@mailwoman/corpus 9.4.0 → 10.1.0

package/out/recipes/bare/country.js

@@ -0,0 +1,77 @@
+import { COUNTRY_SURFACE_FORMS, CountryNames, matchCountry } from "@mailwoman/codex/country";
+import { mulberry32 as makeMulberry32 } from "@mailwoman/core/utils";
+import { alignAndWrite, recipeSourceID } from "#recipes/scaffold";
+import { defaultRecipeSource } from "#recipes/sources";
+import { SourceRegister } from "#registers";
+import { SurfaceOrigin } from "#types";
+/**
+ * Resolved once, because `source_id` includes the source as its prefix and a pair
+ * that disagreed would name a source no row of this output declares.
+ */
+const BARE_COUNTRY_SOURCE = defaultRecipeSource("synth-bare-country");
+const BARE_COUNTRY_PROVENANCE = {
+    register: SourceRegister.Codex,
+    surface: SurfaceOrigin.Attested,
+    // The codex table is the record.
+    // Its rows have no `source_id` of their own.
+    baseSourceID: null,
+};
+const MIN_NAME_LENGTH = 4;
+function* bareCountrySurfaces() {
+    const seen = new Set();
+    for (const [iso2, forms] of Object.entries(COUNTRY_SURFACE_FORMS)) {
+        for (const surface of forms) {
+            if (surface.length < MIN_NAME_LENGTH || seen.has(surface))
+                continue;
+            seen.add(surface);
+            yield { surface, iso2 };
+        }
+    }
+    for (const name of CountryNames) {
+        if (name.length < MIN_NAME_LENGTH || seen.has(name))
+            continue;
+        const matched = matchCountry(name);
+        if (!matched)
+            continue;
+        seen.add(name);
+        yield { surface: name, iso2: matched.iso2 };
+    }
+}
+/**
+ * Generates queries that consist only of a country name, drawn from ISO names
+ * and curated surface forms of at least four characters, with no country codes.
+ */
+export const bareCountryRecipe = {
+    name: "bare-country",
+    description: "Emits a country name as the whole query, drawn from ISO names and curated endonyms but never from country codes",
+    mode: "generate",
+    async run(opts, write) {
+        makeMulberry32(opts.seed);
+        let read = 0;
+        let emitted = 0;
+        let skipped = 0;
+        for (const { surface, iso2 } of bareCountrySurfaces()) {
+            read++;
+            const components = { country: surface };
+            const source_id = recipeSourceID(BARE_COUNTRY_SOURCE, { country: surface, cc: iso2, v: String(read) });
+            const canonical = {
+                raw: surface,
+                components,
+                country: iso2,
+                locale: "und",
+                source: BARE_COUNTRY_SOURCE,
+                source_id,
+                corpus_version: "0.11.0",
+                license: "Synthetic — bare-country; surfaces from the codex country table (ISO 3166 + curated forms)",
+            };
+            if (alignAndWrite(write, canonical, "bare-country", BARE_COUNTRY_PROVENANCE)) {
+                emitted++;
+            }
+            else {
+                skipped++;
+            }
+        }
+        return { read, emitted, skipped };
+    },
+};
+//# sourceMappingURL=country.js.map

package/out/recipes/bare/country.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"country.js","sourceRoot":"","sources":["../../../lib/recipes/bare/country.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,qBAAqB,EAAE,YAAY,EAAE,YAAY,EAAE,MAAM,0BAA0B,CAAA;AAC5F,OAAO,EAAE,UAAU,IAAI,cAAc,EAAE,MAAM,uBAAuB,CAAA;AAEpE,OAAO,EAAE,aAAa,EAAqB,cAAc,EAAE,MAAM,mBAAmB,CAAA;AACpF,OAAO,EAAE,mBAAmB,EAAE,MAAM,kBAAkB,CAAA;AACtD,OAAO,EAAE,cAAc,EAAE,MAAM,YAAY,CAAA;AAC3C,OAAO,EAAE,aAAa,EAAE,MAAM,QAAQ,CAAA;AAEtC;;;GAGG;AACH,MAAM,mBAAmB,GAAG,mBAAmB,CAAC,oBAAoB,CAAC,CAAA;AAErE,MAAM,uBAAuB,GAAG;IAC/B,QAAQ,EAAE,cAAc,CAAC,KAAK;IAC9B,OAAO,EAAE,aAAa,CAAC,QAAQ;IAC/B,iCAAiC;IACjC,6CAA6C;IAC7C,YAAY,EAAE,IAAI;CAClB,CAAA;AAED,MAAM,eAAe,GAAG,CAAC,CAAA;AAEzB,QAAQ,CAAC,CAAC,mBAAmB;IAC5B,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAA;IAE9B,KAAK,MAAM,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,qBAAqB,CAAC,EAAE,CAAC;QACnE,KAAK,MAAM,OAAO,IAAI,KAAK,EAAE,CAAC;YAC7B,IAAI,OAAO,CAAC,MAAM,GAAG,eAAe,IAAI,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC;gBAAE,SAAQ;YACnE,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,CAAA;YACjB,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,CAAA;QACxB,CAAC;IACF,CAAC;IAED,KAAK,MAAM,IAAI,IAAI,YAAY,EAAE,CAAC;QACjC,IAAI,IAAI,CAAC,MAAM,GAAG,eAAe,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC;YAAE,SAAQ;QAC7D,MAAM,OAAO,GAAG,YAAY,CAAC,IAAI,CAAC,CAAA;QAElC,IAAI,CAAC,OAAO;YAAE,SAAQ;QACtB,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAA;QACd,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,EAAE,CAAA;IAC5C,CAAC;AACF,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,MAAM,iBAAiB,GAAiB;IAC9C,IAAI,EAAE,cAAc;IACpB,WAAW,EACV,iHAAiH;IAClH,IAAI,EAAE,UAAU;IAChB,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,KAAK,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,mBAAmB,EAAE,EAAE,CAAC;YACvD,IAAI,EAAE,CAAA;YAEN,MAAM,UAAU,GAA2B,EAAE,OAAO,EAAE,OAAO,EAAE,CAAA;YAC/D,MAAM,SAAS,GAAG,cAAc,CAAC,mBAAmB,EAAE,EAAE,OAAO,EAAE,OAAO,EAAE,EAAE,EAAE,IAAI,EAAE,CAAC,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;YAEtG,MAAM,SAAS,GAAG;gBACjB,GAAG,EAAE,OAAO;gBACZ,UAAU;gBACV,OAAO,EAAE,IAAI;gBACb,MAAM,EAAE,KAAK;gBACb,MAAM,EAAE,mBAAmB;gBAC3B,SAAS;gBACT,cAAc,EAAE,QAAQ;gBACxB,OAAO,EAAE,4FAA4F;aACrG,CAAA;YAED,IAAI,aAAa,CAAC,KAAK,EAAE,SAAS,EAAE,cAAc,EAAE,uBAAuB,CAAC,EAAE,CAAC;gBAC9E,OAAO,EAAE,CAAA;YACV,CAAC;iBAAM,CAAC;gBACP,OAAO,EAAE,CAAA;YACV,CAAC;QACF,CAAC;QAED,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,CAAA;IAClC,CAAC;CACD,CAAA"}

