@mailwoman/corpus 4.16.1 → 5.0.0

package/out/src/wof-json.d.ts

@@ -15,8 +15,8 @@
  *
  *   This module provides:
  *
- *   - `WofRecord`: the lightweight per-feature shape both adapters carry in their ancestry index.
- *   - `walkFeatures`: streaming directory walk → parsed `WofRecord`s (skips alt files, bad JSON,
+ *   - `WOFRecord`: the lightweight per-feature shape both adapters carry in their ancestry index.
+ *   - `walkFeatures`: streaming directory walk → parsed `WOFRecord`s (skips alt files, bad JSON,
  *       deprecated records).
  *   - `buildAncestryIndex`: in-memory ancestry chain construction (`Map<id, ancestors[]>`).
  *   - `extractNameVariants`: pulls `name:*` localized name lists off a feature's properties.
@@ -29,16 +29,16 @@
  *   silently emitted zero rows from the real corpus.
  */
 /** A WOF GeoJSON feature, as published by the per-record bundles. */
-export interface WofFeature {
+export interface WOFFeature {
     type?: string;
     id?: number | string;
     properties?: Record<string, unknown> | null;
 }
 /**
- * Lightweight in-memory shape both adapters keep per record. Geometry is intentionally dropped —
- * it's 95% of the file weight and the adapters never consult it.
+ * Lightweight in-memory shape both adapters keep per record. Geometry is intentionally dropped — it's 95% of the file
+ * weight and the adapters never consult it.
  */
-export interface WofRecord {
+export interface WOFRecord {
     id: number;
     parent_id: number | null;
     /** Canonical `wof:name` of the record. */
@@ -49,57 +49,55 @@ export interface WofRecord {
     /**
      * Localized name variants from `name:*` properties.
      *
-     * Keys are the raw `name:eng_x_preferred` form; values are the first non-empty string from the
-     * underlying array (WOF stores variants as arrays even when only one form is present). The
-     * canonical `wof:name` is NOT included here — adapters add a synthetic `"default"` slot for it.
+     * Keys are the raw `name:eng_x_preferred` form; values are the first non-empty string from the underlying array (WOF
+     * stores variants as arrays even when only one form is present). The canonical `wof:name` is NOT included here —
+     * adapters add a synthetic `"default"` slot for it.
      */
     nameVariants: Map<string, string>;
 }
 /**
  * `mz:is_current` ∈ {`1`, `-1`} → keep. `0` → drop.
  *
- * Real WOF postalcode distros tag every row `-1` ("unknown but treated as active"); the Pelias
- * importer accepts `-1` alongside `1`. The previous SpatiaLite-backed adapters filtered on `= 1`
- * only and silently emitted zero rows from the corpus — this loosened predicate is the critical
- * fix.
+ * Real WOF postalcode distros tag every row `-1` ("unknown but treated as active"); the Pelias importer accepts `-1`
+ * alongside `1`. The previous SpatiaLite-backed adapters filtered on `= 1` only and silently emitted zero rows from the
+ * corpus — this loosened predicate is the critical fix.
  */
 export declare function isCurrentFeature(props: Record<string, unknown>): boolean;
 /**
- * Pull `name:*` localized variants off a WOF feature's properties. WOF stores variants as arrays
- * (`["Saint Petersburg"]`); we lift the first non-empty string. Multiple-value variants (rare;
- * usually historical aliases) are not split into separate rows by this helper — adapters can opt in
- * by iterating the underlying array if they need it.
+ * Pull `name:*` localized variants off a WOF feature's properties. WOF stores variants as arrays (`["Saint
+ * Petersburg"]`); we lift the first non-empty string. Multiple-value variants (rare; usually historical aliases) are
+ * not split into separate rows by this helper — adapters can opt in by iterating the underlying array if they need it.
  */
 export declare function extractNameVariants(props: Record<string, unknown>): Map<string, string>;
 /**
  * Turn a `name:*` property key into a hyphen-safe suffix fragment for `source_id`.
  *
- * `"name:eng_x_colloquial"` → `"name-eng-x-colloquial"`. `:` and `_` both become `-` because both
- * collide with the existing source_id separator vocabulary and downstream consumers split on `-`.
