MAAM,CAAC,GAAG,EAAE,MAAM,EAAE,GAAG,UAAU,GAAG,MAAM;CAI1C"}

package/out/tokenizer.js

@@ -30,7 +30,6 @@
  *   - `loadFromFile(path)` — convenience helper that does the read + b64 + load.
  */
 import { SentencePieceProcessor } from "@sctg/sentencepiece-js";
-import { promises as fs } from "node:fs";
 /** SentencePiece's word-boundary marker (U+2581 LOWER ONE EIGHTH BLOCK). */
 export const SPACE_SENTINEL = "▁";
 export class MailwomanTokenizer {
@@ -44,9 +43,15 @@ export class MailwomanTokenizer {
         await processor.loadFromB64StringModel(b64);
         return new MailwomanTokenizer(processor);
     }
-    /** Load from a path to a `tokenizer.model` file on disk. Node-only convenience. */
+    /**
+     * Load from a path to a `tokenizer.model` file on disk. **Node-only** — the dynamic `node:fs`
+     * import keeps this method out of the static dependency graph so the rest of the tokenizer
+     * bundles cleanly for the browser. Calling it in a browser throws at runtime; use
+     * `loadFromBase64` (or the URL-fetching loaders in `@mailwoman/neural-web`) instead.
+     */
     static async loadFromFile(modelPath) {
-        const buf = await fs.readFile(modelPath);
+        const { readFile } = await import(/* webpackIgnore: true */ "node:fs/promises");
+        const buf = await readFile(modelPath);
         return MailwomanTokenizer.loadFromBase64(buf.toString("base64"));
     }
     /**

package/out/tokenizer.js.map

@@ -1 +1 @@
-{"version":3,"file":"tokenizer.js","sourceRoot":"","sources":["../tokenizer.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8BG;AAEH,OAAO,EAAE,sBAAsB,EAAE,MAAM,wBAAwB,CAAA;AAC/D,OAAO,EAAE,QAAQ,IAAI,EAAE,EAAE,MAAM,SAAS,CAAA;AAExC,4EAA4E;AAC5E,MAAM,CAAC,MAAM,cAAc,GAAG,GAAG,CAAA;AAmBjC,MAAM,OAAO,kBAAkB;IACO;IAArC,YAAqC,SAAiC;QAAjC,cAAS,GAAT,SAAS,CAAwB;IAAG,CAAC;IAE1E,qFAAqF;IACrF,MAAM,CAAC,KAAK,CAAC,cAAc,CAAC,GAAW;QACtC,MAAM,SAAS,GAAG,IAAI,sBAAsB,EAAE,CAAA;QAC9C,MAAM,SAAS,CAAC,sBAAsB,CAAC,GAAG,CAAC,CAAA;QAC3C,OAAO,IAAI,kBAAkB,CAAC,SAAS,CAAC,CAAA;IACzC,CAAC;IAED,mFAAmF;IACnF,MAAM,CAAC,KAAK,CAAC,YAAY,CAAC,SAAiB;QAC1C,MAAM,GAAG,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAA;QACxC,OAAO,kBAAkB,CAAC,cAAc,CAAC,GAAG,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAA;IACjE,CAAC;IAED;;;;;;OAMG;IACH,MAAM,CAAC,IAAY;QAClB,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,YAAY,CAAC,IAAI,CAAC,CAAA;QAChD,MAAM,GAAG,GAAG,IAAI,CAAC,SAAS,CAAC,SAAS,CAAC,IAAI,CAAC,CAAA;QAE1C,MAAM,SAAS,GAAqB,EAAE,CAAA;QACtC,IAAI,MAAM,GAAG,CAAC,CAAA;QAEd,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACxC,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAE,CAAA;YACxB,MAAM,EAAE,GAAG,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAA;YACvB,MAAM,WAAW,GAAG,KAAK,CAAC,UAAU,CAAC,cAAc,CAAC,CAAA;YACpD,MAAM,OAAO,GAAG,WAAW,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,cAAc,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,KAAK,CAAA;YAExE,IAAI,WAAW,EAAE,CAAC;gBACjB,OAAO,MAAM,GAAG,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAE,CAAC;oBAAE,MAAM,EAAE,CAAA;YAClE,CAAC;YAED,MAAM,KAAK,GAAG,MAAM,CAAA;YACpB,MAAM,IAAI,OAAO,CAAC,MAAM,CAAA;YACxB,MAAM,GAAG,GAAG,MAAM,CAAA;YAElB,SAAS,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAA;QAC1C,CAAC;QAED,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,GAAG,EAAE,CAAA;IAClC,CAAC;IAED,oFAAoF;IACpF,MAAM,CAAC,GAA0B;QAChC,MAAM,GAAG,GAAG,GAAG,YAAY,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QAClE,OAAO,IAAI,CAAC,SAAS,CAAC,SAAS,CAAC,GAAG,CAAW,CAAA;IAC/C,CAAC;CACD"}
+{"version":3,"file":"tokenizer.js","sourceRoot":"","sources":["../tokenizer.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8BG;AAEH,OAAO,EAAE,sBAAsB,EAAE,MAAM,wBAAwB,CAAA;AAE/D,4EAA4E;AAC5E,MAAM,CAAC,MAAM,cAAc,GAAG,GAAG,CAAA;AAmBjC,MAAM,OAAO,kBAAkB;IACO;IAArC,YAAqC,SAAiC;QAAjC,cAAS,GAAT,SAAS,CAAwB;IAAG,CAAC;IAE1E,qFAAqF;IACrF,MAAM,CAAC,KAAK,CAAC,cAAc,CAAC,GAAW;QACtC,MAAM,SAAS,GAAG,IAAI,sBAAsB,EAAE,CAAA;QAC9C,MAAM,SAAS,CAAC,sBAAsB,CAAC,GAAG,CAAC,CAAA;QAC3C,OAAO,IAAI,kBAAkB,CAAC,SAAS,CAAC,CAAA;IACzC,CAAC;IAED;;;;;OAKG;IACH,MAAM,CAAC,KAAK,CAAC,YAAY,CAAC,SAAiB;QAC1C,MAAM,EAAE,QAAQ,EAAE,GAAG,MAAM,MAAM,CAAC,yBAAyB,CAAC,kBAAkB,CAAC,CAAA;QAC/E,MAAM,GAAG,GAAG,MAAM,QAAQ,CAAC,SAAS,CAAC,CAAA;QACrC,OAAO,kBAAkB,CAAC,cAAc,CAAC,GAAG,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAA;IACjE,CAAC;IAED;;;;;;OAMG;IACH,MAAM,CAAC,IAAY;QAClB,MAAM,MAAM,GAAG,IAAI,CAAC,SAAS,CAAC,YAAY,CAAC,IAAI,CAAC,CAAA;QAChD,MAAM,GAAG,GAAG,IAAI,CAAC,SAAS,CAAC,SAAS,CAAC,IAAI,CAAC,CAAA;QAE1C,MAAM,SAAS,GAAqB,EAAE,CAAA;QACtC,IAAI,MAAM,GAAG,CAAC,CAAA;QAEd,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACxC,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAE,CAAA;YACxB,MAAM,EAAE,GAAG,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAA;YACvB,MAAM,WAAW,GAAG,KAAK,CAAC,UAAU,CAAC,cAAc,CAAC,CAAA;YACpD,MAAM,OAAO,GAAG,WAAW,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,cAAc,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,KAAK,CAAA;YAExE,IAAI,WAAW,EAAE,CAAC;gBACjB,OAAO,MAAM,GAAG,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAE,CAAC;oBAAE,MAAM,EAAE,CAAA;YAClE,CAAC;YAED,MAAM,KAAK,GAAG,MAAM,CAAA;YACpB,MAAM,IAAI,OAAO,CAAC,MAAM,CAAA;YACxB,MAAM,GAAG,GAAG,MAAM,CAAA;YAElB,SAAS,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAA;QAC1C,CAAC;QAED,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,GAAG,EAAE,CAAA;IAClC,CAAC;IAED,oFAAoF;IACpF,MAAM,CAAC,GAA0B;QAChC,MAAM,GAAG,GAAG,GAAG,YAAY,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QAClE,OAAO,IAAI,CAAC,SAAS,CAAC,SAAS,CAAC,GAAG,CAAW,CAAA;IAC/C,CAAC;CACD"}

