@mailwoman/corpus 9.4.0 → 10.1.0

package/out/adapters/ban/adapter.d.ts

@@ -1,44 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   `ban`: Base Adresse Nationale CSV adapter (FR street-level).
- *
- *   Input: a CSV dump from `adresse.data.gouv.fr` (semicolon-separated, ~25M rows nationally). The
- *   adapter only reads the small set of columns needed for the corpus:
- *
- *   - `numero` → `house_number`
- *   - `rep` → repetition index ("bis", "ter") appended to house_number
- *   - `nom_voie` → `street` (full road name; includes the prefix "Rue", "Avenue", etc.)
- *   - `code_postal` → `postcode`
- *   - `nom_commune` → `locality`
- *
- *   `region` and `country` are not in BAN. The adapter stamps `country: "FR"` on every row; region is
- *   left for the wof-postalcode + wof-admin cross-reference at corpus build time (a future pass;
- *   for Phase 1 the row's region is simply absent).
- *
- *   License: the official BAN (adresse.data.gouv.fr) is DUAL-licensed — Licence Ouverte 2.0 (Etalab,
- *   attribution-only) OR ODbL (share-alike). We ELECT Licence Ouverte 2.0 (issue #26 Tier B:
- *   allowed for training with attribution; the ODbL option's share-alike obligation would defeat
- *   the proprietary-weights goal). Stamped onto every row as `Licence Ouverte 2.0` — NOT the older
- *   conservative `ODbL-1.0` label, which wrongly read as Tier-C-denied in the corpus license audit.
- *   The model card MUST carry the BAN attribution (Tier B obligation).
- *
- *   The adapter is streaming-aware: `CSVSpliterator.fromAsync` reads the `;`-delimited dump row by
- *   row, so a 25M-row file never sits in memory. Honors `opts.limit` for fixture / smoke runs,
- *   `opts.signal` for cancellation, and `opts.country` for a self-consistency check (errors if
- *   country !== FR).
- */
-import type { CorpusAdapter } from "#types";
-/**
- * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
- * it came from.
- */
-export declare const BAN_ADAPTER_ID = "ban";
-export declare function createBanAdapter(): CorpusAdapter;
-/**
- * The configured adapter instance registered with the corpus builder.
- */
-export declare const banAdapter: CorpusAdapter;
-//# sourceMappingURL=adapter.d.ts.map

package/out/adapters/ban/adapter.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../lib/adapters/ban/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA+BG;AAQH,OAAO,KAAK,EAAgC,aAAa,EAAE,MAAM,QAAQ,CAAA;AAEzE;;;GAGG;AACH,eAAO,MAAM,cAAc,QAAQ,CAAA;AAsDnC,wBAAgB,gBAAgB,IAAI,aAAa,CAqFhD;AAED;;GAEG;AACH,eAAO,MAAM,UAAU,eAAqB,CAAA"}

