@mailwoman/corpus 9.4.0 → 10.0.0

package/lib/tools/sub-venue/wikidata.ts

@@ -1,85 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   @file The Wikidata leg of the sub-venue lexicon — turning the designator-label SPARQL payload into
- *   surfaces.
- *
- *   Reads an already-parsed payload rather than issuing a query: the acquisition lives in
- *   `fetch/wikidata-subvenue.ts`, and keeping the conversion a pure function of parsed input is what
- *   lets the builder stay deterministic and testable with no network and no fixture on disk.
- *
- *   Wikidata gives a CONCEPT NAME per language, not a designator as addressed. Q849706's Spanish label
- *   is `terminal aeroportuaria`; the addressed form is `Terminal`. So every surface produced here lands
- *   `curated: false`, and the head-noun derivation is what proposes the addressed form from it.
- */
-
-import { normalizeSurface } from "#tools/sub-venue/surfaces"
-import { CONCEPT_QIDS, type SubVenueSurface } from "#tools/sub-venue/table"
-
-/**
- * The SPARQL results envelope, narrowed to the columns the designator-label query produces.
- */
-interface SPARQLBinding {
-	item?: { value: string }
-	lang?: { value: string }
-	label?: { value: string }
-	kind?: { value: string }
-}
-
-interface SPARQLEnvelope {
-	results?: { bindings?: SPARQLBinding[] }
-}
-
-/**
- * Turn the Wikidata designator-label payload into surfaces.
- *
- * A row is dropped when its language tag is empty (an untagged literal, which Wikidata occasionally carries), when the
- * QID maps to no designator in {@link CONCEPT_QIDS}, or when the normalized phrase is empty. Everything that survives
- * lands `curated: false` — see the module docstring.
- */
-export function surfacesFromWikidata(
-	payload: unknown,
-	conceptQIDs: Readonly<Record<string, string>> = CONCEPT_QIDS
-): SubVenueSurface[] {
-	const byQID = new Map(Object.entries(conceptQIDs).map(([id, qid]) => [qid, id]))
-	const envelope = payload as SPARQLEnvelope
-	const seen = new Set<string>()
-	const out: SubVenueSurface[] = []
-
-	for (const binding of envelope.results?.bindings ?? []) {
-		const qid = binding.item?.value?.split("/").pop()
-		const recordID = qid ? byQID.get(qid) : undefined
-		const lang = binding.lang?.value
-		const raw = binding.label?.value
-
-		if (!recordID || !lang || !raw) continue
-
-		const phrase = normalizeSurface(raw)
-
-		if (!phrase) continue
-
-		const source = binding.kind?.value === "alt" ? "wikidata:alt" : "wikidata:label"
-		// A concept can carry the same string as both a label and an alias, and across dialect subtags
-		// (`zh`, `zh-cn`, `zh-hans` all say 航站楼). Key the dedupe on the tuple that identifies a row.
-		const key = `${phrase}\0${recordID}\0${lang}\0${source}`
-
-		if (seen.has(key)) continue
-		seen.add(key)
-
-		out.push({
-			phrase,
-			recordID,
-			recordKind: "designator",
-			lang,
-			region: "",
-			source,
-			curated: false,
-			observations: 0,
-			context: {},
-		})
-	}
-
-	return out
-}