+ * `"name:eng_x_colloquial"` → `"name-eng-x-colloquial"`. `:` and `_` both become `-` because both collide with the
+ * existing source_id separator vocabulary and downstream consumers split on `-`.
  */
 export declare function normalizeNameKey(rawKey: string): string;
 /**
- * Stream every canonical GeoJSON file under `repoDir` and yield parsed `WofRecord`s.
+ * Stream every canonical GeoJSON file under `repoDir` and yield parsed `WOFRecord`s.
  *
- * `repoDir` may point at a single cloned `whosonfirst-data-*` repo OR at a parent directory holding
- * several such repos (the corpus pipeline clones all four into a shared `wof/repos/` root and runs
- * the adapter against that root). `**\/*.geojson` walks the whole tree; `-alt-` siblings are
- * skipped since they're alternate-geometry exports, not new records.
+ * `repoDir` may point at a single cloned `whosonfirst-data-*` repo OR at a parent directory holding several such repos
+ * (the corpus pipeline clones all four into a shared `wof/repos/` root and runs the adapter against that root).
+ * `**\/*.geojson` walks the whole tree; `-alt-` siblings are skipped since they're alternate-geometry exports, not new
+ * records.
  *
- * Errors per-file (unreadable, malformed JSON, missing properties) are swallowed so one bad file
- * doesn't poison a 3 GB walk. Adapters can add stricter validation downstream if they need it.
+ * Errors per-file (unreadable, malformed JSON, missing properties) are swallowed so one bad file doesn't poison a 3 GB
+ * walk. Adapters can add stricter validation downstream if they need it.
  */
 export declare function walkFeatures(repoDir: string, opts?: {
     signal?: AbortSignal;
-}): AsyncIterable<WofRecord>;
+}): AsyncIterable<WOFRecord>;
 /**
- * Build an in-memory ancestry index: `Map<wof_id, [parent, grandparent, ...]>` walking `parent_id`
- * upward and stopping at the first missing link. A cycle guard halts at any re-visit (defensive —
- * WOF data is acyclic by construction but corrupt fixtures shouldn't infinite-loop the adapter).
+ * Build an in-memory ancestry index: `Map<wof_id, [parent, grandparent, ...]>` walking `parent_id` upward and stopping
+ * at the first missing link. A cycle guard halts at any re-visit (defensive — WOF data is acyclic by construction but
+ * corrupt fixtures shouldn't infinite-loop the adapter).
  *
- * Records whose ancestors aren't in `byId` (e.g. an FR locality whose region wasn't included in the
- * cloned repo set) get a shorter chain; the variant emission gracefully degrades.
+ * Records whose ancestors aren't in `byID` (e.g. an FR locality whose region wasn't included in the cloned repo set)
+ * get a shorter chain; the variant emission gracefully degrades.
  */
-export type AncestryIndex = Map<number, WofRecord[]>;
-export declare function buildAncestryIndex(byId: Map<number, WofRecord>): AncestryIndex;
+export type AncestryIndex = Map<number, WOFRecord[]>;
+export declare function buildAncestryIndex(byID: Map<number, WOFRecord>): AncestryIndex;
 //# sourceMappingURL=wof-json.d.ts.map