package/out/adapters/ban/adapter.js

@@ -1,149 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   `ban`: Base Adresse Nationale CSV adapter (FR street-level).
- *
- *   Input: a CSV dump from `adresse.data.gouv.fr` (semicolon-separated, ~25M rows nationally). The
- *   adapter only reads the small set of columns needed for the corpus:
- *
- *   - `numero` → `house_number`
- *   - `rep` → repetition index ("bis", "ter") appended to house_number
- *   - `nom_voie` → `street` (full road name; includes the prefix "Rue", "Avenue", etc.)
- *   - `code_postal` → `postcode`
- *   - `nom_commune` → `locality`
- *
- *   `region` and `country` are not in BAN. The adapter stamps `country: "FR"` on every row; region is
- *   left for the wof-postalcode + wof-admin cross-reference at corpus build time (a future pass;
- *   for Phase 1 the row's region is simply absent).
- *
- *   License: the official BAN (adresse.data.gouv.fr) is DUAL-licensed — Licence Ouverte 2.0 (Etalab,
- *   attribution-only) OR ODbL (share-alike). We ELECT Licence Ouverte 2.0 (issue #26 Tier B:
- *   allowed for training with attribution; the ODbL option's share-alike obligation would defeat
- *   the proprietary-weights goal). Stamped onto every row as `Licence Ouverte 2.0` — NOT the older
- *   conservative `ODbL-1.0` label, which wrongly read as Tier-C-denied in the corpus license audit.
- *   The model card MUST carry the BAN attribution (Tier B obligation).
- *
- *   The adapter is streaming-aware: `CSVSpliterator.fromAsync` reads the `;`-delimited dump row by
- *   row, so a 25M-row file never sits in memory. Honors `opts.limit` for fixture / smoke runs,
- *   `opts.signal` for cancellation, and `opts.country` for a self-consistency check (errors if
- *   country !== FR).
- */
-import { isPresent } from "@mailwoman/core/objects";
-import { reconcileComponents } from "@mailwoman/formatter";
-import { CSVSpliterator } from "spliterator";
-import { decomposeFrStreet } from "#adapters/ban/street-decompose";
-import { stableSourceID } from "#adapters/utils";
-/**
- * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
- * it came from.
- */
-export const BAN_ADAPTER_ID = "ban";
-/**
- * Compose `house_number` from `numero` + `rep`. BAN uses `rep` for repetition indices ("bis", "ter", "quater") that
- * follow the house number. Result: `"10 bis"`, `"45"`, etc.
- */
-function composeHouseNumber(numero, rep) {
-    const n = numero.trim();
-    const r = rep.trim();
-    if (!n)
-        return "";
-    return r ? `${n} ${r}` : n;
-}
-/**
- * Compose the raw FR-style address line. Two common BAN-derived shapes:
- *
- * "10 bis Avenue des Champs-Élysées, 75008 Paris" "45 Cours Lafayette, 69003 Lyon"
- *
- * FR convention puts postcode on the same line as the locality, comma-separated from the street. The adapter renders
- * that directly rather than relying on OpenCage's template — BAN already gives us the canonical FR strings; the
- * template would round-trip identically.
- */
-function composeRaw(house, street, postcode, locality) {
-    const parts = [];
-    const streetPart = [house, street].filter(isPresent).join(" ").trim();
-    if (streetPart) {
-        parts.push(streetPart);
-    }
-    const cityPart = [postcode, locality].filter(isPresent).join(" ").trim();
-    if (cityPart) {
-        parts.push(cityPart);
-    }
-    return parts.join(", ").replaceAll(/\s+/g, " ").trim();
-}
-export function createBanAdapter() {
-    return {
-        id: BAN_ADAPTER_ID,
-        defaultLicense: "Licence Ouverte 2.0",
-        description: "Base Adresse Nationale (FR): house-number-level street addresses (~25M rows).",
-        async *rows(opts) {
-            if (opts.country && opts.country !== "FR") {
-                throw new Error(`ban adapter: only FR supported, got country=${opts.country}`);
-            }
-            const rows = CSVSpliterator.fromAsync(opts.inputPath, {
-                mode: "object",
-                normalizeKeys: false,
-                columnDelimiter: ";",
-                enableQuoteHandling: true,
-            });
-            let emitted = 0;
-            for await (const record of rows) {
-                if (opts.signal?.aborted)
-                    break;
-                if (opts.limit !== undefined && emitted >= opts.limit)
-                    break;
-                const house = composeHouseNumber(record.numero ?? "", record.rep ?? "");
-                const street = (record.nom_voie ?? "").trim();
-                const postcode = (record.code_postal ?? "").trim();
-                const locality = (record.nom_commune ?? "").trim();
-                if (!street || !locality)
-                    continue;
-                if (!house && !postcode)
-                    continue;
-                const decomposed = decomposeFrStreet(street);
-                const components = {};
-                if (house) {
-                    components.house_number = house;
-                }
-                if (decomposed.prefix) {
-                    components.street_prefix = decomposed.prefix;
-                }
-                if (decomposed.street) {
-                    components.street = decomposed.street;
-                }
-                if (postcode) {
-                    components.postcode = postcode;
-                }
-                if (locality) {
-                    components.locality = locality;
-                }
-                const raw = composeRaw(house, street, postcode, locality);
-                if (!raw)
-                    continue;
-                const aligned = reconcileComponents(components, raw);
-                if (!Object.keys(aligned).length)
-                    continue;
-                const sourceID = record.id?.trim()
-                    ? `${BAN_ADAPTER_ID}-${record.id.trim()}`
-                    : stableSourceID(BAN_ADAPTER_ID, aligned);
-                yield {
-                    raw,
-                    components: aligned,
-                    country: "FR",
-                    locale: "fr-FR",
-                    source: BAN_ADAPTER_ID,
-                    source_id: sourceID,
-                    corpus_version: "",
-                    license: "Licence Ouverte 2.0",
-                };
-                emitted++;
-            }
-        },
-    };
-}
-/**
- * The configured adapter instance registered with the corpus builder.
- */
-export const banAdapter = createBanAdapter();
-//# sourceMappingURL=adapter.js.map

