@mailwoman/corpus 5.2.0 → 5.3.0

package/out/src/shard-recipes/cz-pcfirst-preposition.d.ts

@@ -0,0 +1,18 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   `cz-pcfirst-preposition` — the Czech pc-first prepositional-locality shard, fourth orthography
+ *   of the #901 leading-name-boundary family. The #897 close-out read all 8 residual CZ rows as one
+ *   class: a LEADING postcode mis-assigned as house_number while the multi-word "nad/pod/u X"
+ *   locality shatters ("51244 Rokytnice nad Jizerou, Dolní Rokytnice 111" → street
+ *   'RokytnicenadJizerou' + house_number '51244'). That leading-5-digit confusion is the
+ *   anchor-pollution class whose decode-time OVERRIDE was correctly killed in #723 — this shard is
+ *   the model-first fix as DATA: real prepositional localities in the order that breaks, so the
+ *   model learns that a leading postcode before a multi-word name is a postcode. pc-first leads the
+ *   cycle (the lesson); canonical and city-first keep the polarity balanced (the v1.9.9 lesson).
+ */
+import { type ShardRecipe } from "./scaffold.js";
+export declare const czPcFirstPrepositionRecipe: ShardRecipe;
+//# sourceMappingURL=cz-pcfirst-preposition.d.ts.map

package/out/src/shard-recipes/cz-pcfirst-preposition.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"cz-pcfirst-preposition.d.ts","sourceRoot":"","sources":["../../../src/shard-recipes/cz-pcfirst-preposition.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAEH,OAAO,EAA6C,KAAK,WAAW,EAAiB,MAAM,eAAe,CAAA;AAE1G,eAAO,MAAM,0BAA0B,EAAE,WA6DxC,CAAA"}

package/out/src/shard-recipes/cz-pcfirst-preposition.js

@@ -0,0 +1,76 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   `cz-pcfirst-preposition` — the Czech pc-first prepositional-locality shard, fourth orthography
+ *   of the #901 leading-name-boundary family. The #897 close-out read all 8 residual CZ rows as one
+ *   class: a LEADING postcode mis-assigned as house_number while the multi-word "nad/pod/u X"
+ *   locality shatters ("51244 Rokytnice nad Jizerou, Dolní Rokytnice 111" → street
+ *   'RokytnicenadJizerou' + house_number '51244'). That leading-5-digit confusion is the
+ *   anchor-pollution class whose decode-time OVERRIDE was correctly killed in #723 — this shard is
+ *   the model-first fix as DATA: real prepositional localities in the order that breaks, so the
+ *   model learns that a leading postcode before a multi-word name is a postcode. pc-first leads the
+ *   cycle (the lesson); canonical and city-first keep the polarity balanced (the v1.9.9 lesson).
+ */
+import { alignAndWrite, makeMulberry32, readTuples, shardSourceID } from "./scaffold.js";
+export const czPcFirstPrepositionRecipe = {
+    name: "cz-pcfirst-preposition",
+    description: "CZ pc-first + prepositional locality (#901 family): '«pc» «city nad X», «st» «n»' — the #723 class as data",
+    mode: "tuples",
+    async run(opts, write) {
+        makeMulberry32(opts.seed);
+        let read = 0;
+        let emitted = 0;
+        let skipped = 0;
+        for await (const t of readTuples(opts.input)) {
+            read++;
+            const street = String(t.street ?? "").trim();
+            const city = String(t.locality ?? "").trim();
+            const number = String(t.number ?? "").trim();
+            const postcode = String(t.postcode ?? "").trim();
+            if (!street || !city || !number || !postcode) {
+                skipped++;
+                continue;
+            }
+            const order = read % 3;
+            let raw;
+            const components = {
+                street,
+                house_number: number,
+                postcode,
+                locality: city,
+            };
+            if (order === 0) {
+                raw = `${postcode} ${city}, ${street} ${number}`;
+            }
+            else if (order === 1) {
+                raw = `${street} ${number}, ${postcode} ${city}`;
+            }
+            else {
+                raw = `${city}, ${postcode}, ${street} ${number}`;
+            }
+            const source_id = shardSourceID("synth-cz-pcfirst-preposition", {
+                ...components,
+                o: String(order),
+                v: String(read),
+            });
+            const canonical = {
+                raw,
+                components,
+                country: "CZ",
+                locale: "cs-CZ",
+                source: "synth-cz-pcfirst-preposition",
+                source_id,
+                corpus_version: "0.10.0",
+                license: "Synthetic — cz-pcfirst-preposition; (street, number, postcode, city) from OpenAddresses CZ (per-source attribution in the model card)",
+            };
+            if (alignAndWrite(write, canonical, "cz-pcfirst-preposition"))
+                emitted++;
+            else
+                skipped++;
+        }
+        return { read, emitted, skipped };
+    },
+};
+//# sourceMappingURL=cz-pcfirst-preposition.js.map