package/out/src/wof-json.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"wof-json.d.ts","sourceRoot":"","sources":["../../src/wof-json.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6BG;AAKH,qEAAqE;AACrE,MAAM,WAAW,UAAU;IAC1B,IAAI,CAAC,EAAE,MAAM,CAAA;IACb,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,CAAA;IACpB,UAAU,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,IAAI,CAAA;CAC3C;AAED;;;GAGG;AACH,MAAM,WAAW,SAAS;IACzB,EAAE,EAAE,MAAM,CAAA;IACV,SAAS,EAAE,MAAM,GAAG,IAAI,CAAA;IACxB,0CAA0C;IAC1C,IAAI,EAAE,MAAM,CAAA;IACZ,SAAS,EAAE,MAAM,CAAA;IACjB,6CAA6C;IAC7C,OAAO,EAAE,MAAM,CAAA;IACf;;;;;;OAMG;IACH,YAAY,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;CACjC;AAED;;;;;;;GAOG;AACH,wBAAgB,gBAAgB,CAAC,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,OAAO,CAIxE;AAED;;;;;GAKG;AACH,wBAAgB,mBAAmB,CAAC,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAYvF;AAED;;;;;GAKG;AACH,wBAAgB,gBAAgB,CAAC,MAAM,EAAE,MAAM,GAAG,MAAM,CAEvD;AAwCD;;;;;;;;;;GAUG;AACH,wBAAuB,YAAY,CAAC,OAAO,EAAE,MAAM,EAAE,IAAI,GAAE;IAAE,MAAM,CAAC,EAAE,WAAW,CAAA;CAAO,GAAG,aAAa,CAAC,SAAS,CAAC,CA8BlH;AAED;;;;;;;GAOG;AACH,MAAM,MAAM,aAAa,GAAG,GAAG,CAAC,MAAM,EAAE,SAAS,EAAE,CAAC,CAAA;AAEpD,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,GAAG,CAAC,MAAM,EAAE,SAAS,CAAC,GAAG,aAAa,CAiB9E"}
+{"version":3,"file":"wof-json.d.ts","sourceRoot":"","sources":["../../src/wof-json.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6BG;AAMH,qEAAqE;AACrE,MAAM,WAAW,UAAU;IAC1B,IAAI,CAAC,EAAE,MAAM,CAAA;IACb,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,CAAA;IACpB,UAAU,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,IAAI,CAAA;CAC3C;AAED;;;GAGG;AACH,MAAM,WAAW,SAAS;IACzB,EAAE,EAAE,MAAM,CAAA;IACV,SAAS,EAAE,MAAM,GAAG,IAAI,CAAA;IACxB,0CAA0C;IAC1C,IAAI,EAAE,MAAM,CAAA;IACZ,SAAS,EAAE,MAAM,CAAA;IACjB,6CAA6C;IAC7C,OAAO,EAAE,MAAM,CAAA;IACf;;;;;;OAMG;IACH,YAAY,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;CACjC;AAED;;;;;;GAMG;AACH,wBAAgB,gBAAgB,CAAC,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,OAAO,CAKxE;AAED;;;;GAIG;AACH,wBAAgB,mBAAmB,CAAC,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAevF;AAED;;;;;GAKG;AACH,wBAAgB,gBAAgB,CAAC,MAAM,EAAE,MAAM,GAAG,MAAM,CAEvD;AA4CD;;;;;;;;;;GAUG;AACH,wBAAuB,YAAY,CAAC,OAAO,EAAE,MAAM,EAAE,IAAI,GAAE;IAAE,MAAM,CAAC,EAAE,WAAW,CAAA;CAAO,GAAG,aAAa,CAAC,SAAS,CAAC,CAkClH;AAED;;;;;;;GAOG;AACH,MAAM,MAAM,aAAa,GAAG,GAAG,CAAC,MAAM,EAAE,SAAS,EAAE,CAAC,CAAA;AAEpD,wBAAgB,kBAAkB,CAAC,IAAI,EAAE,GAAG,CAAC,MAAM,EAAE,SAAS,CAAC,GAAG,aAAa,CAsB9E"}

package/out/src/wof-json.js

@@ -15,8 +15,8 @@
  *
  *   This module provides:
  *
- *   - `WofRecord`: the lightweight per-feature shape both adapters carry in their ancestry index.
- *   - `walkFeatures`: streaming directory walk → parsed `WofRecord`s (skips alt files, bad JSON,
+ *   - `WOFRecord`: the lightweight per-feature shape both adapters carry in their ancestry index.
+ *   - `walkFeatures`: streaming directory walk → parsed `WOFRecord`s (skips alt files, bad JSON,
  *       deprecated records).
  *   - `buildAncestryIndex`: in-memory ancestry chain construction (`Map<id, ancestors[]>`).
  *   - `extractNameVariants`: pulls `name:*` localized name lists off a feature's properties.
@@ -28,15 +28,14 @@
  *   treated as active"), which is why the previous SpatiaLite adapter's `is_current = 1` filter
  *   silently emitted zero rows from the real corpus.
  */
-import FastGlob from "fast-glob";
 import { readFile } from "node:fs/promises";
+import FastGlob from "fast-glob";
 /**
  * `mz:is_current` ∈ {`1`, `-1`} → keep. `0` → drop.