package/out/adapters/ban/adapter.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../lib/adapters/ban/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA+BG;AAEH,OAAO,EAAE,SAAS,EAAE,MAAM,yBAAyB,CAAA;AACnD,OAAO,EAAE,mBAAmB,EAAE,MAAM,sBAAsB,CAAA;AAC1D,OAAO,EAAE,cAAc,EAAE,MAAM,aAAa,CAAA;AAE5C,OAAO,EAAE,iBAAiB,EAAE,MAAM,gCAAgC,CAAA;AAClE,OAAO,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAA;AAGhD;;;GAGG;AACH,MAAM,CAAC,MAAM,cAAc,GAAG,KAAK,CAAA;AAenC;;;GAGG;AACH,SAAS,kBAAkB,CAAC,MAAc,EAAE,GAAW;IACtD,MAAM,CAAC,GAAG,MAAM,CAAC,IAAI,EAAE,CAAA;IACvB,MAAM,CAAC,GAAG,GAAG,CAAC,IAAI,EAAE,CAAA;IAEpB,IAAI,CAAC,CAAC;QAAE,OAAO,EAAE,CAAA;IAEjB,OAAO,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAA;AAC3B,CAAC;AAED;;;;;;;;GAQG;AACH,SAAS,UAAU,CAAC,KAAa,EAAE,MAAc,EAAE,QAAgB,EAAE,QAAgB;IACpF,MAAM,KAAK,GAAa,EAAE,CAAA;IAC1B,MAAM,UAAU,GAAG,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;IAErE,IAAI,UAAU,EAAE,CAAC;QAChB,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;IACvB,CAAC;IAED,MAAM,QAAQ,GAAG,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;IAExE,IAAI,QAAQ,EAAE,CAAC;QACd,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;IACrB,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,UAAU,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;AACvD,CAAC;AAED,MAAM,UAAU,gBAAgB;IAC/B,OAAO;QACN,EAAE,EAAE,cAAc;QAClB,cAAc,EAAE,qBAAqB;QACrC,WAAW,EAAE,+EAA+E;QAE5F,KAAK,CAAC,CAAC,IAAI,CAAC,IAAoB;YAC/B,IAAI,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;gBAC3C,MAAM,IAAI,KAAK,CAAC,+CAA+C,IAAI,CAAC,OAAO,EAAE,CAAC,CAAA;YAC/E,CAAC;YAED,MAAM,IAAI,GAAG,cAAc,CAAC,SAAS,CAAC,IAAI,CAAC,SAAS,EAAE;gBACrD,IAAI,EAAE,QAAQ;gBACd,aAAa,EAAE,KAAK;gBACpB,eAAe,EAAE,GAAG;gBACpB,mBAAmB,EAAE,IAAI;aACzB,CAAC,CAAA;YAEF,IAAI,OAAO,GAAG,CAAC,CAAA;YAEf,IAAI,KAAK,EAAE,MAAM,MAAM,IAAI,IAA6B,EAAE,CAAC;gBAC1D,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;oBAAE,MAAK;gBAE/B,IAAI,IAAI,CAAC,KAAK,KAAK,SAAS,IAAI,OAAO,IAAI,IAAI,CAAC,KAAK;oBAAE,MAAK;gBAE5D,MAAM,KAAK,GAAG,kBAAkB,CAAC,MAAM,CAAC,MAAM,IAAI,EAAE,EAAE,MAAM,CAAC,GAAG,IAAI,EAAE,CAAC,CAAA;gBACvE,MAAM,MAAM,GAAG,CAAC,MAAM,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;gBAC7C,MAAM,QAAQ,GAAG,CAAC,MAAM,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;gBAClD,MAAM,QAAQ,GAAG,CAAC,MAAM,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;gBAElD,IAAI,CAAC,MAAM,IAAI,CAAC,QAAQ;oBAAE,SAAQ;gBAElC,IAAI,CAAC,KAAK,IAAI,CAAC,QAAQ;oBAAE,SAAQ;gBAEjC,MAAM,UAAU,GAAG,iBAAiB,CAAC,MAAM,CAAC,CAAA;gBAE5C,MAAM,UAAU,GAA+B,EAAE,CAAA;gBAEjD,IAAI,KAAK,EAAE,CAAC;oBACX,UAAU,CAAC,YAAY,GAAG,KAAK,CAAA;gBAChC,CAAC;gBAED,IAAI,UAAU,CAAC,MAAM,EAAE,CAAC;oBACvB,UAAU,CAAC,aAAa,GAAG,UAAU,CAAC,MAAM,CAAA;gBAC7C,CAAC;gBAED,IAAI,UAAU,CAAC,MAAM,EAAE,CAAC;oBACvB,UAAU,CAAC,MAAM,GAAG,UAAU,CAAC,MAAM,CAAA;gBACtC,CAAC;gBAED,IAAI,QAAQ,EAAE,CAAC;oBACd,UAAU,CAAC,QAAQ,GAAG,QAAQ,CAAA;gBAC/B,CAAC;gBAED,IAAI,QAAQ,EAAE,CAAC;oBACd,UAAU,CAAC,QAAQ,GAAG,QAAQ,CAAA;gBAC/B,CAAC;gBAED,MAAM,GAAG,GAAG,UAAU,CAAC,KAAK,EAAE,MAAM,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAA;gBAEzD,IAAI,CAAC,GAAG;oBAAE,SAAQ;gBAElB,MAAM,OAAO,GAAG,mBAAmB,CAAC,UAAU,EAAE,GAAG,CAAC,CAAA;gBAEpD,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,MAAM;oBAAE,SAAQ;gBAE1C,MAAM,QAAQ,GAAG,MAAM,CAAC,EAAE,EAAE,IAAI,EAAE;oBACjC,CAAC,CAAC,GAAG,cAAc,IAAI,MAAM,CAAC,EAAE,CAAC,IAAI,EAAE,EAAE;oBACzC,CAAC,CAAC,cAAc,CAAC,cAAc,EAAE,OAAO,CAAC,CAAA;gBAE1C,MAAM;oBACL,GAAG;oBACH,UAAU,EAAE,OAAO;oBACnB,OAAO,EAAE,IAAI;oBACb,MAAM,EAAE,OAAO;oBACf,MAAM,EAAE,cAAc;oBACtB,SAAS,EAAE,QAAQ;oBACnB,cAAc,EAAE,EAAE;oBAClB,OAAO,EAAE,qBAAqB;iBAC9B,CAAA;gBAED,OAAO,EAAE,CAAA;YACV,CAAC;QACF,CAAC;KACD,CAAA;AACF,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,UAAU,GAAG,gBAAgB,EAAE,CAAA"}