package/out/viterbi.d.ts

@@ -0,0 +1,76 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Linear-chain CRF Viterbi decoder in TypeScript.
+ *
+ *   Replaces per-token argmax in the classifier when transition scores are available. Mirrors the
+ *   Python training-time / eval-time path so JS runtime decode agrees with the model card's
+ *   metrics.
+ *
+ *   Two transition matrix modes:
+ *
+ *   1. **Structural-only** (no weights changes required) — build from the BIO label vocabulary using
+ *        `buildBioTransitionMask()`. Forbids `O → I-X`, `B-X → I-Y` (X ≠ Y), and sequence-start →
+ *        `I-X`. Permits everything else. This alone prevents orphan-I decoding ("Saint Petersburg →
+ *        Petersburg" bug) at runtime — a strict improvement over argmax.
+ *   2. **Learned** (requires a future weights release that ships `crf-transitions.json`) — load the
+ *        trained transition matrix from the model card. Adds learned soft priors on top of the
+ *        structural mask. Currently not exported from the training-side ONNX bundle.
+ */
+/**
+ * Build the BIO structural transition mask given the label vocabulary in order.
+ *
+ * Rules:
+ *
+ * - `X → O` always permitted (0)
+ * - `X → B-Y` always permitted (0)
+ * - `X → I-Y` permitted only if `X` is `B-Y` or `I-Y` (0); otherwise -inf
+ *
+ * Returns a `numLabels × numLabels` matrix where `mask[from][to]` is the additive log-score (0 for
+ * permitted, NEG_INF for forbidden).
+ */
+export declare function buildBioTransitionMask(labels: readonly string[]): number[][];
+/** Returns the per-label vector of valid start-of-sequence transitions (0 or -inf). */
+export declare function buildBioStartMask(labels: readonly string[]): number[];
+/**
+ * End-of-sequence transitions. By default all labels are valid endings (returns zeros). Override if
+ * the trained model has learned end transitions.
+ */
+export declare function buildBioEndMask(labels: readonly string[]): number[];
+export interface ViterbiInput {
+    /** `emissions[t][k]` — log-emission for label k at timestep t. Pass raw logits or log-softmaxes. */
+    emissions: number[][];
+    /** `transitions[from][to]` — additive log-score. Use `buildBioTransitionMask` if unsure. */
+    transitions: number[][];
+    /** Per-label log-score for being the FIRST label. */
+    startTransitions?: number[];
+    /** Per-label log-score for being the LAST label. */
+    endTransitions?: number[];
+}
+export interface ViterbiResult {
+    /** Best label index per timestep. */
+    path: number[];
+    /** Total path score (log-prob). */
+    score: number;
+}
+/**
+ * Viterbi decode: find the highest-scoring label sequence under the CRF.
+ *
+ * Time: O(seq_len × num_labels²). Space: O(seq_len × num_labels) for the backpointer table.
+ */
+export declare function viterbi(input: ViterbiInput): ViterbiResult;
+/**
+ * Convenience: argmax over per-token softmax (existing behavior). Provided so callers can opt in to
+ * Viterbi only when transitions are available, falling back to this cleanly.
+ */
+export declare function perTokenArgmax(emissions: readonly number[][]): number[];
+/**
+ * Softmax of a logit row (returns probabilities summing to 1).
+ *
+ * Used to compute per-token confidence after Viterbi picks the label sequence — the confidence is
+ * the softmax probability of the Viterbi-chosen label at that timestep.
+ */
+export declare function softmax(row: readonly number[]): number[];
+//# sourceMappingURL=viterbi.d.ts.map