  *
- * Real WOF postalcode distros tag every row `-1` ("unknown but treated as active"); the Pelias
- * importer accepts `-1` alongside `1`. The previous SpatiaLite-backed adapters filtered on `= 1`
- * only and silently emitted zero rows from the corpus — this loosened predicate is the critical
- * fix.
+ * Real WOF postalcode distros tag every row `-1` ("unknown but treated as active"); the Pelias importer accepts `-1`
+ * alongside `1`. The previous SpatiaLite-backed adapters filtered on `= 1` only and silently emitted zero rows from the
+ * corpus — this loosened predicate is the critical fix.
  */
 export function isCurrentFeature(props) {
     const raw = props["mz:is_current"];
@@ -44,10 +43,9 @@ export function isCurrentFeature(props) {
     return n !== 0;
 }
 /**
- * Pull `name:*` localized variants off a WOF feature's properties. WOF stores variants as arrays
- * (`["Saint Petersburg"]`); we lift the first non-empty string. Multiple-value variants (rare;
- * usually historical aliases) are not split into separate rows by this helper — adapters can opt in
- * by iterating the underlying array if they need it.
+ * Pull `name:*` localized variants off a WOF feature's properties. WOF stores variants as arrays (`["Saint
+ * Petersburg"]`); we lift the first non-empty string. Multiple-value variants (rare; usually historical aliases) are
+ * not split into separate rows by this helper — adapters can opt in by iterating the underlying array if they need it.
  */
 export function extractNameVariants(props) {
     const out = new Map();
@@ -67,8 +65,8 @@ export function extractNameVariants(props) {
 /**
  * Turn a `name:*` property key into a hyphen-safe suffix fragment for `source_id`.
  *
- * `"name:eng_x_colloquial"` → `"name-eng-x-colloquial"`. `:` and `_` both become `-` because both
- * collide with the existing source_id separator vocabulary and downstream consumers split on `-`.
+ * `"name:eng_x_colloquial"` → `"name-eng-x-colloquial"`. `:` and `_` both become `-` because both collide with the
+ * existing source_id separator vocabulary and downstream consumers split on `-`.
  */
 export function normalizeNameKey(rawKey) {
     return rawKey.replace(/[:_]/g, "-");
@@ -78,8 +76,8 @@ function recordFromFeature(feature) {
     if (!feature || feature.type !== "Feature" || !feature.properties)
         return null;
     const props = feature.properties;
-    const rawId = typeof feature.id === "number" ? feature.id : props["wof:id"];
-    const id = typeof rawId === "number" ? rawId : typeof rawId === "string" ? Number(rawId) : NaN;
+    const rawID = typeof feature.id === "number" ? feature.id : props["wof:id"];
+    const id = typeof rawID === "number" ? rawID : typeof rawID === "string" ? Number(rawID) : NaN;
     if (!Number.isFinite(id))
         return null;
     const name = props["wof:name"];
@@ -109,15 +107,15 @@ function recordFromFeature(feature) {
     };
 }
 /**
- * Stream every canonical GeoJSON file under `repoDir` and yield parsed `WofRecord`s.
+ * Stream every canonical GeoJSON file under `repoDir` and yield parsed `WOFRecord`s.
  *
- * `repoDir` may point at a single cloned `whosonfirst-data-*` repo OR at a parent directory holding
- * several such repos (the corpus pipeline clones all four into a shared `wof/repos/` root and runs
- * the adapter against that root). `**\/*.geojson` walks the whole tree; `-alt-` siblings are
- * skipped since they're alternate-geometry exports, not new records.
+ * `repoDir` may point at a single cloned `whosonfirst-data-*` repo OR at a parent directory holding several such repos
+ * (the corpus pipeline clones all four into a shared `wof/repos/` root and runs the adapter against that root).
+ * `**\/*.geojson` walks the whole tree; `-alt-` siblings are skipped since they're alternate-geometry exports, not new
+ * records.
  *
- * Errors per-file (unreadable, malformed JSON, missing properties) are swallowed so one bad file
- * doesn't poison a 3 GB walk. Adapters can add stricter validation downstream if they need it.
+ * Errors per-file (unreadable, malformed JSON, missing properties) are swallowed so one bad file doesn't poison a 3 GB
+ * walk. Adapters can add stricter validation downstream if they need it.