package/out/adapters/ban/street-decompose.d.ts

@@ -1,28 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   Decompose a French street name into Stage 3 components. French convention puts the street type as
- *   a leading word: "Rue de Rivoli", "Avenue des Champs-Élysées", "Bd Voltaire".
- *
- *   The street type becomes street_prefix in our schema. The remaining tokens form the street name.
- *
- *   Examples: "Rue de Rivoli" → { prefix: "Rue", street: "de Rivoli" } "Avenue des Champs-Élysées" →
- *   { prefix: "Avenue", street: "des Champs-Élysées" } "Boulevard Voltaire" → { prefix:
- *   "Boulevard", street: "Voltaire" }
- *
- *   Sources street types from `core/data/libpostal/dictionaries/fr/street_types.txt`.
- */
-export interface DecomposedFrStreet {
-    prefix: string | null;
-    street: string;
-}
-/**
- * Decompose a French street name into prefix (leading type word) and street name.
- *
- * If the first 1-2 tokens match a known street type (allowing for multi-word like "ancien chemin"), they become the
- * prefix. Returns `{ prefix: null, street: original }` if no match.
- */
-export declare function decomposeFrStreet(fullname: string): DecomposedFrStreet;
-//# sourceMappingURL=street-decompose.d.ts.map

package/out/adapters/ban/street-decompose.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"street-decompose.d.ts","sourceRoot":"","sources":["../../../lib/adapters/ban/street-decompose.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AAWH,MAAM,WAAW,kBAAkB;IAClC,MAAM,EAAE,MAAM,GAAG,IAAI,CAAA;IACrB,MAAM,EAAE,MAAM,CAAA;CACd;AAED;;;;;GAKG;AACH,wBAAgB,iBAAiB,CAAC,QAAQ,EAAE,MAAM,GAAG,kBAAkB,CA4BtE"}

package/out/adapters/ban/street-decompose.js

@@ -1,51 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   Decompose a French street name into Stage 3 components. French convention puts the street type as
- *   a leading word: "Rue de Rivoli", "Avenue des Champs-Élysées", "Bd Voltaire".
- *
- *   The street type becomes street_prefix in our schema. The remaining tokens form the street name.
- *
- *   Examples: "Rue de Rivoli" → { prefix: "Rue", street: "de Rivoli" } "Avenue des Champs-Élysées" →
- *   { prefix: "Avenue", street: "des Champs-Élysées" } "Boulevard Voltaire" → { prefix:
- *   "Boulevard", street: "Voltaire" }
- *
- *   Sources street types from `core/data/libpostal/dictionaries/fr/street_types.txt`.
- */
-import { loadLibpostalDictionary } from "#adapters/utils";
-/**
- * Tokens a BAN street needs before a type/article/name decomposition is attempted.
- */
-const MIN_TOKENS_FOR_DECOMPOSE = 3;
-const STREET_TYPES_FR = await loadLibpostalDictionary("fr", "street_types.txt");
-/**
- * Decompose a French street name into prefix (leading type word) and street name.
- *
- * If the first 1-2 tokens match a known street type (allowing for multi-word like "ancien chemin"), they become the
- * prefix. Returns `{ prefix: null, street: original }` if no match.
- */
-export function decomposeFrStreet(fullname) {
-    const trimmed = fullname.trim();
-    if (!trimmed)
-        return { prefix: null, street: "" };
-    const tokens = trimmed.split(/\s+/);
-    if (tokens.length < 2)
-        return { prefix: null, street: trimmed };
-    const norm = (s) => s.toLowerCase().replace(/[.,;]$/, "");
-    // Try 2-word prefix first (e.g. "ancien chemin")
-    if (tokens.length >= MIN_TOKENS_FOR_DECOMPOSE) {
-        const twoWord = norm(tokens[0]) + " " + norm(tokens[1]);
-        if (STREET_TYPES_FR.has(twoWord)) {
-            return { prefix: tokens.slice(0, 2).join(" "), street: tokens.slice(2).join(" ") };
-        }
-    }
-    // Then try 1-word prefix
-    const first = norm(tokens[0]);
-    if (STREET_TYPES_FR.has(first)) {
-        return { prefix: tokens[0], street: tokens.slice(1).join(" ") };
-    }
-    return { prefix: null, street: trimmed };
-}
-//# sourceMappingURL=street-decompose.js.map

package/out/adapters/ban/street-decompose.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"street-decompose.js","sourceRoot":"","sources":["../../../lib/adapters/ban/street-decompose.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AAEH,OAAO,EAAE,uBAAuB,EAAE,MAAM,iBAAiB,CAAA;AAEzD;;GAEG;AACH,MAAM,wBAAwB,GAAG,CAAC,CAAA;AAElC,MAAM,eAAe,GAAG,MAAM,uBAAuB,CAAC,IAAI,EAAE,kBAAkB,CAAC,CAAA;AAO/E;;;;;GAKG;AACH,MAAM,UAAU,iBAAiB,CAAC,QAAgB;IACjD,MAAM,OAAO,GAAG,QAAQ,CAAC,IAAI,EAAE,CAAA;IAE/B,IAAI,CAAC,OAAO;QAAE,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,EAAE,EAAE,CAAA;IAEjD,MAAM,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC,KAAK,CAAC,CAAA;IAEnC,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,CAAA;IAE/D,MAAM,IAAI,GAAG,CAAC,CAAS,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAA;IAEjE,iDAAiD;IACjD,IAAI,MAAM,CAAC,MAAM,IAAI,wBAAwB,EAAE,CAAC;QAC/C,MAAM,OAAO,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC,GAAG,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC,CAAA;QAEzD,IAAI,eAAe,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;YAClC,OAAO,EAAE,MAAM,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,MAAM,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAA;QACnF,CAAC;IACF,CAAC;IAED,yBAAyB;IACzB,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC,CAAA;IAE9B,IAAI,eAAe,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;QAChC,OAAO,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC,CAAE,EAAE,MAAM,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAA;IACjE,CAAC;IAED,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,CAAA;AACzC,CAAC"}