package/out/viterbi.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"viterbi.d.ts","sourceRoot":"","sources":["../viterbi.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;GAoBG;AAIH;;;;;;;;;;;GAWG;AACH,wBAAgB,sBAAsB,CAAC,MAAM,EAAE,SAAS,MAAM,EAAE,GAAG,MAAM,EAAE,EAAE,CAa5E;AAED,uFAAuF;AACvF,wBAAgB,iBAAiB,CAAC,MAAM,EAAE,SAAS,MAAM,EAAE,GAAG,MAAM,EAAE,CAErE;AAED;;;GAGG;AACH,wBAAgB,eAAe,CAAC,MAAM,EAAE,SAAS,MAAM,EAAE,GAAG,MAAM,EAAE,CAEnE;AAYD,MAAM,WAAW,YAAY;IAC5B,oGAAoG;IACpG,SAAS,EAAE,MAAM,EAAE,EAAE,CAAA;IACrB,4FAA4F;IAC5F,WAAW,EAAE,MAAM,EAAE,EAAE,CAAA;IACvB,qDAAqD;IACrD,gBAAgB,CAAC,EAAE,MAAM,EAAE,CAAA;IAC3B,oDAAoD;IACpD,cAAc,CAAC,EAAE,MAAM,EAAE,CAAA;CACzB;AAED,MAAM,WAAW,aAAa;IAC7B,qCAAqC;IACrC,IAAI,EAAE,MAAM,EAAE,CAAA;IACd,mCAAmC;IACnC,KAAK,EAAE,MAAM,CAAA;CACb;AAED;;;;GAIG;AACH,wBAAgB,OAAO,CAAC,KAAK,EAAE,YAAY,GAAG,aAAa,CA4D1D;AAED;;;GAGG;AACH,wBAAgB,cAAc,CAAC,SAAS,EAAE,SAAS,MAAM,EAAE,EAAE,GAAG,MAAM,EAAE,CAYvE;AAED;;;;;GAKG;AACH,wBAAgB,OAAO,CAAC,GAAG,EAAE,SAAS,MAAM,EAAE,GAAG,MAAM,EAAE,CAMxD"}