  */
 export async function* walkFeatures(repoDir, opts = {}) {
     const stream = FastGlob.stream(["**/*.geojson"], {
@@ -151,14 +149,14 @@ export async function* walkFeatures(repoDir, opts = {}) {
             yield rec;
     }
 }
-export function buildAncestryIndex(byId) {
+export function buildAncestryIndex(byID) {
     const index = new Map();
-    for (const [id, rec] of byId) {
+    for (const [id, rec] of byID) {
         const chain = [];
         const guard = new Set([id]);
         let cur = rec.parent_id;
         while (cur !== null && cur > 0) {
-            const parent = byId.get(cur);
+            const parent = byID.get(cur);
             if (!parent)
                 break;
             if (guard.has(parent.id))

package/out/src/wof-json.js.map

@@ -1 +1 @@
-{"version":3,"file":"wof-json.js","sourceRoot":"","sources":["../../src/wof-json.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6BG;AAEH,OAAO,QAAQ,MAAM,WAAW,CAAA;AAChC,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAA;AA+B3C;;;;;;;GAOG;AACH,MAAM,UAAU,gBAAgB,CAAC,KAA8B;IAC9D,MAAM,GAAG,GAAG,KAAK,CAAC,eAAe,CAAC,CAAA;IAClC,MAAM,CAAC,GAAG,OAAO,GAAG,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,GAAG,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IACnF,OAAO,CAAC,KAAK,CAAC,CAAA;AACf,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,mBAAmB,CAAC,KAA8B;IACjE,MAAM,GAAG,GAAG,IAAI,GAAG,EAAkB,CAAA;IACrC,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC;QAClD,IAAI,CAAC,GAAG,CAAC,UAAU,CAAC,OAAO,CAAC;YAAE,SAAQ;QACtC,MAAM,SAAS,GAAG,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC;YACrC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAe,EAAE,CAAC,OAAO,CAAC,KAAK,QAAQ,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC;YAC9E,CAAC,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC;gBACrD,CAAC,CAAC,KAAK;gBACP,CAAC,CAAC,SAAS,CAAA;QACb,IAAI,SAAS;YAAE,GAAG,CAAC,GAAG,CAAC,GAAG,EAAE,SAAS,CAAC,IAAI,EAAE,CAAC,CAAA;IAC9C,CAAC;IACD,OAAO,GAAG,CAAA;AACX,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,gBAAgB,CAAC,MAAc;IAC9C,OAAO,MAAM,CAAC,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC,CAAA;AACpC,CAAC;AAED,0FAA0F;AAC1F,SAAS,iBAAiB,CAAC,OAAmB;IAC7C,IAAI,CAAC,OAAO,IAAI,OAAO,CAAC,IAAI,KAAK,SAAS,IAAI,CAAC,OAAO,CAAC,UAAU;QAAE,OAAO,IAAI,CAAA;IAC9E,MAAM,KAAK,GAAG,OAAO,CAAC,UAAU,CAAA;IAEhC,MAAM,KAAK,GAAG,OAAO,OAAO,CAAC,EAAE,KAAK,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAA;IAC3E,MAAM,EAAE,GAAG,OAAO,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAA;IAC9F,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAC;QAAE,OAAO,IAAI,CAAA;IAErC,MAAM,IAAI,GAAG,KAAK,CAAC,UAAU,CAAC,CAAA;IAC9B,IAAI,OAAO,IAAI,KAAK,QAAQ,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE;QAAE,OAAO,IAAI,