package/out/adapters/fcc-bdc/adapter.d.ts

@@ -1,68 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   `fcc-bdc`: FCC Broadband Data Collection (BDC) — Fabric-derived location consumer.
- *
- *   The first member of Phase 1.6's "adversarial sources" class. BDC ships the public-domain US
- *   broadband-serviceable-location (BSL) fabric: ~120M addresses keyed by stable `location_id`,
- *   carrying `address_primary` + `city` + `state` + `zip` + `zip_suffix`. Compared to the clean
- *   gazetteer rows from WOF / TIGER / BAN, BDC carries the chaos of address data that has passed
- *   through several layers of human entry + automated geocoding + revision: abbreviation drift,
- *   inconsistent unit designators, "RR" / "HC" / "PSC" rural-route shapes, embedded apartment /
- *   suite numbers that did not survive the address parser cleanly. This is the highest-signal,
- *   hardest-to-normalize address corpus in the federal public-domain catalog.
- *
- *   Following the `tiger` / `wof-admin` pattern, this adapter consumes a SQLite database the operator
- *   pre-builds via the isp-nexus BDC ETL (`/srv/isp-nexus/sync/fcc/bdc/`) or any equivalent
- *   host-side pipeline. The mailwoman side does not download or parse the raw CSV/ZIP distribution
- *   directly — that keeps the adapter narrow and the BDC ingest pluggable.
- *
- *   The SQLite schema is documented in README.md and modeled after `NTIARecord`
- *   (`isp-nexus/fcc/bdc/data-collection.ts`): one row per `location_id`. The adapter splits
- *   `address_primary` into `house_number` (leading numeric prefix, if any) + `street` (everything
- *   after), and combines `zip` + `zip_suffix` into the canonical USPS `postcode` slot.
- *
- *   One CanonicalRow per fabric record. Unlike `tiger` (multiple postcode variants per segment) or
- *   `wof-admin` (multiple hierarchy variants per place), BDC records already represent fully
- *   specified addresses; no fan-out is warranted. Adversarial composition (Phase 1.6 §2.1) is the
- *   mechanism for deriving multiple training rows per BDC record.
- *
- *   License: stamped `"Public Domain"` per the BDC fabric's US federal-government distribution terms.
- *   The CostQuest Fabric source data has its own license; consumers who substitute that path should
- *   re-stamp accordingly.
- */
-import type { CorpusAdapter } from "#types";
-/**
- * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
- * it came from.
- */
-export declare const FCC_BDC_ADAPTER_ID = "fcc-bdc";
-/**
- * License carried by this source (Public Domain), attached to each row so downstream consumers inherit the terms rather
- * than having to look them up.
- */
-export declare const FCC_BDC_DEFAULT_LICENSE = "Public Domain";
-/**
- * Combine `zip` + optional `zip_suffix` into the canonical USPS postcode surface form.
- *
- * NTIARecord doc is ambiguous about whether `zip_suffix` is the 4-digit extension alone or the full ZIP+4 string. This
- * handles both:
- *
- * - Bare 4-digit extension (`zip="94103"`, `zip_suffix="1234"`) → `"94103-1234"`
- * - Already-joined form (`zip_suffix="94103-1234"`) → returned as-is
- * - No suffix → bare `zip`
- *
- * Empty / whitespace-only suffix is treated as missing.
- */
-export declare function buildPostcode(zip: string, suffix: string | null): string;
-/**
- * Build a BDC adapter. Pure factory so multiple instances can be created in tests.
- */
-export declare function createFccBdcAdapter(): CorpusAdapter;
-/**
- * The configured adapter instance registered with the corpus builder.
- */
-export declare const fccBdcAdapter: CorpusAdapter;
-//# sourceMappingURL=adapter.d.ts.map

package/out/adapters/fcc-bdc/adapter.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../lib/adapters/fcc-bdc/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAkCG;AAQH,OAAO,KAAK,EAAgC,aAAa,EAAE,MAAM,QAAQ,CAAA;AAEzE;;;GAGG;AACH,eAAO,MAAM,kBAAkB,YAAY,CAAA;AAC3C;;;GAGG;AACH,eAAO,MAAM,uBAAuB,kBAAkB,CAAA;AActD;;;;;;;;;;;GAWG;AACH,wBAAgB,aAAa,CAAC,GAAG,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,GAAG,IAAI,GAAG,MAAM,CAWxE;AAED;;GAEG;AACH,wBAAgB,mBAAmB,IAAI,aAAa,CAyEnD;AAED;;GAEG;AACH,eAAO,MAAM,aAAa,eAAwB,CAAA"}