package/out/viterbi.js

@@ -0,0 +1,163 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Linear-chain CRF Viterbi decoder in TypeScript.
+ *
+ *   Replaces per-token argmax in the classifier when transition scores are available. Mirrors the
+ *   Python training-time / eval-time path so JS runtime decode agrees with the model card's
+ *   metrics.
+ *
+ *   Two transition matrix modes:
+ *
+ *   1. **Structural-only** (no weights changes required) — build from the BIO label vocabulary using
+ *        `buildBioTransitionMask()`. Forbids `O → I-X`, `B-X → I-Y` (X ≠ Y), and sequence-start →
+ *        `I-X`. Permits everything else. This alone prevents orphan-I decoding ("Saint Petersburg →
+ *        Petersburg" bug) at runtime — a strict improvement over argmax.
+ *   2. **Learned** (requires a future weights release that ships `crf-transitions.json`) — load the
+ *        trained transition matrix from the model card. Adds learned soft priors on top of the
+ *        structural mask. Currently not exported from the training-side ONNX bundle.
+ */
+const NEG_INF = -1e9;
+/**
+ * Build the BIO structural transition mask given the label vocabulary in order.
+ *
+ * Rules:
+ *
+ * - `X → O` always permitted (0)
+ * - `X → B-Y` always permitted (0)
+ * - `X → I-Y` permitted only if `X` is `B-Y` or `I-Y` (0); otherwise -inf
+ *
+ * Returns a `numLabels × numLabels` matrix where `mask[from][to]` is the additive log-score (0 for
+ * permitted, NEG_INF for forbidden).
+ */
+export function buildBioTransitionMask(labels) {
+    const n = labels.length;
+    const mask = [];
+    for (let from = 0; from < n; from++) {
+        const row = new Array(n);
+        const fromLabel = labels[from];
+        for (let to = 0; to < n; to++) {
+            const toLabel = labels[to];
+            row[to] = isValidTransition(fromLabel, toLabel) ? 0 : NEG_INF;
+        }
+        mask.push(row);
+    }
+    return mask;
+}
+/** Returns the per-label vector of valid start-of-sequence transitions (0 or -inf). */
+export function buildBioStartMask(labels) {
+    return labels.map((l) => (l.startsWith("I-") ? NEG_INF : 0));
+}
+/**
+ * End-of-sequence transitions. By default all labels are valid endings (returns zeros). Override if
+ * the trained model has learned end transitions.
+ */
+export function buildBioEndMask(labels) {
+    return labels.map(() => 0);
+}
+function isValidTransition(from, to) {
+    if (to === "O")
+        return true;
+    if (to.startsWith("B-"))
+        return true;
+    if (to.startsWith("I-")) {
+        const tag = to.slice(2);
+        return from === `B-${tag}` || from === `I-${tag}`;
+    }
+    return true;
+}
+/**
+ * Viterbi decode: find the highest-scoring label sequence under the CRF.
+ *
+ * Time: O(seq_len × num_labels²). Space: O(seq_len × num_labels) for the backpointer table.
+ */
+export function viterbi(input) {
+    const { emissions, transitions } = input;
+    const T = emissions.length;
+    if (T === 0)
+        return { path: [], score: 0 };
+    const numLabels = emissions[0].length;
+    const startTrans = input.startTransitions ?? new Array(numLabels).fill(0);
+    const endTrans = input.endTransitions ?? new Array(numLabels).fill(0);
+    // dp[t][k] = best log-score ending at (timestep t, label k)
+    const dp = [];
+    const back = [];
+    // t = 0
+    const first = new Array(numLabels);
+    for (let k = 0; k < numLabels; k++) {
+        first[k] = startTrans[k] + emissions[0][k];
+    }
+    dp.push(first);
+    back.push(new Array(numLabels).fill(-1));
+    for (let t = 1; t < T; t++) {
+        const cur = new Array(numLabels);
+        const ptr = new Array(numLabels);
+        for (let k = 0; k < numLabels; k++) {
+            let bestScore = NEG_INF;
+            let bestPrev = 0;
+            for (let j = 0; j < numLabels; j++) {
+                const s = dp[t - 1][j] + transitions[j][k];
+                if (s > bestScore) {
+                    bestScore = s;
+                    bestPrev = j;
+                }
+            }
+            cur[k] = bestScore + emissions[t][k];
+            ptr[k] = bestPrev;
+        }
+        dp.push(cur);
+        back.push(ptr);
+    }
+    // Pick the best ending state.
+    let bestEndScore = NEG_INF;
+    let bestEnd = 0;
+    for (let k = 0; k < numLabels; k++) {
+        const s = dp[T - 1][k] + endTrans[k];
+        if (s > bestEndScore) {
+            bestEndScore = s;
+            bestEnd = k;
+        }
+    }
+    // Trace back.
+    const path = new Array(T);
+    path[T - 1] = bestEnd;
+    for (let t = T - 1; t > 0; t--) {
+        path[t - 1] = back[t][path[t]];
+    }
+    return { path, score: bestEndScore };
+}
+/**
+ * Convenience: argmax over per-token softmax (existing behavior). Provided so callers can opt in to
+ * Viterbi only when transitions are available, falling back to this cleanly.
+ */
+export function perTokenArgmax(emissions) {
+    return emissions.map((row) => {
+        let bestIdx = 0;
+        let bestVal = row[0];
+        for (let k = 1; k < row.length; k++) {
+            if (row[k] > bestVal) {
+                bestVal = row[k];
+                bestIdx = k;
+            }
+        }
+        return bestIdx;
+    });
+}
+/**
+ * Softmax of a logit row (returns probabilities summing to 1).
+ *
+ * Used to compute per-token confidence after Viterbi picks the label sequence — the confidence is
+ * the softmax probability of the Viterbi-chosen label at that timestep.
+ */
+export function softmax(row) {
+    let max = row[0];
+    for (let i = 1; i < row.length; i++)
+        if (row[i] > max)
+            max = row[i];
+    const exps = row.map((v) => Math.exp(v - max));
+    const sum = exps.reduce((a, b) => a + b, 0);
+    return exps.map((e) => e / sum);
+}
+//# sourceMappingURL=viterbi.js.map