package/out/src/shard-recipes/cz-pcfirst-preposition.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"cz-pcfirst-preposition.js","sourceRoot":"","sources":["../../../src/shard-recipes/cz-pcfirst-preposition.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAEH,OAAO,EAAE,aAAa,EAAE,cAAc,EAAE,UAAU,EAAoB,aAAa,EAAE,MAAM,eAAe,CAAA;AAE1G,MAAM,CAAC,MAAM,0BAA0B,GAAgB;IACtD,IAAI,EAAE,wBAAwB;IAC9B,WAAW,EACV,4GAA4G;IAC7G,IAAI,EAAE,QAAQ;IACd,KAAK,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK;QACpB,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QACzB,IAAI,IAAI,GAAG,CAAC,CAAA;QACZ,IAAI,OAAO,GAAG,CAAC,CAAA;QACf,IAAI,OAAO,GAAG,CAAC,CAAA;QAEf,IAAI,KAAK,EAAE,MAAM,CAAC,IAAI,UAAU,CAAC,IAAI,CAAC,KAAM,CAAC,EAAE,CAAC;YAC/C,IAAI,EAAE,CAAA;YACN,MAAM,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;YAC5C,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;YAC5C,MAAM,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;YAC5C,MAAM,QAAQ,GAAG,MAAM,CAAC,CAAC,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;YAEhD,IAAI,CAAC,MAAM,IAAI,CAAC,IAAI,IAAI,CAAC,MAAM,IAAI,CAAC,QAAQ,EAAE,CAAC;gBAC9C,OAAO,EAAE,CAAA;gBACT,SAAQ;YACT,CAAC;YACD,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,CAAA;YACtB,IAAI,GAAW,CAAA;YACf,MAAM,UAAU,GAA2B;gBAC1C,MAAM;gBACN,YAAY,EAAE,MAAM;gBACpB,QAAQ;gBACR,QAAQ,EAAE,IAAI;aACd,CAAA;YAED,IAAI,KAAK,KAAK,CAAC,EAAE,CAAC;gBACjB,GAAG,GAAG,GAAG,QAAQ,IAAI,IAAI,KAAK,MAAM,IAAI,MAAM,EAAE,CAAA;YACjD,CAAC;iBAAM,IAAI,KAAK,KAAK,CAAC,EAAE,CAAC;gBACxB,GAAG,GAAG,GAAG,MAAM,IAAI,MAAM,KAAK,QAAQ,IAAI,IAAI,EAAE,CAAA;YACjD,CAAC;iBAAM,CAAC;gBACP,GAAG,GAAG,GAAG,IAAI,KAAK,QAAQ,KAAK,MAAM,IAAI,MAAM,EAAE,CAAA;YAClD,CAAC;YACD,MAAM,SAAS,GAAG,aAAa,CAAC,8BAA8B,EAAE;gBAC/D,GAAG,UAAU;gBACb,CAAC,EAAE,MAAM,CAAC,KAAK,CAAC;gBAChB,CAAC,EAAE,MAAM,CAAC,IAAI,CAAC;aACf,CAAC,CAAA;YACF,MAAM,SAAS,GAAG;gBACjB,GAAG;gBACH,UAAU;gBACV,OAAO,EAAE,IAAI;gBACb,MAAM,EAAE,OAAO;gBACf,MAAM,EAAE,8BAA8B;gBACtC,SAAS;gBACT,cAAc,EAAE,QAAQ;gBACxB,OAAO,EACN,uIAAuI;aACxI,CAAA;YAED,IAAI,aAAa,CAAC,KAAK,EAAE,SAAS,EAAE,wBAAwB,CAAC;gBAAE,OAAO,EAAE,CAAA;;gBACnE,OAAO,EAAE,CAAA;QACf,CAAC;QAED,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,CAAA;IAClC,CAAC;CACD,CAAA"}