package/out/recipes/bare/postcode/eval.d.ts

@@ -0,0 +1,29 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Reserved bare-postcode capability cases. These strings are excluded from every country in the
+ *   training recipe and are read only by evaluation. Reservation is by normalized input string,
+ *   because `NNN NN` is shared by CZ, SK, SE and GR. excluding only a Czech row could still train
+ *   the model on the identical Swedish input.
+ */
+export type BarePostcodeEvalFamily = "nnn_nn" | "nnnn_ll";
+export interface BarePostcodeEvalCase {
+    input: string;
+    country: "CZ" | "NL" | "SK";
+    family: BarePostcodeEvalFamily;
+}
+/**
+ * The 56 postcode strings measured across v5.0.0 through v5.5.0 and reserved from later training.
+ */
+export declare const BARE_POSTCODE_EVAL_CASES: readonly BarePostcodeEvalCase[];
+/**
+ * Normalize the model-visible postcode value for reservation checks.
+ */
+export declare function normalizeBarePostcodeSurface(surface: string): string;
+/**
+ * Whether a postcode string is reserved exclusively for the capability evaluation.
+ */
+export declare function isReservedBarePostcode(surface: string): boolean;
+//# sourceMappingURL=eval.d.ts.map

package/out/recipes/bare/postcode/eval.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"eval.d.ts","sourceRoot":"","sources":["../../../../lib/recipes/bare/postcode/eval.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,MAAM,MAAM,sBAAsB,GAAG,QAAQ,GAAG,SAAS,CAAA;AAEzD,MAAM,WAAW,oBAAoB;IACpC,KAAK,EAAE,MAAM,CAAA;IACb,OAAO,EAAE,IAAI,GAAG,IAAI,GAAG,IAAI,CAAA;IAC3B,MAAM,EAAE,sBAAsB,CAAA;CAC9B;AA0DD;;GAEG;AACH,eAAO,MAAM,wBAAwB,EAAE,SAAS,oBAAoB,EAInE,CAAA;AAED;;GAEG;AACH,wBAAgB,4BAA4B,CAAC,OAAO,EAAE,MAAM,GAAG,MAAM,CAEpE;AAID;;GAEG;AACH,wBAAgB,sBAAsB,CAAC,OAAO,EAAE,MAAM,GAAG,OAAO,CAE/D"}

package/out/recipes/bare/postcode/eval.js

@@ -0,0 +1,85 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Reserved bare-postcode capability cases. These strings are excluded from every country in the
+ *   training recipe and are read only by evaluation. Reservation is by normalized input string,
+ *   because `NNN NN` is shared by CZ, SK, SE and GR. excluding only a Czech row could still train
+ *   the model on the identical Swedish input.
+ */
+const SPACED_FIVE_CZ = [
+    "120 00",
+    "130 00",
+    "140 00",
+    "150 00",
+    "160 00",
+    "170 00",
+    "180 00",
+    "190 00",
+    "602 00",
+    "603 00",
+    "612 00",
+    "616 00",
+    "621 00",
+    "628 00",
+    "635 00",
+    "639 00",
+    "702 00",
+    "708 00",
+    "710 00",
+    "721 00",
+    "779 00",
+    "746 01",
+    "747 05",
+    "750 02",
+];
+const SPACED_FIVE_SK = ["811 02", "821 01", "831 02", "841 01", "851 01", "949 01", "974 01", "010 01"];
+const DUTCH = [
+    "1011 AB",
+    "1013 BG",
+    "1015 CJ",
+    "1017 DK",
+    "1018 EL",
+    "1052 GM",
+    "1053 HN",
+    "1054 JP",
+    "2011 KR",
+    "2012 LS",
+    "2511 MT",
+    "2513 NV",
+    "3011 PW",
+    "3012 RX",
+    "3511 SA",
+    "3512 TB",
+    "4811 VC",
+    "5611 WD",
+    "6211 XE",
+    "6511 ZF",
+    "7511 AG",
+    "8011 BH",
+    "9711 CJ",
+    "9712 DK",
+];
+/**
+ * The 56 postcode strings measured across v5.0.0 through v5.5.0 and reserved from later training.
+ */
+export const BARE_POSTCODE_EVAL_CASES = [
+    ...SPACED_FIVE_CZ.map((input) => ({ input, country: "CZ", family: "nnn_nn" })),
+    ...SPACED_FIVE_SK.map((input) => ({ input, country: "SK", family: "nnn_nn" })),
+    ...DUTCH.map((input) => ({ input, country: "NL", family: "nnnn_ll" })),
+];
+/**
+ * Normalize the model-visible postcode value for reservation checks.
+ */
+export function normalizeBarePostcodeSurface(surface) {
+    return surface.trim().toUpperCase().replaceAll(/\s+/gu, "");
+}
+const RESERVED_SURFACES = new Set(BARE_POSTCODE_EVAL_CASES.map(({ input }) => normalizeBarePostcodeSurface(input)));
+/**
+ * Whether a postcode string is reserved exclusively for the capability evaluation.
+ */
+export function isReservedBarePostcode(surface) {
+    return RESERVED_SURFACES.has(normalizeBarePostcodeSurface(surface));
+}
+//# sourceMappingURL=eval.js.map