CAAA;IAEzD,MAAM,SAAS,GAAG,KAAK,CAAC,eAAe,CAAC,CAAA;IACxC,IAAI,OAAO,SAAS,KAAK,QAAQ,IAAI,CAAC,SAAS;QAAE,OAAO,IAAI,CAAA;IAE5D,MAAM,OAAO,GAAG,KAAK,CAAC,aAAa,CAAC,CAAA;IACpC,IAAI,OAAO,OAAO,KAAK,QAAQ,IAAI,CAAC,OAAO;QAAE,OAAO,IAAI,CAAA;IAExD,IAAI,CAAC,gBAAgB,CAAC,KAAK,CAAC;QAAE,OAAO,IAAI,CAAA;IAEzC,MAAM,SAAS,GAAG,KAAK,CAAC,eAAe,CAAC,CAAA;IACxC,MAAM,SAAS,GACd,OAAO,SAAS,KAAK,QAAQ;QAC5B,CAAC,CAAC,SAAS;QACX,CAAC,CAAC,OAAO,SAAS,KAAK,QAAQ,IAAI,SAAS,CAAC,IAAI,EAAE;YAClD,CAAC,CAAC,MAAM,CAAC,SAAS,CAAC;YACnB,CAAC,CAAC,IAAI,CAAA;IAET,OAAO;QACN,EAAE;QACF,SAAS,EAAE,MAAM,CAAC,QAAQ,CAAC,SAAmB,CAAC,CAAC,CAAC,CAAE,SAAoB,CAAC,CAAC,CAAC,IAAI;QAC9E,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE;QACjB,SAAS;QACT,OAAO;QACP,YAAY,EAAE,mBAAmB,CAAC,KAAK,CAAC;KACxC,CAAA;AACF,CAAC;AAED;;;;;;;;;;GAUG;AACH,MAAM,CAAC,KAAK,SAAS,CAAC,CAAC,YAAY,CAAC,OAAe,EAAE,OAAiC,EAAE;IACvF,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,cAAc,CAAC,EAAE;QAChD,GAAG,EAAE,OAAO;QACZ,QAAQ,EAAE,IAAI;QACd,SAAS,EAAE,IAAI;QACf,cAAc,EAAE,IAAI;KACpB,CAAC,CAAA;IAEF,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAClC,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;YAAE,OAAM;QAChC,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,CAAA;QAC9B,IAAI,QAAQ,CAAC,QAAQ,CAAC,OAAO,CAAC;YAAE,SAAQ;QAExC,IAAI,IAAY,CAAA;QAChB,IAAI,CAAC;YACJ,IAAI,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAA;QACxC,CAAC;QAAC,MAAM,CAAC;YACR,SAAQ;QACT,CAAC;QAED,IAAI,MAAkB,CAAA;QACtB,IAAI,CAAC;YACJ,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAe,CAAA;QACxC,CAAC;QAAC,MAAM,CAAC;YACR,SAAQ;QACT,CAAC;QAED,MAAM,GAAG,GAAG,iBAAiB,CAAC,MAAM,CAAC,CAAA;QACrC,IAAI,GAAG;YAAE,MAAM,GAAG,CAAA;IACnB,CAAC;AACF,CAAC;AAYD,MAAM,UAAU,kBAAkB,CAAC,IAA4B;IAC9D,MAAM,KAAK,GAAkB,IAAI,GAAG,EAAE,CAAA;IACtC,KAAK,MAAM,CAAC,EAAE,EAAE,GAAG,CAAC,IAAI,IAAI,EAAE,CAAC;QAC9B,MAAM,KAAK,GAAgB,EAAE,CAAA;QAC7B,MAAM,KAAK,GAAG,IAAI,GAAG,CAAS,CAAC,EAAE,CAAC,CAAC,CAAA;QACnC,IAAI,GAAG,GAAkB,GAAG,CAAC,SAAS,CAAA;QACtC,OAAO,GAAG,KAAK,IAAI,IAAI,GAAG,GAAG,CAAC,EAAE,CAAC;YAChC,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;YAC5B,IAAI,CAAC,MAAM;gBAAE,MAAK;YAClB,IAAI,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC;gBAAE,MAAK;YAC/B,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YAClB,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAA;YACpB,GAAG,GAAG,MAAM,CAAC,SAAS,CAAA;QACvB,CAAC;QACD,KAAK,CAAC,GAAG,CAAC,EAAE,EAAE,KAAK,CAAC,CAAA;IACrB,CAAC;IACD,OAAO,KAAK,CAAA;AACb,CAAC"}