package/out/src/shard-recipes/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/shard-recipes/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAeH,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,eAAe,CAAA;AAMhD,cAAc,eAAe,CAAA;AAuB7B,4BAA4B;AAC5B,eAAO,MAAM,aAAa,EAAE,WAAW,CAAC,MAAM,EAAE,WAAW,CAA4C,CAAA;AAEvG,+CAA+C;AAC/C,wBAAgB,cAAc,CAAC,IAAI,EAAE,MAAM,GAAG,WAAW,GAAG,SAAS,CAEpE;AAED,yCAAyC;AACzC,wBAAgB,gBAAgB,IAAI,SAAS,WAAW,EAAE,CAEzD"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/shard-recipes/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAiBH,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,eAAe,CAAA;AAOhD,cAAc,eAAe,CAAA;AA0B7B,4BAA4B;AAC5B,eAAO,MAAM,aAAa,EAAE,WAAW,CAAC,MAAM,EAAE,WAAW,CAA4C,CAAA;AAEvG,+CAA+C;AAC/C,wBAAgB,cAAc,CAAC,IAAI,EAAE,MAAM,GAAG,WAAW,GAAG,SAAS,CAEpE;AAED,yCAAyC;AACzC,wBAAgB,gBAAgB,IAAI,SAAS,WAAW,EAAE,CAEzD"}

package/out/src/shard-recipes/index.js

@@ -10,6 +10,7 @@
 import { anchorAbsorptionRecipe } from "./anchor-absorption.js";
 import { boundaryStressRecipe } from "./boundary-stress.js";
 import { countryBalancedRecipe } from "./country-balanced.js";
+import { czPcFirstPrepositionRecipe } from "./cz-pcfirst-preposition.js";
 import { frAdminSplitRecipe } from "./fr-admin-split.js";
 import { frBareStreetRecipe } from "./fr-bare-street.js";
 import { frOrderRecipe } from "./fr-order.js";
@@ -17,9 +18,11 @@ import { germanRecipe } from "./german.js";
 import { houseVenueRecipe } from "./house-venue.js";
 import { intersectionRecipe } from "./intersection.js";
 import { localeRecipe } from "./locale.js";
+import { noStreetLedRecipe } from "./no-street-led.js";
 import { noStreetRecipe } from "./no-street.js";
 import { poBoxCedexRecipe } from "./po-box-cedex.js";
 import { poBoxRecipe } from "./po-box.js";
+import { siBareVillageRecipe } from "./si-bare-village.js";
 import { streetAffixRecipe } from "./street-affix.js";
 import { streetBareRecipe } from "./street-bare.js";
 import { streetRecipe } from "./street.js";
@@ -41,6 +44,9 @@ const RECIPES = [
     frOrderRecipe,
     frAdminSplitRecipe,
     frBareStreetRecipe,
+    czPcFirstPrepositionRecipe,
+    noStreetLedRecipe,
+    siBareVillageRecipe,
     countryBalancedRecipe,
     boundaryStressRecipe,
     anchorAbsorptionRecipe,