package/out/recipes/bare/postcode/eval.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"eval.js","sourceRoot":"","sources":["../../../../lib/recipes/bare/postcode/eval.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAUH,MAAM,cAAc,GAAG;IACtB,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;IACR,QAAQ;CACC,CAAA;AAEV,MAAM,cAAc,GAAG,CAAC,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,CAAU,CAAA;AAEhH,MAAM,KAAK,GAAG;IACb,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;IACT,SAAS;CACA,CAAA;AAEV;;GAEG;AACH,MAAM,CAAC,MAAM,wBAAwB,GAAoC;IACxE,GAAG,cAAc,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,IAAa,EAAE,MAAM,EAAE,QAAiB,EAAE,CAAC,CAAC;IAChG,GAAG,cAAc,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,IAAa,EAAE,MAAM,EAAE,QAAiB,EAAE,CAAC,CAAC;IAChG,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,IAAa,EAAE,MAAM,EAAE,SAAkB,EAAE,CAAC,CAAC;CACxF,CAAA;AAED;;GAEG;AACH,MAAM,UAAU,4BAA4B,CAAC,OAAe;IAC3D,OAAO,OAAO,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,EAAE,CAAC,CAAA;AAC5D,CAAC;AAED,MAAM,iBAAiB,GAAG,IAAI,GAAG,CAAC,wBAAwB,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,4BAA4B,CAAC,KAAK,CAAC,CAAC,CAAC,CAAA;AAEnH;;GAEG;AACH,MAAM,UAAU,sBAAsB,CAAC,OAAe;IACrD,OAAO,iBAAiB,CAAC,GAAG,CAAC,4BAA4B,CAAC,OAAO,CAAC,CAAC,CAAA;AACpE,CAAC"}

package/out/recipes/bare/postcode.d.ts

@@ -0,0 +1,41 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ * `bare-postcode` — a postcode without another address component, in the country's format.
+ * Every surface the recipe renders is run through `detectKnownFormats` and refused unless the detector
+ * calls it a postcode, so the rendering table and `known-formats.ts`'s patterns must agree.
+ */
+import type { PathBuilderLike } from "path-ts";
+import { type CorpusRecipe } from "#recipes/scaffold";
+/**
+ * Choose distinct postcodes without inheriting the publisher's row order.
+ *
+ * A prior sort makes the result independent of input order.
+ * The seeded sample spreads reproducibly across the complete set.
+ */
+export declare function selectPostcodes(codes: Iterable<string>, limit: number, seed: number): string[];
+/**
+ * Return every required input path that is absent, preserving declaration order for diagnostics.
+ */
+export declare function findMissingPostcodeSources<P extends PathBuilderLike>(paths: readonly P[], exists?: (path: P) => Promise<boolean>): Promise<P[]>;
+/**
+ * Every surface a country writes for one postcode, spaced form first, or `[]`
+ * when this recipe defines no form for that country or the code does not fit.
+ *
+ * Exported so a test can pin that every rendered surface is one {@linkcode detectedAsPostcode} accepts.
+ */
+export declare function renderBarePostcode(country: string, postcode: string): string[];
+/**
+ * Whether `known-formats.ts` reads this surface as a postcode across its whole span.
+ *
+ * A surface the detector does not recognize would train the model on a string
+ * the query-shape prior cannot then support.
+ */
+export declare function detectedAsPostcode(surface: string): boolean;
+/**
+ * Recipe registered with the corpus builder.
+ */
+export declare const barePostcodeRecipe: CorpusRecipe;
+//# sourceMappingURL=postcode.d.ts.map

package/out/recipes/bare/postcode.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"postcode.d.ts","sourceRoot":"","sources":["../../../lib/recipes/bare/postcode.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAQH,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,SAAS,CAAA;AAI9C,OAAO,EAAiB,KAAK,YAAY,EAAkC,MAAM,mBAAmB,CAAA;AA8GpG;;;;;GAKG;AACH,wBAAgB,eAAe,CAAC,KAAK,EAAE,QAAQ,CAAC,MAAM,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,GAAG,MAAM,EAAE,CAM9F;AAED;;GAEG;AACH,wBAAsB,0BAA0B,CAAC,CAAC,SAAS,eAAe,EACzE,KAAK,EAAE,SAAS,CAAC,EAAE,EACnB,MAAM,GAAE,CAAC,IAAI,EAAE,CAAC,KAAK,OAAO,CAAC,OAAO,CAAc,GAChD,OAAO,CAAC,CAAC,EAAE,CAAC,CAId;AAED;;;;;GAKG;AACH,wBAAgB,kBAAkB,CAAC,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,GAAG,MAAM,EAAE,CAM9E;AAED;;;;;GAKG;AACH,wBAAgB,kBAAkB,CAAC,OAAO,EAAE,MAAM,GAAG,OAAO,CAI3D;AAED;;GAEG;AACH,eAAO,MAAM,kBAAkB,EAAE,YAgKhC,CAAA"}