package/out/viterbi.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"viterbi.js","sourceRoot":"","sources":["../viterbi.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;GAoBG;AAEH,MAAM,OAAO,GAAG,CAAC,GAAG,CAAA;AAEpB;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,sBAAsB,CAAC,MAAyB;IAC/D,MAAM,CAAC,GAAG,MAAM,CAAC,MAAM,CAAA;IACvB,MAAM,IAAI,GAAe,EAAE,CAAA;IAC3B,KAAK,IAAI,IAAI,GAAG,CAAC,EAAE,IAAI,GAAG,CAAC,EAAE,IAAI,EAAE,EAAE,CAAC;QACrC,MAAM,GAAG,GAAG,IAAI,KAAK,CAAS,CAAC,CAAC,CAAA;QAChC,MAAM,SAAS,GAAG,MAAM,CAAC,IAAI,CAAE,CAAA;QAC/B,KAAK,IAAI,EAAE,GAAG,CAAC,EAAE,EAAE,GAAG,CAAC,EAAE,EAAE,EAAE,EAAE,CAAC;YAC/B,MAAM,OAAO,GAAG,MAAM,CAAC,EAAE,CAAE,CAAA;YAC3B,GAAG,CAAC,EAAE,CAAC,GAAG,iBAAiB,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAA;QAC9D,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;IACf,CAAC;IACD,OAAO,IAAI,CAAA;AACZ,CAAC;AAED,uFAAuF;AACvF,MAAM,UAAU,iBAAiB,CAAC,MAAyB;IAC1D,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;AAC7D,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,eAAe,CAAC,MAAyB;IACxD,OAAO,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAA;AAC3B,CAAC;AAED,SAAS,iBAAiB,CAAC,IAAY,EAAE,EAAU;IAClD,IAAI,EAAE,KAAK,GAAG;QAAE,OAAO,IAAI,CAAA;IAC3B,IAAI,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC;QAAE,OAAO,IAAI,CAAA;IACpC,IAAI,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC;QACzB,MAAM,GAAG,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,CAAA;QACvB,OAAO,IAAI,KAAK,KAAK,GAAG,EAAE,IAAI,IAAI,KAAK,KAAK,GAAG,EAAE,CAAA;IAClD,CAAC;IACD,OAAO,IAAI,CAAA;AACZ,CAAC;AAoBD;;;;GAIG;AACH,MAAM,UAAU,OAAO,CAAC,KAAmB;IAC1C,MAAM,EAAE,SAAS,EAAE,WAAW,EAAE,GAAG,KAAK,CAAA;IACxC,MAAM,CAAC,GAAG,SAAS,CAAC,MAAM,CAAA;IAC1B,IAAI,CAAC,KAAK,CAAC;QAAE,OAAO,EAAE,IAAI,EAAE,EAAE,EAAE,KAAK,EAAE,CAAC,EAAE,CAAA;IAE1C,MAAM,SAAS,GAAG,SAAS,CAAC,CAAC,CAAE,CAAC,MAAM,CAAA;IACtC,MAAM,UAAU,GAAG,KAAK,CAAC,gBAAgB,IAAI,IAAI,KAAK,CAAS,SAAS,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IACjF,MAAM,QAAQ,GAAG,KAAK,CAAC,cAAc,IAAI,IAAI,KAAK,CAAS,SAAS,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IAE7E,4DAA4D;IAC5D,MAAM,EAAE,GAAe,EAAE,CAAA;IACzB,MAAM,IAAI,GAAe,EAAE,CAAA;IAE3B,QAAQ;IACR,MAAM,KAAK,GAAG,IAAI,KAAK,CAAS,SAAS,CAAC,CAAA;IAC1C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;QACpC,KAAK,CAAC,CAAC,CAAC,GAAG,UAAU,CAAC,CAAC,CAAE,GAAG,SAAS,CAAC,CAAC,CAAE,CAAC,CAAC,CAAE,CAAA;IAC9C,CAAC;IACD,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;IACd,IAAI,CAAC,IAAI,CAAC,IAAI,KAAK,CAAS,SAAS,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IAEhD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC5B,MAAM,GAAG,GAAG,IAAI,KAAK,CAAS,SAAS,CAAC,CAAA;QACxC,MAAM,GAAG,GAAG,IAAI,KAAK,CAAS,SAAS,CAAC,CAAA;QACxC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;YACpC,IAAI,SAAS,GAAG,OAAO,CAAA;YACvB,IAAI,QAAQ,GAAG,CAAC,CAAA;YAChB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;gBACpC,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,CAAE,CAAC,CAAC,CAAE,GAAG,WAAW,CAAC,CAAC,CAAE,CAAC,CAAC,CAAE,CAAA;gBAC9C,IAAI,CAAC,GAAG,SAAS,EAAE,CAAC;oBACnB,SAAS,GAAG,CAAC,CAAA;oBACb,QAAQ,GAAG,CAAC,CAAA;gBACb,CAAC;YACF,CAAC;YACD,GAAG,CAAC,CAAC,CAAC,GAAG,SAAS,GAAG,SAAS,CAAC,CAAC,CAAE,CAAC,CAAC,CAAE,CAAA;YACtC,GAAG,CAAC,CAAC,CAAC,GAAG,QAAQ,CAAA;QAClB,CAAC;QACD,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;QACZ,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;IACf,CAAC;IAED,8BAA8B;IAC9B,IAAI,YAAY,GAAG,OAAO,CAAA;IAC1B,IAAI,OAAO,GAAG,CAAC,CAAA;IACf,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;QACpC,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,CAAE,CAAC,CAAC,CAAE,GAAG,QAAQ,CAAC,CAAC,CAAE,CAAA;QACvC,IAAI,CAAC,GAAG,YAAY,EAAE,CAAC;YACtB,YAAY,GAAG,CAAC,CAAA;YAChB,OAAO,GAAG,CAAC,CAAA;QACZ,CAAC;IACF,CAAC;IAED,cAAc;IACd,MAAM,IAAI,GAAG,IAAI,KAAK,CAAS,CAAC,CAAC,CAAA;IACjC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,OAAO,CAAA;IACrB,KAAK,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAChC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAE,CAAC,IAAI,CAAC,CAAC,CAAE,CAAE,CAAA;IAClC,CAAC;IAED,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,YAAY,EAAE,CAAA;AACrC,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,cAAc,CAAC,SAA8B;IAC5D,OAAO,SAAS,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE;QAC5B,IAAI,OAAO,GAAG,CAAC,CAAA;QACf,IAAI,OAAO,GAAG,GAAG,CAAC,CAAC,CAAE,CAAA;QACrB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACrC,IAAI,GAAG,CAAC,CAAC,CAAE,GAAG,OAAO,EAAE,CAAC;gBACvB,OAAO,GAAG,GAAG,CAAC,CAAC,CAAE,CAAA;gBACjB,OAAO,GAAG,CAAC,CAAA;YACZ,CAAC;QACF,CAAC;QACD,OAAO,OAAO,CAAA;IACf,CAAC,CAAC,CAAA;AACH,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,OAAO,CAAC,GAAsB;IAC7C,IAAI,GAAG,GAAG,GAAG,CAAC,CAAC,CAAE,CAAA;IACjB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,EAAE;QAAE,IAAI,GAAG,CAAC,CAAC,CAAE,GAAG,GAAG;YAAE,GAAG,GAAG,GAAG,CAAC,CAAC,CAAE,CAAA;IACrE,MAAM,IAAI,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAA;IAC9C,MAAM,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAA;IAC3C,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,GAAG,CAAC,CAAA;AAChC,CAAC"}