package/out/src/shard-recipes/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/shard-recipes/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,EAAE,sBAAsB,EAAE,MAAM,wBAAwB,CAAA;AAC/D,OAAO,EAAE,oBAAoB,EAAE,MAAM,sBAAsB,CAAA;AAC3D,OAAO,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAA;AAC7D,OAAO,EAAE,kBAAkB,EAAE,MAAM,qBAAqB,CAAA;AACxD,OAAO,EAAE,kBAAkB,EAAE,MAAM,qBAAqB,CAAA;AACxD,OAAO,EAAE,aAAa,EAAE,MAAM,eAAe,CAAA;AAC7C,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAA;AAC1C,OAAO,EAAE,gBAAgB,EAAE,MAAM,kBAAkB,CAAA;AACnD,OAAO,EAAE,kBAAkB,EAAE,MAAM,mBAAmB,CAAA;AACtD,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAA;AAC1C,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAA;AAC/C,OAAO,EAAE,gBAAgB,EAAE,MAAM,mBAAmB,CAAA;AACpD,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAA;AAEzC,OAAO,EAAE,iBAAiB,EAAE,MAAM,mBAAmB,CAAA;AACrD,OAAO,EAAE,gBAAgB,EAAE,MAAM,kBAAkB,CAAA;AACnD,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAA;AAC1C,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAA;AAEtC,cAAc,eAAe,CAAA;AAE7B,iDAAiD;AACjD,MAAM,OAAO,GAA2B;IACvC,YAAY;IACZ,gBAAgB;IAChB,iBAAiB;IACjB,cAAc;IACd,gBAAgB;IAChB,WAAW;IACX,gBAAgB;IAChB,UAAU;IACV,kBAAkB;IAClB,YAAY;IACZ,YAAY;IACZ,aAAa;IACb,kBAAkB;IAClB,kBAAkB;IAClB,qBAAqB;IACrB,oBAAoB;IACpB,sBAAsB;CACtB,CAAA;AAED,4BAA4B;AAC5B,MAAM,CAAC,MAAM,aAAa,GAAqC,IAAI,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,CAAA;AAEvG,+CAA+C;AAC/C,MAAM,UAAU,cAAc,CAAC,IAAY;IAC1C,OAAO,aAAa,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;AAC/B,CAAC;AAED,yCAAyC;AACzC,MAAM,UAAU,gBAAgB;IAC/B,OAAO,OAAO,CAAA;AACf,CAAC"}
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/shard-recipes/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,EAAE,sBAAsB,EAAE,MAAM,wBAAwB,CAAA;AAC/D,OAAO,EAAE,oBAAoB,EAAE,MAAM,sBAAsB,CAAA;AAC3D,OAAO,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAA;AAC7D,OAAO,EAAE,0BAA0B,EAAE,MAAM,6BAA6B,CAAA;AACxE,OAAO,EAAE,kBAAkB,EAAE,MAAM,qBAAqB,CAAA;AACxD,OAAO,EAAE,kBAAkB,EAAE,MAAM,qBAAqB,CAAA;AACxD,OAAO,EAAE,aAAa,EAAE,MAAM,eAAe,CAAA;AAC7C,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAA;AAC1C,OAAO,EAAE,gBAAgB,EAAE,MAAM,kBAAkB,CAAA;AACnD,OAAO,EAAE,kBAAkB,EAAE,MAAM,mBAAmB,CAAA;AACtD,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAA;AAC1C,OAAO,EAAE,iBAAiB,EAAE,MAAM,oBAAoB,CAAA;AACtD,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAA;AAC/C,OAAO,EAAE,gBAAgB,EAAE,MAAM,mBAAmB,CAAA;AACpD,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAA;AAEzC,OAAO,EAAE,mBAAmB,EAAE,MAAM,sBAAsB,CAAA;AAC1D,OAAO,EAAE,iBAAiB,EAAE,MAAM,mBAAmB,CAAA;AACrD,OAAO,EAAE,gBAAgB,EAAE,MAAM,kBAAkB,CAAA;AACnD,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAA;AAC1C,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAA;AAEtC,cAAc,eAAe,CAAA;AAE7B,iDAAiD;AACjD,MAAM,OAAO,GAA2B;IACvC,YAAY;IACZ,gBAAgB;IAChB,iBAAiB;IACjB,cAAc;IACd,gBAAgB;IAChB,WAAW;IACX,gBAAgB;IAChB,UAAU;IACV,kBAAkB;IAClB,YAAY;IACZ,YAAY;IACZ,aAAa;IACb,kBAAkB;IAClB,kBAAkB;IAClB,0BAA0B;IAC1B,iBAAiB;IACjB,mBAAmB;IACnB,qBAAqB;IACrB,oBAAoB;IACpB,sBAAsB;CACtB,CAAA;AAED,4BAA4B;AAC5B,MAAM,CAAC,MAAM,aAAa,GAAqC,IAAI,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,CAAA;AAEvG,+CAA+C;AAC/C,MAAM,UAAU,cAAc,CAAC,IAAY;IAC1C,OAAO,aAAa,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;AAC/B,CAAC;AAED,yCAAyC;AACzC,MAAM,UAAU,gBAAgB;IAC/B,OAAO,OAAO,CAAA;AACf,CAAC"}

package/out/src/shard-recipes/no-street-led.d.ts

@@ -0,0 +1,20 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   `no-street-led` — the Norwegian street-led shard, third orthography of the #901
+ *   leading-name-boundary family. The post-#920 NO row-read measured the residual as
+ *   ORDER-SENSITIVE decode: street-led forms ("Tangavegen 40, 5620 Tørvikbygd") carry a 30%
+ *   residual vs pc-first's 7% — the diacritic street head (…vegen/…veien with ø/å/æ) is the same
+ *   leading-name-before-comma boundary the FR bare-street and SI village forms exercise, in a
+ *   third orthography. All three real orders cycle per tuple (balanced polarity — the v1.9.9
+ *   lesson: a one-order gradient loses at convergence):
+ *
+ *   1. canonical  "«st» «n», «pc» «city»"   (the 30% residual class — the lesson)
+ *   2. city-first "«city», «pc», «st» «n»"
+ *   3. pc-first   "«pc» «city», «st» «n»"   (the 7% floor — the anchor the others converge to)
+ */
+import { type ShardRecipe } from "./scaffold.js";
+export declare const noStreetLedRecipe: ShardRecipe;
+//# sourceMappingURL=no-street-led.d.ts.map