package/out/recipes/bare/postcode.js

@@ -0,0 +1,279 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ * `bare-postcode` — a postcode without another address component, in the country's format.
+ * Every surface the recipe renders is run through `detectKnownFormats` and refused unless the detector
+ * calls it a postcode, so the rendering table and `known-formats.ts`'s patterns must agree.
+ */
+import { isNLPostcodeKey } from "@mailwoman/codex/nl";
+import { dataRootPath } from "@mailwoman/core/data-root";
+import { pathExists } from "@mailwoman/core/fs/readers";
+import { SeededRandom } from "@mailwoman/core/utils";
+import { computeQueryShape } from "@mailwoman/query-shape";
+import { isPostcodeFormat } from "@mailwoman/query-shape/known-formats";
+import { normalizeGauntletSurface, readGauntletInputs } from "#gauntlet-inputs";
+import { isReservedBarePostcode } from "#recipes/bare/postcode/eval";
+import { alignAndWrite, readCSVRecords, recipeSourceID } from "#recipes/scaffold";
+import { defaultRecipeSource } from "#recipes/sources";
+import { SourceRegister } from "#registers";
+import { SurfaceOrigin } from "#types";
+/**
+ * A postcode in the form its country writes, with no surrounding text.
+ *
+ * The codes come from this repository's postcode-format tables rather than a
+ * national postcode file, so the register is the codex.
+ */
+/**
+ * Resolved once, because `source_id` includes the source as its prefix and a pair
+ * that disagreed would name a source no row of this output declares.
+ */
+const BARE_POSTCODE_SOURCE = defaultRecipeSource("synth-bare-postcode");
+const BARE_POSTCODE_PROVENANCE = {
+    register: SourceRegister.Codex,
+    surface: SurfaceOrigin.Attested,
+    // The codex table is the record.
+    // Its rows have no `source_id` of their own.
+    baseSourceID: null,
+};
+const EXTRACTED = dataRootPath("openaddresses", "extracted");
+/**
+ * Read from the archive rather than guessed, because OpenAddresses keeps the Swedish
+ * spelling on two of these (`savsjö`, `Österåker`) while folding the rest to ascii.
+ */
+const SWEDISH_MUNICIPALITIES = [
+    "alingsas",
+    "gislaved",
+    "goteborg",
+    "helsingborg",
+    "hoganas",
+    "kalmar",
+    "kristinehamn",
+    "malmo",
+    "nacka",
+    "savsjö",
+    "stockholm",
+    "uppsala",
+    "vasteras",
+    "vaxholm",
+    "vaxjo",
+    "Österåker",
+];
+/**
+ * Greece is deliberately absent: its only member declares a `postcode` column with no values over
+ * 10,877 rows, though `gr_postcode` shares `NNN NN` with CZ/SK/SE, so no row here claims to be Greek.
+ */
+const SOURCES = [
+    { csv: EXTRACTED("cz", "countrywide.csv"), country: "CZ" },
+    { csv: EXTRACTED("sk", "countrywide.csv"), country: "SK" },
+    { csv: EXTRACTED("nl", "countrywide.csv"), country: "NL" },
+    ...SWEDISH_MUNICIPALITIES.map((name) => ({
+        csv: EXTRACTED("se", `municipality_of_${name}.csv`),
+        country: "SE",
+    })),
+];
+/**
+ * The postcode format to use on its own and the locale to stamp.
+ *
+ * This map includes countries whose bare postcode collides with a house number.
+ * `render` answers with the spaced form first.
+ * GB is absent because `SW1A 1AA` starts with letters.
+ */
+const WRITTEN_FORMS = new Map([
+    // `NNN NN`, written with the space.
+    // The four countries share the shape, so `detectKnownFormats` labels all four the same.
+    ["CZ", { locale: "cs-CZ", render: spacedThree }],
+    ["SK", { locale: "sk-SK", render: spacedThree }],
+    ["SE", { locale: "sv-SE", render: spacedThree }],
+    ["GR", { locale: "el-GR", render: spacedThree }],
+    // Both spellings of `nnnn LL` are attested.
+    // The spaced form is the failing case.
+    [
+        "NL",
+        {
+            locale: "nl-NL",
+            render: (compact) => (isNLPostcodeKey(compact) ? [`${compact.slice(0, 4)} ${compact.slice(4)}`, compact] : []),
+        },
+    ],
+]);
+/**
+ * `10000` → `["100 00", "10000"]`.
+ *
+ * Sweden and Greece group five digits after the third, as Czechia and Slovakia do.
+ * The compact form remains because sources store it that way.
+ */
+function spacedThree(compact) {
+    if (!/^\d{5}$/.test(compact))
+        return [];
+    return [`${compact.slice(0, 3)} ${compact.slice(3)}`, compact];
+}
+/**
+ * Choose distinct postcodes without inheriting the publisher's row order.
+ *
+ * A prior sort makes the result independent of input order.
+ * The seeded sample spreads reproducibly across the complete set.
+ */
+export function selectPostcodes(codes, limit, seed) {
+    const pool = [...new Set(codes)].toSorted();
+    if (limit >= pool.length)
+        return pool;
+    return new SeededRandom(seed).sample(pool, Math.max(0, limit));
+}
+/**
+ * Return every required input path that is absent, preserving declaration order for diagnostics.
+ */
+export async function findMissingPostcodeSources(paths, exists = pathExists) {
+    const results = await Promise.all(paths.map(async (path) => ({ path, exists: await exists(path) })));
+    return results.filter((result) => !result.exists).map(({ path }) => path);
+}
+/**
+ * Every surface a country writes for one postcode, spaced form first, or `[]`
+ * when this recipe defines no form for that country or the code does not fit.
+ *
+ * Exported so a test can pin that every rendered surface is one {@linkcode detectedAsPostcode} accepts.
+ */
+export function renderBarePostcode(country, postcode) {
+    const form = WRITTEN_FORMS.get(country.trim().toUpperCase());
+    if (!form)
+        return [];
+    return form.render(postcode.trim().toUpperCase().replaceAll(/\s+/gu, ""));
+}
+/**
+ * Whether `known-formats.ts` reads this surface as a postcode across its whole span.
+ *
+ * A surface the detector does not recognize would train the model on a string
+ * the query-shape prior cannot then support.
+ */
+export function detectedAsPostcode(surface) {
+    return computeQueryShape(surface).knownFormats.some((hit) => isPostcodeFormat(hit.format) && hit.span.start === 0 && hit.span.end === surface.length);
+}
+/**
+ * Recipe registered with the corpus builder.
+ */
+export const barePostcodeRecipe = {
+    name: "bare-postcode",
+    description: "A postcode alone, in its country's written form — the shape no other recipe output carries (CZ/SK/SE/NL)",
+    mode: "generate",
+    options: [
+        {
+            flag: "--count",
+            description: "Cap on emitted rows; each country contributes an equal share until its codes run out",