+{"version":3,"file":"wof-json.js","sourceRoot":"","sources":["../../src/wof-json.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6BG;AAEH,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAA;AAE3C,OAAO,QAAQ,MAAM,WAAW,CAAA;AA+BhC;;;;;;GAMG;AACH,MAAM,UAAU,gBAAgB,CAAC,KAA8B;IAC9D,MAAM,GAAG,GAAG,KAAK,CAAC,eAAe,CAAC,CAAA;IAClC,MAAM,CAAC,GAAG,OAAO,GAAG,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,GAAG,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IAEnF,OAAO,CAAC,KAAK,CAAC,CAAA;AACf,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,mBAAmB,CAAC,KAA8B;IACjE,MAAM,GAAG,GAAG,IAAI,GAAG,EAAkB,CAAA;IAErC,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC;QAClD,IAAI,CAAC,GAAG,CAAC,UAAU,CAAC,OAAO,CAAC;YAAE,SAAQ;QACtC,MAAM,SAAS,GAAG,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC;YACrC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAe,EAAE,CAAC,OAAO,CAAC,KAAK,QAAQ,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC;YAC9E,CAAC,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC;gBACrD,CAAC,CAAC,KAAK;gBACP,CAAC,CAAC,SAAS,CAAA;QAEb,IAAI,SAAS;YAAE,GAAG,CAAC,GAAG,CAAC,GAAG,EAAE,SAAS,CAAC,IAAI,EAAE,CAAC,CAAA;IAC9C,CAAC;IAED,OAAO,GAAG,CAAA;AACX,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,gBAAgB,CAAC,MAAc;IAC9C,OAAO,MAAM,CAAC,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC,CAAA;AACpC,CAAC;AAED,0FAA0F;AAC1F,SAAS,iBAAiB,CAAC,OAAmB;IAC7C,IAAI,CAAC,OAAO,IAAI,OAAO,CAAC,IAAI,KAAK,SAAS,IAAI,CAAC,OAAO,CAAC,UAAU;QAAE,OAAO,IAAI,CAAA;IAC9E,MAAM,KAAK,GAAG,OAAO,CAAC,UAAU,CAAA;IAEhC,MAAM,KAAK,GAAG,OAAO,OAAO,CAAC,EAAE,KAAK,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAA;IAC3E,MAAM,EAAE,GAAG,OAAO,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAA;IAE9F,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAC;QAAE,OAAO,IAAI,CAAA;IAErC,MAAM,IAAI,GAAG,KAAK,CAAC,UAAU,CAAC,CAAA;IAE9B,IAAI,OAAO,IAAI,KAAK,QAAQ,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE;QAAE,OAAO,IAAI,CAAA;IAEzD,MAAM,SAAS,GAAG,KAAK,CAAC,eAAe,CAAC,CAAA;IAExC,IAAI,OAAO,SAAS,KAAK,QAAQ,IAAI,CAAC,SAAS;QAAE,OAAO,IAAI,CAAA;IAE5D,MAAM,OAAO,GAAG,KAAK,CAAC,aAAa,CAAC,CAAA;IAEpC,IAAI,OAAO,OAAO,KAAK,QAAQ,IAAI,CAAC,OAAO;QAAE,OAAO,IAAI,CAAA;IAExD,IAAI,CAAC,gBAAgB,CAAC,KAAK,CAAC;QAAE,OAAO,IAAI,CAAA;IAEzC,MAAM,SAAS,GAAG,KAAK,CAAC,eAAe,CAAC,CAAA;IACxC,MAAM,SAAS,GACd,OAAO,SAAS,KAAK,QAAQ;QAC5B,CAAC,CAAC,SAAS;QACX,CAAC,CAAC,OAAO,SAAS,KAAK,QAAQ,IAAI,SAAS,CAAC,IAAI,EAAE;YAClD,CAAC,CAAC,MAAM,CAAC,SAAS,CAAC;YACnB,CAAC,CAAC,IAAI,CAAA;IAET,OAAO;QACN,EAAE;QACF,SAAS,EAAE,MAAM,CAAC,QAAQ,CAAC,SAAmB,CAAC,CAAC,CAAC,CAAE,SAAoB,CAAC,CAAC,CAAC,IAAI;QAC9E,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE;QACjB,SAAS;QACT,OAAO;QACP,YAAY,EAAE,mBAAmB,CAAC,KAAK,CAAC;KACxC,CAAA;AACF,CAAC;AAED;;;;;;;;;;GAUG;AACH,MAAM,CAAC,KAAK,SAAS,CAAC,CAAC,YAAY,CAAC,OAAe,EAAE,OAAiC,EAAE;IACvF,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,cAAc,CAAC,EAAE;QAChD,GAAG,EAAE,OAAO;QACZ,QAAQ,EAAE,IAAI;QACd,SAAS,EAAE,IAAI;QACf,cAAc,EAAE,IAAI;KACpB,CAAC,CAAA;IAEF,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAClC,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;YAAE,OAAM;QAChC,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,CAAA;QAE9B,IAAI,QAAQ,CAAC,QAAQ,CAAC,OAAO,CAAC;YAAE,SAAQ;QAExC,IAAI,IAAY,CAAA;QAEhB,IAAI,CAAC;YACJ,IAAI,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAA;QACxC,CAAC;QAAC,MAAM,CAAC;YACR,SAAQ;QACT,CAAC;QAED,IAAI,MAAkB,CAAA;QAEtB,IAAI,CAAC;YACJ,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAe,CAAA;QACxC,CAAC;QAAC,MAAM,CAAC;YACR,SAAQ;QACT,CAAC;QAED,MAAM,GAAG,GAAG,iBAAiB,CAAC,MAAM,CAAC,CAAA;QAErC,IAAI,GAAG;YAAE,MAAM,GAAG,CAAA;IACnB,CAAC;AACF,CAAC;AAYD,MAAM,UAAU,kBAAkB,CAAC,IAA4B;IAC9D,MAAM,KAAK,GAAkB,IAAI,GAAG,EAAE,CAAA;IAEtC,KAAK,MAAM,CAAC,EAAE,EAAE,GAAG,CAAC,IAAI,IAAI,EAAE,CAAC;QAC9B,MAAM,KAAK,GAAgB,EAAE,CAAA;QAC7B,MAAM,KAAK,GAAG,IAAI,GAAG,CAAS,CAAC,EAAE,CAAC,CAAC,CAAA;QACnC,IAAI,GAAG,GAAkB,GAAG,CAAC,SAAS,CAAA;QAEtC,OAAO,GAAG,KAAK,IAAI,IAAI,GAAG,GAAG,CAAC,EAAE,CAAC;YAChC,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;YAE5B,IAAI,CAAC,MAAM;gBAAE,MAAK;YAElB,IAAI,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC;gBAAE,MAAK;YAC/B,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YAClB,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAA;YACpB,GAAG,GAAG,MAAM,CAAC,SAAS,CAAA;QACvB,CAAC;QACD,KAAK,CAAC,GAAG,CAAC,EAAE,EAAE,KAAK,CAAC,CAAA;IACrB,CAAC;IAED,OAAO,KAAK,CAAA;AACb,CAAC"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@mailwoman/corpus",
-  "version": "4.16.1",
+  "version": "5.0.0",
   "description": "Mailwoman corpus pipeline: BIO-labeled dataset builder for the neural classifier.",
   "license": "AGPL-3.0-only",
   "repository": {
@@ -8,6 +8,12 @@
     "url": "https://github.com/sister-software/mailwoman.git",
     "directory": "corpus"
   },
+  "files": [
+    "out/**/*.js",
+    "out/**/*.js.map",
+    "out/**/*.d.ts",
+    "out/**/*.d.ts.map"
+  ],
   "type": "module",
   "exports": {
     "./package.json": "./package.json",
@@ -15,27 +21,21 @@
     "./types": "./out/src/types.js",
     "./format": "./out/src/format.js"
   },
+  "publishConfig": {
+    "access": "public"
+  },
   "dependencies": {
     "@dsnp/parquetjs": "1.8.7",
-    "@mailwoman/codex": "4.16.1",
-    "@mailwoman/core": "4.16.1",
-    "@mailwoman/formatter": "4.16.1",
+    "@mailwoman/codex": "5.0.0",
+    "@mailwoman/core": "5.0.0",
+    "@mailwoman/formatter": "5.0.0",
     "csv-parse": "^5.6.0",
     "fast-glob": "^3.3.3",
     "fastest-levenshtein": "^1.0.16",
     "lru-cache": "^10.4.3",
-    "spliterator": "^2.2.1"
+    "spliterator": "^3.1.0"
   },
   "devDependencies": {
     "@types/node": ">=24.18.0"
-  },
-  "files": [
-    "out/**/*.js",
-    "out/**/*.js.map",
-    "out/**/*.d.ts",
-    "out/**/*.d.ts.map"
-  ],
-  "publishConfig": {
-    "access": "public"
   }
 }