package/out/src/shard-recipes/no-street-led.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"no-street-led.d.ts","sourceRoot":"","sources":["../../../src/shard-recipes/no-street-led.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAEH,OAAO,EAA6C,KAAK,WAAW,EAAiB,MAAM,eAAe,CAAA;AAE1G,eAAO,MAAM,iBAAiB,EAAE,WAwD/B,CAAA"}

package/out/src/shard-recipes/no-street-led.js

@@ -0,0 +1,74 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   `no-street-led` — the Norwegian street-led shard, third orthography of the #901
+ *   leading-name-boundary family. The post-#920 NO row-read measured the residual as
+ *   ORDER-SENSITIVE decode: street-led forms ("Tangavegen 40, 5620 Tørvikbygd") carry a 30%
+ *   residual vs pc-first's 7% — the diacritic street head (…vegen/…veien with ø/å/æ) is the same
+ *   leading-name-before-comma boundary the FR bare-street and SI village forms exercise, in a
+ *   third orthography. All three real orders cycle per tuple (balanced polarity — the v1.9.9
+ *   lesson: a one-order gradient loses at convergence):
+ *
+ *   1. canonical  "«st» «n», «pc» «city»"   (the 30% residual class — the lesson)
+ *   2. city-first "«city», «pc», «st» «n»"
+ *   3. pc-first   "«pc» «city», «st» «n»"   (the 7% floor — the anchor the others converge to)
+ */
+import { alignAndWrite, makeMulberry32, readTuples, shardSourceID } from "./scaffold.js";
+export const noStreetLedRecipe = {
+    name: "no-street-led",
+    description: "NO street-led boundary form (#901 family): '«st» «n», «pc» «city»' — diacritic street heads",
+    mode: "tuples",
+    async run(opts, write) {
+        makeMulberry32(opts.seed);
+        let read = 0;
+        let emitted = 0;
+        let skipped = 0;
+        for await (const t of readTuples(opts.input)) {
+            read++;
+            const street = String(t.street ?? "").trim();
+            const city = String(t.locality ?? "").trim();
+            const number = String(t.number ?? "").trim();
+            const postcode = String(t.postcode ?? "").trim();
+            if (!street || !city || !number || !postcode) {
+                skipped++;
+                continue;
+            }
+            const order = read % 3;
+            let raw;
+            const components = {
+                street,
+                house_number: number,
+                postcode,
+                locality: city,
+            };
+            if (order === 0) {
+                raw = `${street} ${number}, ${postcode} ${city}`;
+            }
+            else if (order === 1) {
+                raw = `${city}, ${postcode}, ${street} ${number}`;
+            }
+            else {
+                raw = `${postcode} ${city}, ${street} ${number}`;
+            }
+            const source_id = shardSourceID("synth-no-street-led", { ...components, o: String(order), v: String(read) });
+            const canonical = {
+                raw,
+                components,
+                country: "NO",
+                locale: "nb-NO",
+                source: "synth-no-street-led",
+                source_id,
+                corpus_version: "0.10.0",
+                license: "Synthetic — no-street-led; (street, number, postcode, city) from OpenAddresses NO (per-source attribution in the model card)",
+            };
+            if (alignAndWrite(write, canonical, "no-street-led"))
+                emitted++;
+            else
+                skipped++;
+        }
+        return { read, emitted, skipped };
+    },
+};
+//# sourceMappingURL=no-street-led.js.map