+        },
+    ],
+    async run(opts, write) {
+        let read = 0;
+        let emitted = 0;
+        let skipped = 0;
+        let unrecognized = 0;
+        // Preflight the complete input set before writing a row.
+        // A missing municipality would otherwise produce a plausible artifact with
+        // less Swedish coverage than the recipe declares.
+        const missing = await findMissingPostcodeSources(SOURCES.map(({ csv }) => csv));
+        if (missing.length) {
+            throw new Error(`bare-postcode is missing ${missing.length} of ${SOURCES.length} required OpenAddresses files, ` +
+                `starting with ${missing[0]}. Fetch the per-source runs into \`openaddresses/extracted/\` ` +
+                "(anonymous, see packages/corpus/CLAUDE.md).");
+        }
+        // Per-country budget, because supply is uneven: the Netherlands publishes ~460,000
+        // distinct `nnnn LL` codes against Czechia's 2,669 and Slovakia's 1,059, so an uncapped
+        // pass would emit 98.9% Dutch rows and teach the `NNN NN` countries almost no pattern.
+        const countries = [...new Set(SOURCES.map((source) => source.country))];
+        const budget = opts.count ? Math.ceil(opts.count / countries.length) : Number.POSITIVE_INFINITY;
+        const perCountry = new Map(countries.map((country) => [country, 0]));
+        const codesByCountry = new Map(countries.map((country) => [country, new Set()]));
+        // The gauntlet boards are a second held-out register, separate from
+        // `BARE_POSTCODE_EVAL_CASES`, and `cz/bare-postcode.jsonl` predates this recipe.
+        const boardInputs = await readGauntletInputs();
+        let boardRefused = 0;
+        for (const source of SOURCES) {
+            const form = WRITTEN_FORMS.get(source.country);
+            if (!form)
+                continue;
+            for await (const row of readCSVRecords(source.csv)) {
+                read++;
+                const compact = String(row.postcode ?? "")
+                    .trim()
+                    .toUpperCase()
+                    .replaceAll(/\s+/gu, "");
+                const countryCodes = codesByCountry.get(source.country);
+                if (!compact || isReservedBarePostcode(compact) || countryCodes.has(compact)) {
+                    skipped++;
+                    continue;
+                }
+                // A board row is refused on any of its written forms, so the check runs over
+                // each rendered surface rather than the compact code; `normalizeGauntletSurface`
+                // makes `100 00` and `10000` one string.
+                if (form.render(compact).some((surface) => boardInputs.has(normalizeGauntletSurface(surface)))) {
+                    boardRefused++;
+                    skipped++;
+                    continue;
+                }
+                if (!form.render(compact).length) {
+                    skipped++;
+                    continue;
+                }
+                countryCodes.add(compact);
+            }
+        }
+        for (const country of countries) {
+            const form = WRITTEN_FORMS.get(country);
+            const maximumCodes = Number.isFinite(budget) ? Math.ceil(budget / 2) : Number.POSITIVE_INFINITY;
+            const countrySeed = opts.seed ^ (country.charCodeAt(0) << 8) ^ country.charCodeAt(1);
+            const selected = selectPostcodes(codesByCountry.get(country), maximumCodes, countrySeed);
+            for (const compact of selected) {
+                const surfaces = form.render(compact);
+                for (const surface of surfaces) {
+                    if (perCountry.get(country) >= budget)
+                        break;
+                    // A surface the detector does not read as a postcode is a disagreement
+                    // between this table and `known-formats.ts`.
+                    if (!detectedAsPostcode(surface)) {
+                        unrecognized++;
+                        continue;
+                    }
+                    const components = { postcode: surface };
+                    const canonical = {
+                        raw: surface,
+                        components,
+                        country,
+                        locale: form.locale,
+                        source: BARE_POSTCODE_SOURCE,
+                        source_id: recipeSourceID(BARE_POSTCODE_SOURCE, {
+                            ...components,
+                            c: country,
+                        }),
+                        corpus_version: "0.30.0",
+                        license: "Synthetic — bare-postcode; postcodes from OpenAddresses (per-source attribution in the model card)",
+                    };
+                    if (alignAndWrite(write, canonical, "bare-postcode", BARE_POSTCODE_PROVENANCE)) {
+                        emitted++;
+                        perCountry.set(country, perCountry.get(country) + 1);
+                    }
+                    else {
+                        skipped++;
+                    }
+                }
+            }
+        }
+        // A rendered surface the detector refuses is an interface failure: every form here is
+        // one `known-formats.ts` declares a pattern for, so the count is expected to be zero.
+        if (unrecognized > 0) {
+            throw new Error(`${unrecognized} rendered surfaces were not read as a postcode by detectKnownFormats — the ` +
+                "WRITTEN_FORMS table and known-formats.ts's PATTERNS disagree. Reconcile them before building.");
+        }
+        // An empty build is a failure rather than an empty answer: the required
+        // files were preflighted above, so zero rows means their postcode columns
+        // or the written-form rules stopped providing usable data.
+        if (emitted === 0) {
+            throw new Error(`bare-postcode emitted no rows from ${SOURCES.length} readable sources. ` +
+                "The postcode columns or the written-form rules changed.");
+        }
+        // Reported rather than folded into `skipped` so a growing board is visible.
+        // A count of zero where the boards hold postcodes means the check stopped reaching them.
+        console.error(`  bare-postcode: ${boardRefused.toLocaleString()} rows refused as gauntlet board inputs ` +
+            `(${boardInputs.size.toLocaleString()} inputs read)`);
+        return { read, emitted, skipped };
+    },
+};
+//# sourceMappingURL=postcode.js.map