package/out/adapters/fcc-bdc/adapter.js

@@ -1,144 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   `fcc-bdc`: FCC Broadband Data Collection (BDC) — Fabric-derived location consumer.
- *
- *   The first member of Phase 1.6's "adversarial sources" class. BDC ships the public-domain US
- *   broadband-serviceable-location (BSL) fabric: ~120M addresses keyed by stable `location_id`,
- *   carrying `address_primary` + `city` + `state` + `zip` + `zip_suffix`. Compared to the clean
- *   gazetteer rows from WOF / TIGER / BAN, BDC carries the chaos of address data that has passed
- *   through several layers of human entry + automated geocoding + revision: abbreviation drift,
- *   inconsistent unit designators, "RR" / "HC" / "PSC" rural-route shapes, embedded apartment /
- *   suite numbers that did not survive the address parser cleanly. This is the highest-signal,
- *   hardest-to-normalize address corpus in the federal public-domain catalog.
- *
- *   Following the `tiger` / `wof-admin` pattern, this adapter consumes a SQLite database the operator
- *   pre-builds via the isp-nexus BDC ETL (`/srv/isp-nexus/sync/fcc/bdc/`) or any equivalent
- *   host-side pipeline. The mailwoman side does not download or parse the raw CSV/ZIP distribution
- *   directly — that keeps the adapter narrow and the BDC ingest pluggable.
- *
- *   The SQLite schema is documented in README.md and modeled after `NTIARecord`
- *   (`isp-nexus/fcc/bdc/data-collection.ts`): one row per `location_id`. The adapter splits
- *   `address_primary` into `house_number` (leading numeric prefix, if any) + `street` (everything
- *   after), and combines `zip` + `zip_suffix` into the canonical USPS `postcode` slot.
- *
- *   One CanonicalRow per fabric record. Unlike `tiger` (multiple postcode variants per segment) or
- *   `wof-admin` (multiple hierarchy variants per place), BDC records already represent fully
- *   specified addresses; no fan-out is warranted. Adversarial composition (Phase 1.6 §2.1) is the
- *   mechanism for deriving multiple training rows per BDC record.
- *
- *   License: stamped `"Public Domain"` per the BDC fabric's US federal-government distribution terms.
- *   The CostQuest Fabric source data has its own license; consumers who substitute that path should
- *   re-stamp accordingly.
- */
-import { formatAddress, reconcileComponents } from "@mailwoman/formatter";
-import { DatabaseClient } from "@mailwoman/sqlite/client";
-import { splitStreetLine } from "#adapters/utils";
-import { lookupStateAbbreviation } from "#codex/us-fips-state";
-/**
- * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
- * it came from.
- */
-export const FCC_BDC_ADAPTER_ID = "fcc-bdc";
-/**
- * License carried by this source (Public Domain), attached to each row so downstream consumers inherit the terms rather
- * than having to look them up.
- */
-export const FCC_BDC_DEFAULT_LICENSE = "Public Domain";
-/**
- * Combine `zip` + optional `zip_suffix` into the canonical USPS postcode surface form.
- *
- * NTIARecord doc is ambiguous about whether `zip_suffix` is the 4-digit extension alone or the full ZIP+4 string. This
- * handles both:
- *
- * - Bare 4-digit extension (`zip="94103"`, `zip_suffix="1234"`) → `"94103-1234"`
- * - Already-joined form (`zip_suffix="94103-1234"`) → returned as-is
- * - No suffix → bare `zip`
- *
- * Empty / whitespace-only suffix is treated as missing.
- */
-export function buildPostcode(zip, suffix) {
-    const z = zip.trim();
-    if (!z)
-        return "";
-    const s = suffix?.trim() ?? "";
-    if (!s)
-        return z;
-    if (s.includes("-"))
-        return s;
-    return `${z}-${s}`;
-}
-/**
- * Build a BDC adapter. Pure factory so multiple instances can be created in tests.
- */
-export function createFccBdcAdapter() {
-    return {
-        id: FCC_BDC_ADAPTER_ID,
-        defaultLicense: FCC_BDC_DEFAULT_LICENSE,
-        description: "FCC Broadband Data Collection — Fabric-derived BSL addresses (public-domain); SQLite DB the operator builds via the isp-nexus BDC ETL.",
-        async *rows(opts) {
-            if (opts.country && opts.country !== "US") {
-                throw new Error(`fcc-bdc adapter: only US supported, got country=${opts.country}`);
-            }
-            const db = new DatabaseClient(opts.inputPath, { readOnly: true });
-            let emitted = 0;
-            try {
-                const stmt = db.prepare(`SELECT location_id, address_primary, city, state, zip, zip_suffix
-					 FROM bdc_locations
-					 ORDER BY location_id`);
-                for (const row of stmt.iterate()) {
-                    if (opts.signal?.aborted)
-                        return;
-                    if (opts.limit !== undefined && emitted >= opts.limit)
-                        return;
-                    const split = splitStreetLine(row.address_primary ?? "");
-                    if (!split)
-                        continue;
-                    const state = lookupStateAbbreviation(row.state);
-                    if (!state)
-                        continue;
-                    const locality = row.city?.trim();
-                    if (!locality)
-                        continue;
-                    const postcode = buildPostcode(row.zip ?? "", row.zip_suffix ?? null);
-                    if (!postcode)
-                        continue;
-                    const components = {
-                        ...(split.house_number ? { house_number: split.house_number } : {}),
-                        street: split.street,
-                        locality,
-                        region: state.abbreviation,
-                        postcode,
-                    };
-                    const raw = formatAddress(components, "US", { separator: ", " });
-                    if (!raw)
-                        continue;
-                    const aligned = reconcileComponents(components, raw);
-                    if (!Object.keys(aligned).length)
-                        continue;
-                    yield {
-                        raw,
-                        components: aligned,
-                        country: "US",
-                        locale: "en-US",
-                        source: FCC_BDC_ADAPTER_ID,
-                        source_id: `${FCC_BDC_ADAPTER_ID}-${row.location_id}`,
-                        corpus_version: "",
-                        license: FCC_BDC_DEFAULT_LICENSE,
-                    };
-                    emitted++;
-                }
-            }
-            finally {
-                await db.destroy();
-            }
-        },
-    };
-}
-/**
- * The configured adapter instance registered with the corpus builder.
- */
-export const fccBdcAdapter = createFccBdcAdapter();
-//# sourceMappingURL=adapter.js.map