package/out/weights.d.ts

@@ -32,8 +32,26 @@ export interface ResolveWeightsOpts {
 export interface ResolvedWeights {
     modelPath: string;
     tokenizerPath: string;
+    /**
+     * Path to `model-card.json` alongside the resolved model. `undefined` when the caller passed
+     * explicit paths or when the package directory has no card on disk. Read by `loadFromWeights` to
+     * thread the trained label vocabulary into the classifier — see {@link readLabelsFromModelCard}.
+     */
+    modelCardPath?: string;
     /** "explicit" if both paths came from opts; "package:<name>" if resolved via require.resolve. */
     source: string;
 }
 export declare function resolveWeights(opts: ResolveWeightsOpts): ResolvedWeights;
+/**
+ * Read the `labels` array from a `model-card.json` file. Returns `undefined` when the file is
+ * missing, unreadable, malformed, or has no `labels` field — callers should fall back to their
+ * compile-time default in that case (the loader contract: the JS-side default tracks the most
+ * recent shipped stage, so a card without `labels` is always a pre-v0.4.0 card whose label vocab
+ * matches that default by construction).
+ *
+ * Validates shape: must be a non-empty array of strings. Throws on a present-but-malformed `labels`
+ * field — a card that emits e.g. `labels: 21` rather than `labels: [...]` is a corrupted artifact
+ * and should be loud, not silently re-defaulted.
+ */
+export declare function readLabelsFromModelCard(modelCardPath: string | undefined): readonly string[] | undefined;
 //# sourceMappingURL=weights.d.ts.map

package/out/weights.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"weights.d.ts","sourceRoot":"","sources":["../weights.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;GAsBG;AAQH,MAAM,WAAW,kBAAkB;IAClC,wFAAwF;IACxF,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,4EAA4E;IAC5E,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB,iFAAiF;IACjF,aAAa,CAAC,EAAE,MAAM,CAAA;CACtB;AAED,MAAM,WAAW,eAAe;IAC/B,SAAS,EAAE,MAAM,CAAA;IACjB,aAAa,EAAE,MAAM,CAAA;IACrB,iGAAiG;IACjG,MAAM,EAAE,MAAM,CAAA;CACd;AAED,wBAAgB,cAAc,CAAC,IAAI,EAAE,kBAAkB,GAAG,eAAe,CAoCxE"}
+{"version":3,"file":"weights.d.ts","sourceRoot":"","sources":["../weights.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;GAsBG;AAQH,MAAM,WAAW,kBAAkB;IAClC,wFAAwF;IACxF,MAAM,CAAC,EAAE,MAAM,CAAA;IACf,4EAA4E;IAC5E,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB,iFAAiF;IACjF,aAAa,CAAC,EAAE,MAAM,CAAA;CACtB;AAED,MAAM,WAAW,eAAe;IAC/B,SAAS,EAAE,MAAM,CAAA;IACjB,aAAa,EAAE,MAAM,CAAA;IACrB;;;;OAIG;IACH,aAAa,CAAC,EAAE,MAAM,CAAA;IACtB,iGAAiG;IACjG,MAAM,EAAE,MAAM,CAAA;CACd;AAED,wBAAgB,cAAc,CAAC,IAAI,EAAE,kBAAkB,GAAG,eAAe,CA0CxE;AAED;;;;;;;;;;GAUG;AACH,wBAAgB,uBAAuB,CAAC,aAAa,EAAE,MAAM,GAAG,SAAS,GAAG,SAAS,MAAM,EAAE,GAAG,SAAS,CAwBxG"}

package/out/weights.js

@@ -21,7 +21,7 @@
  *   The resolver checks for both files and throws a single actionable error when neither is findable,
  *   naming all the paths it tried.
  */
-import { existsSync } from "node:fs";
+import { existsSync, readFileSync } from "node:fs";
 import { createRequire } from "node:module";
 import { dirname, resolve } from "node:path";
 const req = createRequire(import.meta.url);
@@ -34,7 +34,10 @@ export function resolveWeights(opts) {
             throw new Error(`Explicit tokenizerPath does not exist: ${opts.tokenizerPath}`);
         return { modelPath: opts.modelPath, tokenizerPath: opts.tokenizerPath, source: "explicit" };
     }
-    const locale = opts.locale ?? "en-us";
+    // Package names follow the all-lowercase BCP-47 convention (`neural-weights-en-us`,
+    // `neural-weights-fr-fr`). The CLI's locale validation accepts canonical `en-US` / `fr-FR`
+    // casing, so we normalize here rather than at the callsite.
+    const locale = (opts.locale ?? "en-us").toLowerCase();
     const packageName = `@mailwoman/neural-weights-${locale}`;
     let packageDir;
     try {
@@ -54,6 +57,47 @@ export function resolveWeights(opts) {
             `Run \`scripts/link-dev-weights.sh\` inside the package to symlink dev weights, ` +
             `or pass --model + --tokenizer with explicit paths.`);
     }
-    return { modelPath, tokenizerPath, source: `package:${packageName}` };
+    const modelCardCandidate = resolve(packageDir, "model-card.json");
+    const modelCardPath = existsSync(modelCardCandidate) ? modelCardCandidate : undefined;
+    return { modelPath, tokenizerPath, modelCardPath, source: `package:${packageName}` };
+}
+/**
+ * Read the `labels` array from a `model-card.json` file. Returns `undefined` when the file is
+ * missing, unreadable, malformed, or has no `labels` field — callers should fall back to their
+ * compile-time default in that case (the loader contract: the JS-side default tracks the most
+ * recent shipped stage, so a card without `labels` is always a pre-v0.4.0 card whose label vocab
+ * matches that default by construction).
+ *
+ * Validates shape: must be a non-empty array of strings. Throws on a present-but-malformed `labels`
+ * field — a card that emits e.g. `labels: 21` rather than `labels: [...]` is a corrupted artifact
+ * and should be loud, not silently re-defaulted.
+ */
+export function readLabelsFromModelCard(modelCardPath) {
+    if (!modelCardPath || !existsSync(modelCardPath))
+        return undefined;
+    let raw;
+    try {
+        raw = readFileSync(modelCardPath, "utf8");
+    }
+    catch {
+        return undefined;
+    }
+    let parsed;
+    try {
+        parsed = JSON.parse(raw);
+    }
+    catch {
+        return undefined;
+    }
+    if (typeof parsed !== "object" || parsed === null)
+        return undefined;
+    const labels = parsed.labels;
+    if (labels === undefined)
+        return undefined;
+    if (!Array.isArray(labels) || labels.length === 0 || !labels.every((l) => typeof l === "string")) {
+        throw new Error(`model-card.json at ${modelCardPath} has a malformed \`labels\` field — ` +
+            `expected a non-empty array of strings, got ${JSON.stringify(labels)}.`);
+    }
+    return Object.freeze(labels.slice());
 }
 //# sourceMappingURL=weights.js.map