package/out/src/shard-recipes/no-street-led.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"no-street-led.js","sourceRoot":"","sources":["../../../src/shard-recipes/no-street-led.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAEH,OAAO,EAAE,aAAa,EAAE,cAAc,EAAE,UAAU,EAAoB,aAAa,EAAE,MAAM,eAAe,CAAA;AAE1G,MAAM,CAAC,MAAM,iBAAiB,GAAgB;IAC7C,IAAI,EAAE,eAAe;IACrB,WAAW,EAAE,6FAA6F;IAC1G,IAAI,EAAE,QAAQ;IACd,KAAK,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK;QACpB,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QACzB,IAAI,IAAI,GAAG,CAAC,CAAA;QACZ,IAAI,OAAO,GAAG,CAAC,CAAA;QACf,IAAI,OAAO,GAAG,CAAC,CAAA;QAEf,IAAI,KAAK,EAAE,MAAM,CAAC,IAAI,UAAU,CAAC,IAAI,CAAC,KAAM,CAAC,EAAE,CAAC;YAC/C,IAAI,EAAE,CAAA;YACN,MAAM,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;YAC5C,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;YAC5C,MAAM,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;YAC5C,MAAM,QAAQ,GAAG,MAAM,CAAC,CAAC,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;YAEhD,IAAI,CAAC,MAAM,IAAI,CAAC,IAAI,IAAI,CAAC,MAAM,IAAI,CAAC,QAAQ,EAAE,CAAC;gBAC9C,OAAO,EAAE,CAAA;gBACT,SAAQ;YACT,CAAC;YACD,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,CAAA;YACtB,IAAI,GAAW,CAAA;YACf,MAAM,UAAU,GAA2B;gBAC1C,MAAM;gBACN,YAAY,EAAE,MAAM;gBACpB,QAAQ;gBACR,QAAQ,EAAE,IAAI;aACd,CAAA;YAED,IAAI,KAAK,KAAK,CAAC,EAAE,CAAC;gBACjB,GAAG,GAAG,GAAG,MAAM,IAAI,MAAM,KAAK,QAAQ,IAAI,IAAI,EAAE,CAAA;YACjD,CAAC;iBAAM,IAAI,KAAK,KAAK,CAAC,EAAE,CAAC;gBACxB,GAAG,GAAG,GAAG,IAAI,KAAK,QAAQ,KAAK,MAAM,IAAI,MAAM,EAAE,CAAA;YAClD,CAAC;iBAAM,CAAC;gBACP,GAAG,GAAG,GAAG,QAAQ,IAAI,IAAI,KAAK,MAAM,IAAI,MAAM,EAAE,CAAA;YACjD,CAAC;YACD,MAAM,SAAS,GAAG,aAAa,CAAC,qBAAqB,EAAE,EAAE,GAAG,UAAU,EAAE,CAAC,EAAE,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;YAC5G,MAAM,SAAS,GAAG;gBACjB,GAAG;gBACH,UAAU;gBACV,OAAO,EAAE,IAAI;gBACb,MAAM,EAAE,OAAO;gBACf,MAAM,EAAE,qBAAqB;gBAC7B,SAAS;gBACT,cAAc,EAAE,QAAQ;gBACxB,OAAO,EACN,8HAA8H;aAC/H,CAAA;YAED,IAAI,aAAa,CAAC,KAAK,EAAE,SAAS,EAAE,eAAe,CAAC;gBAAE,OAAO,EAAE,CAAA;;gBAC1D,OAAO,EAAE,CAAA;QACf,CAAC;QAED,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,CAAA;IAClC,CAAC;CACD,CAAA"}

package/out/src/shard-recipes/si-bare-village.d.ts

@@ -0,0 +1,28 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   `si-bare-village` — the Slovenian no-street counter-shard (#901 run-2). Slovenia's rural
+ *   addressing has NO street line: the village name is the street-level token and repeats as the
+ *   locality ("Zabiče 8, 6250 Zabiče"). The v1.9.8 gate FALSIFIED the fr-bare-street shard alone on
+ *   exactly this class (SI resolve −3.4pp; "Apače 108" split into street "Apače 10" + house "8") —
+ *   the bare-street boundary lesson generalizes onto a form where the leading name must keep its
+ *   number whole and the trailing mention must stay locality-bound. This shard is the paired
+ *   counter-distribution: same lesson ("name before number, comma, then admin"), opposite polarity
+ *   on the trailing mention.
+ *
+ *   Three real-order templates cycle per tuple (mirrors the coord-golden orders so the eval and the
+ *   training distribution agree):
+ *
+ *   1. canonical  "«V» «n», «pc» «V»"
+ *   2. bare       "«V» «n», «V»"        (no postcode — the anchor-free form, the fr-bare lesson)
+ *   3. pc-first   "«pc» «V», «V» «n»"
+ *
+ *   Gold spans: leading «V» = street (matches OA ground truth — the village IS the address line),
+ *   «n» = house_number (NEVER split mid-digits), «pc» = postcode (never swallowing the neighbor),
+ *   trailing «V» = locality (the binding the resolver needs).
+ */
+import { type ShardRecipe } from "./scaffold.js";
+export declare const siBareVillageRecipe: ShardRecipe;
+//# sourceMappingURL=si-bare-village.d.ts.map