package/out/recipes/bare/postcode.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"postcode.js","sourceRoot":"","sources":["../../../lib/recipes/bare/postcode.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAA;AACrD,OAAO,EAAE,YAAY,EAAE,MAAM,2BAA2B,CAAA;AACxD,OAAO,EAAE,UAAU,EAAE,MAAM,4BAA4B,CAAA;AACvD,OAAO,EAAE,YAAY,EAAE,MAAM,uBAAuB,CAAA;AACpD,OAAO,EAAE,iBAAiB,EAAE,MAAM,wBAAwB,CAAA;AAC1D,OAAO,EAAE,gBAAgB,EAAE,MAAM,sCAAsC,CAAA;AAGvE,OAAO,EAAE,wBAAwB,EAAE,kBAAkB,EAAE,MAAM,kBAAkB,CAAA;AAC/E,OAAO,EAAE,sBAAsB,EAAE,MAAM,6BAA6B,CAAA;AACpE,OAAO,EAAE,aAAa,EAAqB,cAAc,EAAE,cAAc,EAAE,MAAM,mBAAmB,CAAA;AACpG,OAAO,EAAE,mBAAmB,EAAE,MAAM,kBAAkB,CAAA;AACtD,OAAO,EAAE,cAAc,EAAE,MAAM,YAAY,CAAA;AAC3C,OAAO,EAAE,aAAa,EAAE,MAAM,QAAQ,CAAA;AAEtC;;;;;GAKG;AACH;;;GAGG;AACH,MAAM,oBAAoB,GAAG,mBAAmB,CAAC,qBAAqB,CAAC,CAAA;AAEvE,MAAM,wBAAwB,GAAG;IAChC,QAAQ,EAAE,cAAc,CAAC,KAAK;IAC9B,OAAO,EAAE,aAAa,CAAC,QAAQ;IAC/B,iCAAiC;IACjC,6CAA6C;IAC7C,YAAY,EAAE,IAAI;CAClB,CAAA;AAWD,MAAM,SAAS,GAAG,YAAY,CAAC,eAAe,EAAE,WAAW,CAAC,CAAA;AAE5D;;;GAGG;AACH,MAAM,sBAAsB,GAAG;IAC9B,UAAU;IACV,UAAU;IACV,UAAU;IACV,aAAa;IACb,SAAS;IACT,QAAQ;IACR,cAAc;IACd,OAAO;IACP,OAAO;IACP,QAAQ;IACR,WAAW;IACX,SAAS;IACT,UAAU;IACV,SAAS;IACT,OAAO;IACP,WAAW;CACX,CAAA;AAED;;;GAGG;AACH,MAAM,OAAO,GAAqB;IACjC,EAAE,GAAG,EAAE,SAAS,CAAC,IAAI,EAAE,iBAAiB,CAAC,EAAE,OAAO,EAAE,IAAI,EAAE;IAC1D,EAAE,GAAG,EAAE,SAAS,CAAC,IAAI,EAAE,iBAAiB,CAAC,EAAE,OAAO,EAAE,IAAI,EAAE;IAC1D,EAAE,GAAG,EAAE,SAAS,CAAC,IAAI,EAAE,iBAAiB,CAAC,EAAE,OAAO,EAAE,IAAI,EAAE;IAC1D,GAAG,sBAAsB,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;QACxC,GAAG,EAAE,SAAS,CAAC,IAAI,EAAE,mBAAmB,IAAI,MAAM,CAAC;QACnD,OAAO,EAAE,IAAI;KACb,CAAC,CAAC;CACH,CAAA;AAED;;;;;;GAMG;AACH,MAAM,aAAa,GAAmF,IAAI,GAAG,CAAC;IAC7G,oCAAoC;IACpC,wFAAwF;IACxF,CAAC,IAAI,EAAE,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;IAChD,CAAC,IAAI,EAAE,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;IAChD,CAAC,IAAI,EAAE,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;IAChD,CAAC,IAAI,EAAE,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC;IAChD,4CAA4C;IAC5C,uCAAuC;IACvC;QACC,IAAI;QACJ;YACC,MAAM,EAAE,OAAO;YACf,MAAM,EAAE,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,eAAe,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;SAC9G;KACD;CACD,CAAC,CAAA;AAEF;;;;;GAKG;AACH,SAAS,WAAW,CAAC,OAAe;IACnC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,OAAO,CAAC;QAAE,OAAO,EAAE,CAAA;IAEvC,OAAO,CAAC,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,OAAO,CAAC,CAAA;AAC/D,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,eAAe,CAAC,KAAuB,EAAE,KAAa,EAAE,IAAY;IACnF,MAAM,IAAI,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,QAAQ,EAAE,CAAA;IAE3C,IAAI,KAAK,IAAI,IAAI,CAAC,MAAM;QAAE,OAAO,IAAI,CAAA;IAErC,OAAO,IAAI,YAAY,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC,CAAA;AAC/D,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,0BAA0B,CAC/C,KAAmB,EACnB,MAAM,GAAkC,UAAU;IAElD,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;IAEpG,OAAO,OAAO,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,EAAE,EAAE,EAAE,CAAC,IAAI,CAAC,CAAA;AAC1E,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,kBAAkB,CAAC,OAAe,EAAE,QAAgB;IACnE,MAAM,IAAI,GAAG,aAAa,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,CAAA;IAE5D,IAAI,CAAC,IAAI;QAAE,OAAO,EAAE,CAAA;IAEpB,OAAO,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAA;AAC1E,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,kBAAkB,CAAC,OAAe;IACjD,OAAO,iBAAiB,CAAC,OAAO,CAAC,CAAC,YAAY,CAAC,IAAI,CAClD,CAAC,GAAG,EAAE,EAAE,CAAC,gBAAgB,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,KAAK,KAAK,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,KAAK,OAAO,CAAC,MAAM,CAChG,CAAA;AACF,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,kBAAkB,GAAiB;IAC/C,IAAI,EAAE,eAAe;IACrB,WAAW,EACV,0GAA0G;IAC3G,IAAI,EAAE,UAAU;IAChB,OAAO,EAAE;QACR;YACC,IAAI,EAAE,SAAS;YACf,WAAW,EAAE,sFAAsF;SACnG;KACD;IACD,KAAK,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK;QACpB,IAAI,IAAI,