package/out/adapters/fcc-bdc/adapter.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../lib/adapters/fcc-bdc/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAkCG;AAGH,OAAO,EAAE,aAAa,EAAE,mBAAmB,EAAE,MAAM,sBAAsB,CAAA;AACzE,OAAO,EAAE,cAAc,EAAE,MAAM,0BAA0B,CAAA;AAEzD,OAAO,EAAE,eAAe,EAAE,MAAM,iBAAiB,CAAA;AACjD,OAAO,EAAE,uBAAuB,EAAE,MAAM,sBAAsB,CAAA;AAG9D;;;GAGG;AACH,MAAM,CAAC,MAAM,kBAAkB,GAAG,SAAS,CAAA;AAC3C;;;GAGG;AACH,MAAM,CAAC,MAAM,uBAAuB,GAAG,eAAe,CAAA;AActD;;;;;;;;;;;GAWG;AACH,MAAM,UAAU,aAAa,CAAC,GAAW,EAAE,MAAqB;IAC/D,MAAM,CAAC,GAAG,GAAG,CAAC,IAAI,EAAE,CAAA;IAEpB,IAAI,CAAC,CAAC;QAAE,OAAO,EAAE,CAAA;IACjB,MAAM,CAAC,GAAG,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,CAAA;IAE9B,IAAI,CAAC,CAAC;QAAE,OAAO,CAAC,CAAA;IAEhB,IAAI,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,OAAO,CAAC,CAAA;IAE7B,OAAO,GAAG,CAAC,IAAI,CAAC,EAAE,CAAA;AACnB,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,mBAAmB;IAClC,OAAO;QACN,EAAE,EAAE,kBAAkB;QACtB,cAAc,EAAE,uBAAuB;QACvC,WAAW,EACV,wIAAwI;QAEzI,KAAK,CAAC,CAAC,IAAI,CAAC,IAAoB;YAC/B,IAAI,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,OAAO,KAAK,IAAI,EAAE,CAAC;gBAC3C,MAAM,IAAI,KAAK,CAAC,mDAAmD,IAAI,CAAC,OAAO,EAAE,CAAC,CAAA;YACnF,CAAC;YAED,MAAM,EAAE,GAAG,IAAI,cAAc,CAAc,IAAI,CAAC,SAAS,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAA;YAC9E,IAAI,OAAO,GAAG,CAAC,CAAA;YAEf,IAAI,CAAC;gBACJ,MAAM,IAAI,GAAG,EAAE,CAAC,OAAO,CACtB;;2BAEsB,CACtB,CAAA;gBAED,KAAK,MAAM,GAAG,IAAI,IAAI,CAAC,OAAO,EAAsC,EAAE,CAAC;oBACtE,IAAI,IAAI,CAAC,MAAM,EAAE,OAAO;wBAAE,OAAM;oBAEhC,IAAI,IAAI,CAAC,KAAK,KAAK,SAAS,IAAI,OAAO,IAAI,IAAI,CAAC,KAAK;wBAAE,OAAM;oBAE7D,MAAM,KAAK,GAAG,eAAe,CAAC,GAAG,CAAC,eAAe,IAAI,EAAE,CAAC,CAAA;oBAExD,IAAI,CAAC,KAAK;wBAAE,SAAQ;oBACpB,MAAM,KAAK,GAAG,uBAAuB,CAAC,GAAG,CAAC,KAAK,CAAC,CAAA;oBAEhD,IAAI,CAAC,KAAK;wBAAE,SAAQ;oBACpB,MAAM,QAAQ,GAAG,GAAG,CAAC,IAAI,EAAE,IAAI,EAAE,CAAA;oBAEjC,IAAI,CAAC,QAAQ;wBAAE,SAAQ;oBACvB,MAAM,QAAQ,GAAG,aAAa,CAAC,GAAG,CAAC,GAAG,IAAI,EAAE,EAAE,GAAG,CAAC,UAAU,IAAI,IAAI,CAAC,CAAA;oBAErE,IAAI,CAAC,QAAQ;wBAAE,SAAQ;oBAEvB,MAAM,UAAU,GAA+B;wBAC9C,GAAG,CAAC,KAAK,CAAC,YAAY,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,KAAK,CAAC,YAAY,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;wBACnE,MAAM,EAAE,KAAK,CAAC,MAAM;wBACpB,QAAQ;wBACR,MAAM,EAAE,KAAK,CAAC,YAAY;wBAC1B,QAAQ;qBACR,CAAA;oBAED,MAAM,GAAG,GAAG,aAAa,CAAC,UAAU,EAAE,IAAI,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAA;oBAEhE,IAAI,CAAC,GAAG;wBAAE,SAAQ;oBAClB,MAAM,OAAO,GAAG,mBAAmB,CAAC,UAAU,EAAE,GAAG,CAAC,CAAA;oBAEpD,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,MAAM;wBAAE,SAAQ;oBAE1C,MAAM;wBACL,GAAG;wBACH,UAAU,EAAE,OAAO;wBACnB,OAAO,EAAE,IAAI;wBACb,MAAM,EAAE,OAAO;wBACf,MAAM,EAAE,kBAAkB;wBAC1B,SAAS,EAAE,GAAG,kBAAkB,IAAI,GAAG,CAAC,WAAW,EAAE;wBACrD,cAAc,EAAE,EAAE;wBAClB,OAAO,EAAE,uBAAuB;qBAChC,CAAA;oBAED,OAAO,EAAE,CAAA;gBACV,CAAC;YACF,CAAC;oBAAS,CAAC;gBACV,MAAM,EAAE,CAAC,OAAO,EAAE,CAAA;YACnB,CAAC;QACF,CAAC;KACD,CAAA;AACF,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,aAAa,GAAG,mBAAmB,EAAE,CAAA"}