package/out/src/shard-recipes/si-bare-village.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"si-bare-village.d.ts","sourceRoot":"","sources":["../../../src/shard-recipes/si-bare-village.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;GAwBG;AAEH,OAAO,EAA6C,KAAK,WAAW,EAAiB,MAAM,eAAe,CAAA;AAE1G,eAAO,MAAM,mBAAmB,EAAE,WAwDjC,CAAA"}

package/out/src/shard-recipes/si-bare-village.js

@@ -0,0 +1,82 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   `si-bare-village` — the Slovenian no-street counter-shard (#901 run-2). Slovenia's rural
+ *   addressing has NO street line: the village name is the street-level token and repeats as the
+ *   locality ("Zabiče 8, 6250 Zabiče"). The v1.9.8 gate FALSIFIED the fr-bare-street shard alone on
+ *   exactly this class (SI resolve −3.4pp; "Apače 108" split into street "Apače 10" + house "8") —
+ *   the bare-street boundary lesson generalizes onto a form where the leading name must keep its
+ *   number whole and the trailing mention must stay locality-bound. This shard is the paired
+ *   counter-distribution: same lesson ("name before number, comma, then admin"), opposite polarity
+ *   on the trailing mention.
+ *
+ *   Three real-order templates cycle per tuple (mirrors the coord-golden orders so the eval and the
+ *   training distribution agree):
+ *
+ *   1. canonical  "«V» «n», «pc» «V»"
+ *   2. bare       "«V» «n», «V»"        (no postcode — the anchor-free form, the fr-bare lesson)
+ *   3. pc-first   "«pc» «V», «V» «n»"
+ *
+ *   Gold spans: leading «V» = street (matches OA ground truth — the village IS the address line),
+ *   «n» = house_number (NEVER split mid-digits), «pc» = postcode (never swallowing the neighbor),
+ *   trailing «V» = locality (the binding the resolver needs).
+ */
+import { alignAndWrite, makeMulberry32, readTuples, shardSourceID } from "./scaffold.js";
+export const siBareVillageRecipe = {
+    name: "si-bare-village",
+    description: "SI no-street village form (#901 run-2): '«V» «n», «pc» «V»' — the fr-bare-street counter-distribution",
+    mode: "tuples",
+    async run(opts, write) {
+        makeMulberry32(opts.seed);
+        let read = 0;
+        let emitted = 0;
+        let skipped = 0;
+        for await (const t of readTuples(opts.input)) {
+            read++;
+            const village = String(t.locality ?? "").trim();
+            const number = String(t.number ?? "").trim();
+            const postcode = String(t.postcode ?? "").trim();
+            if (!village || !number || !postcode) {
+                skipped++;
+                continue;
+            }
+            const order = read % 3;
+            let raw;
+            const components = {
+                street: village,
+                house_number: number,
+                locality: village,
+            };
+            if (order === 0) {
+                components.postcode = postcode;
+                raw = `${village} ${number}, ${postcode} ${village}`;
+            }
+            else if (order === 1) {
+                raw = `${village} ${number}, ${village}`;
+            }
+            else {
+                components.postcode = postcode;
+                raw = `${postcode} ${village}, ${village} ${number}`;
+            }
+            const source_id = shardSourceID("synth-si-bare-village", { ...components, o: String(order), v: String(read) });
+            const canonical = {
+                raw,
+                components,
+                country: "SI",
+                locale: "sl-SI",
+                source: "synth-si-bare-village",
+                source_id,
+                corpus_version: "0.9.9",
+                license: "Synthetic — si-bare-village; (village, number, postcode) from OpenAddresses SI (per-source attribution in the model card)",
+            };
+            if (alignAndWrite(write, canonical, "si-bare-village"))
+                emitted++;
+            else
+                skipped++;
+        }
+        return { read, emitted, skipped };
+    },
+};
+//# sourceMappingURL=si-bare-village.js.map