GAAG,CAAC,CAAA;QACZ,IAAI,OAAO,GAAG,CAAC,CAAA;QACf,IAAI,OAAO,GAAG,CAAC,CAAA;QACf,IAAI,YAAY,GAAG,CAAC,CAAA;QAEpB,yDAAyD;QACzD,2EAA2E;QAC3E,kDAAkD;QAClD,MAAM,OAAO,GAAG,MAAM,0BAA0B,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,EAAE,GAAG,EAAE,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,CAAA;QAE/E,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;YACpB,MAAM,IAAI,KAAK,CACd,4BAA4B,OAAO,CAAC,MAAM,OAAO,OAAO,CAAC,MAAM,iCAAiC;gBAC/F,iBAAiB,OAAO,CAAC,CAAC,CAAC,gEAAgE;gBAC3F,6CAA6C,CAC9C,CAAA;QACF,CAAC;QAED,mFAAmF;QACnF,wFAAwF;QACxF,uFAAuF;QACvF,MAAM,SAAS,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAA;QACvE,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,iBAAiB,CAAA;QAC/F,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,CAAC,CAAA;QACpE,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,OAAO,EAAE,IAAI,GAAG,EAAU,CAAC,CAAC,CAAC,CAAA;QAExF,oEAAoE;QACpE,iFAAiF;QACjF,MAAM,WAAW,GAAG,MAAM,kBAAkB,EAAE,CAAA;QAC9C,IAAI,YAAY,GAAG,CAAC,CAAA;QAEpB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC9B,MAAM,IAAI,GAAG,aAAa,CAAC,GAAG,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;YAE9C,IAAI,CAAC,IAAI;gBAAE,SAAQ;YAEnB,IAAI,KAAK,EAAE,MAAM,GAAG,IAAI,cAAc,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC;gBACpD,IAAI,EAAE,CAAA;gBAEN,MAAM,OAAO,GAAG,MAAM,CAAC,GAAG,CAAC,QAAQ,IAAI,EAAE,CAAC;qBACxC,IAAI,EAAE;qBACN,WAAW,EAAE;qBACb,UAAU,CAAC,OAAO,EAAE,EAAE,CAAC,CAAA;gBAEzB,MAAM,YAAY,GAAG,cAAc,CAAC,GAAG,CAAC,MAAM,CAAC,OAAO,CAAE,CAAA;gBAExD,IAAI,CAAC,OAAO,IAAI,sBAAsB,CAAC,OAAO,CAAC,IAAI,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;oBAC9E,OAAO,EAAE,CAAA;oBAET,SAAQ;gBACT,CAAC;gBAED,6EAA6E;gBAC7E,iFAAiF;gBACjF,yCAAyC;gBACzC,IAAI,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,WAAW,CAAC,GAAG,CAAC,wBAAwB,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;oBAChG,YAAY,EAAE,CAAA;oBAEd,OAAO,EAAE,CAAA;oBAET,SAAQ;gBACT,CAAC;gBAED,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,CAAC;oBAClC,OAAO,EAAE,CAAA;oBAET,SAAQ;gBACT,CAAC;gBAED,YAAY,CAAC,GAAG,CAAC,OAAO,CAAC,CAAA;YAC1B,CAAC;QACF,CAAC;QAED,KAAK,MAAM,OAAO,IAAI,SAAS,EAAE,CAAC;YACjC,MAAM,IAAI,GAAG,aAAa,CAAC,GAAG,CAAC,OAAO,CAAE,CAAA;YACxC,MAAM,YAAY,GAAG,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,iBAAiB,CAAA;YAC/F,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,GAAG,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,CAAA;YACpF,MAAM,QAAQ,GAAG,eAAe,CAAC,cAAc,CAAC,GAAG,CAAC,OAAO,CAAE,EAAE,YAAY,EAAE,WAAW,CAAC,CAAA;YAEzF,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;gBAChC,MAAM,QAAQ,GAAG,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;gBAErC,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;oBAChC,IAAI,UAAU,CAAC,GAAG,CAAC,OAAO,CAAE,IAAI,MAAM;wBAAE,MAAK;oBAE7C,uEAAuE;oBACvE,6CAA6C;oBAC7C,IAAI,CAAC,kBAAkB,CAAC,OAAO,CAAC,EAAE,CAAC;wBAClC,YAAY,EAAE,CAAA;wBAEd,SAAQ;oBACT,CAAC;oBAED,MAAM,UAAU,GAAG,EAAE,QAAQ,EAAE,OAAO,EAAE,CAAA;oBAExC,MAAM,SAAS,GAAG;wBACjB,GAAG,EAAE,OAAO;wBACZ,UAAU;wBACV,OAAO;wBACP,MAAM,EAAE,IAAI,CAAC,MAAM;wBACnB,MAAM,EAAE,oBAAoB;wBAC5B,SAAS,EAAE,cAAc,CAAC,oBAAoB,EAAE;4BAC/C,GAAG,UAAU;4BACb,CAAC,EAAE,OAAO;yBACV,CAAC;wBACF,cAAc,EAAE,QAAQ;wBACxB,OAAO,EACN,oGAAoG;qBACrG,CAAA;oBAED,IAAI,aAAa,CAAC,KAAK,EAAE,SAAS,EAAE,eAAe,EAAE,wBAAwB,CAAC,EAAE,CAAC;wBAChF,OAAO,EAAE,CAAA;wBACT,UAAU,CAAC,GAAG,CAAC,OAAO,EAAE,UAAU,CAAC,GAAG,CAAC,OAAO,CAAE,GAAG,CAAC,CAAC,CAAA;oBACtD,CAAC;yBAAM,CAAC;wBACP,OAAO,EAAE,CAAA;oBACV,CAAC;gBACF,CAAC;YACF,CAAC;QACF,CAAC;QAED,sFAAsF;QACtF,sFAAsF;QACtF,IAAI,YAAY,GAAG,CAAC,EAAE,CAAC;YACtB,MAAM,IAAI,KAAK,CACd,GAAG,YAAY,6EAA6E;gBAC3F,+FAA+F,CAChG,CAAA;QACF,CAAC;QAED,wEAAwE;QACxE,0EAA0E;QAC1E,2DAA2D;QAC3D,IAAI,OAAO,KAAK,CAAC,EAAE,CAAC;YACnB,MAAM,IAAI,KAAK,CACd,sCAAsC,OAAO,CAAC,MAAM,qBAAqB;gBACxE,yDAAyD,CAC1D,CAAA;QACF,CAAC;QAED,4EAA4E;QAC5E,yFAAyF;QACzF,OAAO,CAAC,KAAK,CACZ,oBAAoB,YAAY,CAAC,cAAc,EAAE,yCAAyC;YACzF,IAAI,WAAW,CAAC,IAAI,CAAC,cAAc,EAAE,eAAe,CACrD,CAAA;QAED,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,CAAA;IAClC,CAAC;CACD,CAAA"}