package/out/adapters/geonames-postal/adapter.d.ts

@@ -1,54 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   `geonames-postal`: GeoNames postal-code dump consumer (https://www.geonames.org/, CC-BY-4.0).
- *
- *   The GeoNames postal export (`https://download.geonames.org/export/zip/<CC>.zip`) is a clean,
- *   per-country `postcode → place → admin1` table with the place + region NAMES inline (no aux-file
- *   join needed). It broadens the corpus's postcode→locality→region coverage to ~80 countries, well
- *   beyond `wof-postalcode`/the coordinate-first table — forward coverage for the multi-locale
- *   goal.
- *
- *   Input: a per-country postal dump (`<CC>.txt`, 12 tab-separated columns, no header): country,
- *   postcode, place, admin1_name, admin1_code, admin2__, admin3__, lat, lon, accuracy.
- *
- *   Output: per row, postcode-FIRST (international) variants — the common order for the non-US
- *   locales this fills (US postcodes are already covered by TIGER/WOF, which use postcode-LAST):
- *
- *   1. `{ postcode, locality }` → "AD100 Canillo"
- *   2. `{ postcode, locality, region }` → "AD100 Canillo, Canillo" Prefer configuring this adapter for
- *        non-US countries; for US, the postcode-last sources are the right order. License:
- *        `"CC-BY-4.0"` per row (attribute "GeoNames").
- */
-import type { CorpusAdapter } from "#types";
-/**
- * Registry id for this adapter. Stamped into every row it emits, so a corpus record can be traced back to the dataset
- * it came from.
- */
-export declare const GEONAMES_POSTAL_ADAPTER_ID = "geonames-postal";
-/**
- * License carried by this source (CC-BY-4.0), attached to each row so downstream consumers inherit the terms rather
- * than having to look them up.
- */
-export declare const GEONAMES_POSTAL_DEFAULT_LICENSE = "CC-BY-4.0";
-/**
- * GeoNames postal-dump columns (0-based): country, postcode, place, admin1_name, admin1_code, admin2_name, ….
- *
- * Shared with `tools/postcode-triples.ts`, which additionally reads `admin2Name` — the CITY for the IN/MX/PT-shaped
- * exports whose `place` column is a street or colonia.
- */
-export declare const GEONAMES_POSTAL_COLUMNS: {
-    readonly country: 0;
-    readonly postcode: 1;
-    readonly place: 2;
-    readonly admin1Name: 3;
-    readonly admin2Name: 5;
-};
-export declare function createGeonamesPostalAdapter(): CorpusAdapter;
-/**
- * The configured adapter instance registered with the corpus builder.
- */
-export declare const geonamesPostalAdapter: CorpusAdapter;
-//# sourceMappingURL=adapter.d.ts.map

package/out/adapters/geonames-postal/adapter.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../lib/adapters/geonames-postal/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;GAuBG;AAMH,OAAO,KAAK,EAAgC,aAAa,EAAE,MAAM,QAAQ,CAAA;AAEzE;;;GAGG;AACH,eAAO,MAAM,0BAA0B,oBAAoB,CAAA;AAC3D;;;GAGG;AACH,eAAO,MAAM,+BAA+B,cAAc,CAAA;AAE1D;;;;;GAKG;AACH,eAAO,MAAM,uBAAuB;;;;;;CAA+E,CAAA;AAEnH,wBAAgB,2BAA2B,IAAI,aAAa,CAkE3D;AAED;;GAEG;AACH,eAAO,MAAM,qBAAqB,eAAgC,CAAA"}