package/out/src/shard-recipes/si-bare-village.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"si-bare-village.js","sourceRoot":"","sources":["../../../src/shard-recipes/si-bare-village.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;GAwBG;AAEH,OAAO,EAAE,aAAa,EAAE,cAAc,EAAE,UAAU,EAAoB,aAAa,EAAE,MAAM,eAAe,CAAA;AAE1G,MAAM,CAAC,MAAM,mBAAmB,GAAgB;IAC/C,IAAI,EAAE,iBAAiB;IACvB,WAAW,EAAE,uGAAuG;IACpH,IAAI,EAAE,QAAQ;IACd,KAAK,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK;QACpB,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QACzB,IAAI,IAAI,GAAG,CAAC,CAAA;QACZ,IAAI,OAAO,GAAG,CAAC,CAAA;QACf,IAAI,OAAO,GAAG,CAAC,CAAA;QAEf,IAAI,KAAK,EAAE,MAAM,CAAC,IAAI,UAAU,CAAC,IAAI,CAAC,KAAM,CAAC,EAAE,CAAC;YAC/C,IAAI,EAAE,CAAA;YACN,MAAM,OAAO,GAAG,MAAM,CAAC,CAAC,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;YAC/C,MAAM,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;YAC5C,MAAM,QAAQ,GAAG,MAAM,CAAC,CAAC,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;YAEhD,IAAI,CAAC,OAAO,IAAI,CAAC,MAAM,IAAI,CAAC,QAAQ,EAAE,CAAC;gBACtC,OAAO,EAAE,CAAA;gBACT,SAAQ;YACT,CAAC;YACD,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,CAAA;YACtB,IAAI,GAAW,CAAA;YACf,MAAM,UAAU,GAA2B;gBAC1C,MAAM,EAAE,OAAO;gBACf,YAAY,EAAE,MAAM;gBACpB,QAAQ,EAAE,OAAO;aACjB,CAAA;YAED,IAAI,KAAK,KAAK,CAAC,EAAE,CAAC;gBACjB,UAAU,CAAC,QAAQ,GAAG,QAAQ,CAAA;gBAC9B,GAAG,GAAG,GAAG,OAAO,IAAI,MAAM,KAAK,QAAQ,IAAI,OAAO,EAAE,CAAA;YACrD,CAAC;iBAAM,IAAI,KAAK,KAAK,CAAC,EAAE,CAAC;gBACxB,GAAG,GAAG,GAAG,OAAO,IAAI,MAAM,KAAK,OAAO,EAAE,CAAA;YACzC,CAAC;iBAAM,CAAC;gBACP,UAAU,CAAC,QAAQ,GAAG,QAAQ,CAAA;gBAC9B,GAAG,GAAG,GAAG,QAAQ,IAAI,OAAO,KAAK,OAAO,IAAI,MAAM,EAAE,CAAA;YACrD,CAAC;YACD,MAAM,SAAS,GAAG,aAAa,CAAC,uBAAuB,EAAE,EAAE,GAAG,UAAU,EAAE,CAAC,EAAE,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;YAC9G,MAAM,SAAS,GAAG;gBACjB,GAAG;gBACH,UAAU;gBACV,OAAO,EAAE,IAAI;gBACb,MAAM,EAAE,OAAO;gBACf,MAAM,EAAE,uBAAuB;gBAC/B,SAAS;gBACT,cAAc,EAAE,OAAO;gBACvB,OAAO,EACN,2HAA2H;aAC5H,CAAA;YAED,IAAI,aAAa,CAAC,KAAK,EAAE,SAAS,EAAE,iBAAiB,CAAC;gBAAE,OAAO,EAAE,CAAA;;gBAC5D,OAAO,EAAE,CAAA;QACf,CAAC;QAED,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,CAAA;IAClC,CAAC;CACD,CAAA"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@mailwoman/corpus",
-  "version": "5.2.0",
+  "version": "5.3.0",
   "description": "Mailwoman corpus pipeline: BIO-labeled dataset builder for the neural classifier.",
   "license": "AGPL-3.0-only OR LicenseRef-Commercial",
   "repository": {
@@ -26,9 +26,9 @@
   },
   "dependencies": {
     "@dsnp/parquetjs": "1.8.7",
-    "@mailwoman/codex": "5.2.0",
-    "@mailwoman/core": "5.2.0",
-    "@mailwoman/formatter": "5.2.0",
+    "@mailwoman/codex": "5.3.0",
+    "@mailwoman/core": "5.3.0",
+    "@mailwoman/formatter": "5.3.0",
     "csv-parse": "^5.6.0",
     "fast-glob": "^3.3.3",
     "fastest-levenshtein": "^1.0.16",