package/out/recipes/boundary-stress.d.ts

@@ -3,18 +3,18 @@
  * @license AGPL-3.0
  * @author Teffen Ellis, et al.
  *
- *   `boundary-stress` slice recipe (#375) — the boundary-instability augmentation. Self-generates
- *   `--count` rows from {@link synthesizeBoundaryStressRow}'s weighted template mix (the v1.7.0,
- *   DeepSeek-tuned 2026-06-18 composition), aligns each to BIO, and emits a labeled JSONL. The
- *   change for the taxonomy's #1 parser family (the boundary-wobble class). Ported from
- *   scripts/build-boundary-stress-slice.mjs.
+ *   `boundary-stress` recipe, the boundary-instability augmentation. Self-generates `--count` rows
+ *   from {@link synthesizeBoundaryStressRow}'s weighted template mix. It aligns each to BIO and emits a
+ *   labeled jsonl.
  *
- *   `synthesizeBoundaryStressRow` is NOT re-exported from the corpus index — imported directly here.
+ *   `synthesizeBoundaryStressRow` is imported directly here.
  */
 import { type CorpusRecipe } from "#recipes/scaffold";
 /**
- * Slice recipe registered with the corpus builder — see the file header for the parse behaviour it exists to exercise,
- * and `description` below for the surface form it generates.
+ * Recipe registered with the corpus builder.
+ *
+ * See the file header for the parse behavior this recipe exercises and `description`
+ * below for the surface form it generates.
  */
 export declare const boundaryStressRecipe: CorpusRecipe;
 //# sourceMappingURL=boundary-stress.d.ts.map

package/out/recipes/boundary-stress.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"boundary-stress.d.ts","sourceRoot":"","sources":["../../lib/recipes/boundary-stress.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAIH,OAAO,EAAE,KAAK,YAAY,EAAiB,MAAM,mBAAmB,CAAA;AAoCpE;;;GAGG;AACH,eAAO,MAAM,oBAAoB,EAAE,YA6ClC,CAAA"}
+{"version":3,"file":"boundary-stress.d.ts","sourceRoot":"","sources":["../../lib/recipes/boundary-stress.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAKH,OAAO,EAAE,KAAK,YAAY,EAAkB,MAAM,mBAAmB,CAAA;AA2CrE;;;;;GAKG;AACH,eAAO,MAAM,oBAAoB,EAAE,YAsDlC,CAAA"}