package/out/weights.js.map

@@ -1 +1 @@
-{"version":3,"file":"weights.js","sourceRoot":"","sources":["../weights.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;GAsBG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAA;AACpC,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAA;AAC3C,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAA;AAE5C,MAAM,GAAG,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;AAkB1C,MAAM,UAAU,cAAc,CAAC,IAAwB;IACtD,MAAM,KAAK,GAAa,EAAE,CAAA;IAE1B,IAAI,IAAI,CAAC,SAAS,IAAI,IAAI,CAAC,aAAa,EAAE,CAAC;QAC1C,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,SAAS,CAAC;YAAE,MAAM,IAAI,KAAK,CAAC,sCAAsC,IAAI,CAAC,SAAS,EAAE,CAAC,CAAA;QACxG,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,aAAa,CAAC;YAAE,MAAM,IAAI,KAAK,CAAC,0CAA0C,IAAI,CAAC,aAAa,EAAE,CAAC,CAAA;QACpH,OAAO,EAAE,SAAS,EAAE,IAAI,CAAC,SAAS,EAAE,aAAa,EAAE,IAAI,CAAC,aAAa,EAAE,MAAM,EAAE,UAAU,EAAE,CAAA;IAC5F,CAAC;IAED,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,IAAI,OAAO,CAAA;IACrC,MAAM,WAAW,GAAG,6BAA6B,MAAM,EAAE,CAAA;IACzD,IAAI,UAAkB,CAAA;IACtB,IAAI,CAAC;QACJ,MAAM,WAAW,GAAG,GAAG,CAAC,OAAO,CAAC,GAAG,WAAW,eAAe,CAAC,CAAA;QAC9D,UAAU,GAAG,OAAO,CAAC,WAAW,CAAC,CAAA;IAClC,CAAC;IAAC,MAAM,CAAC;QACR,MAAM,IAAI,KAAK,CACd,qBAAqB,WAAW,iCAAiC,WAAW,IAAI;YAC/E,oDAAoD,CACrD,CAAA;IACF,CAAC;IAED,MAAM,SAAS,GAAG,IAAI,CAAC,SAAS,IAAI,OAAO,CAAC,UAAU,EAAE,YAAY,CAAC,CAAA;IACrE,MAAM,aAAa,GAAG,IAAI,CAAC,aAAa,IAAI,OAAO,CAAC,UAAU,EAAE,iBAAiB,CAAC,CAAA;IAClF,KAAK,CAAC,IAAI,CAAC,SAAS,EAAE,aAAa,CAAC,CAAA;IAEpC,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,IAAI,CAAC,UAAU,CAAC,aAAa,CAAC,EAAE,CAAC;QAC1D,MAAM,IAAI,KAAK,CACd,mBAAmB,WAAW,gBAAgB,UAAU,gCAAgC;YACvF,aAAa,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI;YACnC,iFAAiF;YACjF,oDAAoD,CACrD,CAAA;IACF,CAAC;IAED,OAAO,EAAE,SAAS,EAAE,aAAa,EAAE,MAAM,EAAE,WAAW,WAAW,EAAE,EAAE,CAAA;AACtE,CAAC"}
+{"version":3,"file":"weights.js","sourceRoot":"","sources":["../weights.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;GAsBG;AAEH,OAAO,EAAE,UAAU,EAAE,YAAY,EAAE,MAAM,SAAS,CAAA;AAClD,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAA;AAC3C,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAA;AAE5C,MAAM,GAAG,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;AAwB1C,MAAM,UAAU,cAAc,CAAC,IAAwB;IACtD,MAAM,KAAK,GAAa,EAAE,CAAA;IAE1B,IAAI,IAAI,CAAC,SAAS,IAAI,IAAI,CAAC,aAAa,EAAE,CAAC;QAC1C,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,SAAS,CAAC;YAAE,MAAM,IAAI,KAAK,CAAC,sCAAsC,IAAI,CAAC,SAAS,EAAE,CAAC,CAAA;QACxG,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,aAAa,CAAC;YAAE,MAAM,IAAI,KAAK,CAAC,0CAA0C,IAAI,CAAC,aAAa,EAAE,CAAC,CAAA;QACpH,OAAO,EAAE,SAAS,EAAE,IAAI,CAAC,SAAS,EAAE,aAAa,EAAE,IAAI,CAAC,aAAa,EAAE,MAAM,EAAE,UAAU,EAAE,CAAA;IAC5F,CAAC;IAED,oFAAoF;IACpF,2FAA2F;IAC3F,4DAA4D;IAC5D,MAAM,MAAM,GAAG,CAAC,IAAI,CAAC,MAAM,IAAI,OAAO,CAAC,CAAC,WAAW,EAAE,CAAA;IACrD,MAAM,WAAW,GAAG,6BAA6B,MAAM,EAAE,CAAA;IACzD,IAAI,UAAkB,CAAA;IACtB,IAAI,CAAC;QACJ,MAAM,WAAW,GAAG,GAAG,CAAC,OAAO,CAAC,GAAG,WAAW,eAAe,CAAC,CAAA;QAC9D,UAAU,GAAG,OAAO,CAAC,WAAW,CAAC,CAAA;IAClC,CAAC;IAAC,MAAM,CAAC;QACR,MAAM,IAAI,KAAK,CACd,qBAAqB,WAAW,iCAAiC,WAAW,IAAI;YAC/E,oDAAoD,CACrD,CAAA;IACF,CAAC;IAED,MAAM,SAAS,GAAG,IAAI,CAAC,SAAS,IAAI,OAAO,CAAC,UAAU,EAAE,YAAY,CAAC,CAAA;IACrE,MAAM,aAAa,GAAG,IAAI,CAAC,aAAa,IAAI,OAAO,CAAC,UAAU,EAAE,iBAAiB,CAAC,CAAA;IAClF,KAAK,CAAC,IAAI,CAAC,SAAS,EAAE,aAAa,CAAC,CAAA;IAEpC,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,IAAI,CAAC,UAAU,CAAC,aAAa,CAAC,EAAE,CAAC;QAC1D,MAAM,IAAI,KAAK,CACd,mBAAmB,WAAW,gBAAgB,UAAU,gCAAgC;YACvF,aAAa,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI;YACnC,iFAAiF;YACjF,oDAAoD,CACrD,CAAA;IACF,CAAC;IAED,MAAM,kBAAkB,GAAG,OAAO,CAAC,UAAU,EAAE,iBAAiB,CAAC,CAAA;IACjE,MAAM,aAAa,GAAG,UAAU,CAAC,kBAAkB,CAAC,CAAC,CAAC,CAAC,kBAAkB,CAAC,CAAC,CAAC,SAAS,CAAA;IAErF,OAAO,EAAE,SAAS,EAAE,aAAa,EAAE,aAAa,EAAE,MAAM,EAAE,WAAW,WAAW,EAAE,EAAE,CAAA;AACrF,CAAC;AAED;;;;;;;;;;GAUG;AACH,MAAM,UAAU,uBAAuB,CAAC,aAAiC;IACxE,IAAI,CAAC,aAAa,IAAI,CAAC,UAAU,CAAC,aAAa,CAAC;QAAE,OAAO,SAAS,CAAA;IAClE,IAAI,GAAW,CAAA;IACf,IAAI,CAAC;QACJ,GAAG,GAAG,YAAY,CAAC,aAAa,EAAE,MAAM,CAAC,CAAA;IAC1C,CAAC;IAAC,MAAM,CAAC;QACR,OAAO,SAAS,CAAA;IACjB,CAAC;IACD,IAAI,MAAe,CAAA;IACnB,IAAI,CAAC;QACJ,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAA;IACzB,CAAC;IAAC,MAAM,CAAC;QACR,OAAO,SAAS,CAAA;IACjB,CAAC;IACD,IAAI,OAAO,MAAM,KAAK,QAAQ,IAAI,MAAM,KAAK,IAAI;QAAE,OAAO,SAAS,CAAA;IACnE,MAAM,MAAM,GAAI,MAA+B,CAAC,MAAM,CAAA;IACtD,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,SAAS,CAAA;IAC1C,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,QAAQ,CAAC,EAAE,CAAC;QAClG,MAAM,IAAI,KAAK,CACd,sBAAsB,aAAa,sCAAsC;YACxE,8CAA8C,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,GAAG,CACxE,CAAA;IACF,CAAC;IACD,OAAO,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,KAAK,EAAE,CAAsB,CAAA;AAC1D,CAAC"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@mailwoman/neural",
-  "version": "2.0.6",
+  "version": "2.2.0",
   "description": "Mailwoman neural classifier runtime: SentencePiece tokenizer + ONNX inference + decoder wiring.",
   "license": "AGPL-3.0-only",
   "repository": {
@@ -12,10 +12,12 @@
   "exports": {
     "./package.json": "./package.json",
     ".": "./out/index.js",
-    "./tokenizer": "./out/tokenizer.js"
+    "./tokenizer": "./out/tokenizer.js",
+    "./weights": "./out/weights.js",
+    "./browser": "./out/browser.js"
   },
   "dependencies": {
-    "@mailwoman/core": "2.0.6",
+    "@mailwoman/core": "2.2.0",
     "@sctg/sentencepiece-js": "^1.3.3",
     "onnxruntime-node": "^1.26.0"
   },
@@ -25,6 +27,7 @@
     "out/**/*.d.ts",
     "out/**/*.d.ts.map"
   ],
+  "sideEffects": false,
   "publishConfig": {
     "access": "public"
   }