package/lib/tools/sub-venue-lexicon.ts

@@ -1,429 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   @file Build the sub-venue designator lexicon (#35) — the vocabulary a corpus slice and, eventually,
- *   the span proposer read to recognize `Terminal 5`, `North Terminal`, `Concourse B`, `ターミナル1` as
- *   venue-INTERIOR structure. This is the assembly: the record schema lives in `sub-venue/table.ts`, the
- *   implementation in its siblings, and the curation decisions in `sub-venue-promotions.ts`.
- *
- *   Reads the fetch outputs (`mailwoman corpus fetch wikidata-subvenue`, a JSONL of
- *   `@mailwoman/osm/sdk`'s `SubVenueSourceRow`s per region, and the Overture slice of `poi.db` via
- *   `overture-subvenue.ts`) and emits one committed JSON table.
- *
- *   ── Determinism ──────────────────────────────────────────────────────────────────────────────────
- *   {@link buildSubVenueLexicon} is a PURE function of its inputs with a stable sort on every array, so
- *   a regenerate against the same fetch outputs is byte-identical. No timestamp is emitted for the same
- *   reason `taxonomy.json` carries none — a clock in the artifact makes every regenerate a diff.
- *   Vintages live in `sources[]`, taken from the fetch manifests.
- *
- *   ── Where the stages live ────────────────────────────────────────────────────────────────────────
- *   Each stage carries the measurements that shaped it; the order they run in is
- *   {@link buildSubVenueLexicon}'s own docstring, and it is required.
- *
- *   - `sub-venue/table.ts` — the emitted record schema plus the shipped seed vocabulary.
- *   - `sub-venue/surfaces.ts` — phrase normalization, the phrase → record index, and the name-match
- *     operator that gates the harvest.
- *   - `sub-venue/wikidata.ts` — the designator-label SPARQL payload turned into surfaces.
- *   - `sub-venue/head-nouns.ts` — the addressed form derived from an encyclopaedic label.
- *   - `sub-venue/harvest.ts` — the harvestable row shape, its JSONL reader, and the attestation pass.
- *
- *   ── What `curated: false` means, and how a surface stops being it ────────────────────────────────
- *   Every machine-derived surface lands `curated: false`, and {@link SubVenueLexiconTable} consumers
- *   that gate parsing MUST filter to `curated: true`. A surface becomes curated ONLY by matching a
- *   {@link SubVenuePromotion} in `sub-venue-promotions.ts` — a per-designator, per-LOCALE decision
- *   carrying the census that backs it. Promotion is per-locale because the same token is a designator in
- *   one language and a disaster in another: `hall` is `Halle 2` at Frankfurt and `Village Hall` at 3,205
- *   British bus stops.
- */
-
-import { readLocalJSONFile, statPath } from "@mailwoman/core/fs/readers"
-import { writeLocalFile } from "@mailwoman/core/fs/writers"
-import { isoDate } from "@mailwoman/core/utils"
-import { basename, join } from "path-ts"
-
-import { SUBVENUE_PROMOTIONS, type SubVenuePromotion } from "#tools/sub-venue-promotions"
-import { extractAttestedPhrases, readSubVenueJSONL, type SubVenueHarvestRow } from "#tools/sub-venue/harvest"
-import { deriveHeadNounSurfaces } from "#tools/sub-venue/head-nouns"
-import { buildSurfaceIndex } from "#tools/sub-venue/surfaces"
-import {
-	CONCEPT_QIDS,
-	type IdentifierShape,
-	LexiconTier,
-	PROPOSED_DESIGNATORS,
-	SHIPPED_DESIGNATOR_SEED,
-	SHIPPED_MODIFIER_SEED,
-	type SubVenueDesignator,
-	type SubVenueLexiconSource,
-	type SubVenueLexiconTable,
-	type SubVenueModifier,
-	type SubVenueSurface,
-	SUBVENUE_LEXICON_VERSION,
-} from "#tools/sub-venue/table"
-import { surfacesFromWikidata } from "#tools/sub-venue/wikidata"
-
-export * from "#tools/sub-venue/harvest"
-export * from "#tools/sub-venue/head-nouns"
-export * from "#tools/sub-venue/surfaces"
-export * from "#tools/sub-venue/table"
-export * from "#tools/sub-venue/wikidata"
-
-/**
- * Apply the curation decisions to a surface list, IN PLACE on a copy.
- *
- * A promotion binds `(designatorID, phrase, locale)`. A surface matches when it names the same record with the same
- * phrase and its language is the locale's language OR the untagged `und` — the default `name` tag carries no language,
- * and a German extract's untagged `Halle 2` is German.
- *
- * REGION is the subtle half. A surface attested in an extract carries that extract's region and matches only its own
- * locale. A surface with `region: ""` is region-FREE — a Wikidata label or a derived head noun — and a promotion
- * reaches it only when no REJECTION exists for the same designator, phrase and language anywhere else. That guard is
- * not decoration: `pier` is promoted for en-GB and rejected for en-US, and without it the en-GB decision would curate
- * the region-free English surface and hand `Pier 1 Imports` the promotion en-US was refused. Where no rejection
- * competes — `terminal` in `es`, `ターミナル` in `ja` — the region-free surface is the whole point, since a language's
- * designator does not stop at a border.
- *
- * Rejections mark nothing themselves. They exist in `promotions[]` as the record of a decision taken, so the next
- * reader meets en-GB `hall`'s 3,204 bus stops before re-proposing it, not after.
- */
-export function applyPromotions(
-	surfaces: readonly SubVenueSurface[],
-	promotions: readonly SubVenuePromotion[]
-): SubVenueSurface[] {
-	const language = (locale: string): string => locale.split("-")[0]!
-	const region = (locale: string): string => locale.split("-")[1] ?? ""
-
-	const rejectedLanguages = new Set(
-		promotions
-			.filter((promotion) => promotion.decision === "reject")
-			.map((promotion) => `${promotion.designatorID} ${promotion.phrase} ${language(promotion.locale)}`)
-	)
-
-	const promoted = promotions.filter((promotion) => promotion.decision === "promote")
-
-	return surfaces.map((surface) => {
-		if (surface.curated) return surface
-
-		const hit = promoted.some((promotion) => {
-			if (promotion.designatorID !== surface.recordID || promotion.phrase !== surface.phrase) return false
-
-			const lang = language(promotion.locale)
-
-			if (surface.lang !== lang && surface.lang !== "und") return false
-
-			if (surface.region === region(promotion.locale)) return true
-
-			return surface.region === "" && !rejectedLanguages.has(`${promotion.designatorID} ${promotion.phrase} ${lang}`)
-		})
-
-		return hit ? { ...surface, curated: true } : surface
-	})
-}
-
-/**
- * One harvestable input: rows plus the stamp they carry into the table.
- */
-export interface SubVenueHarvest {
-	rows: readonly SubVenueHarvestRow[]
-	source?: string
-	region?: string
-}
-
-/**
- * Everything {@link buildSubVenueLexicon} needs, already parsed. Keeping the builder off the filesystem is what makes it
- * deterministic and testable without fixtures on disk.
- */
-export interface BuildSubVenueLexiconInput {
-	/**
-	 * The raw `designator-labels.json` SPARQL envelope, or `null` to build the seed-only table.
-	 */
-	wikidata: unknown | null
-	/**
-	 * Every harvestable source, in the order they should contribute. Order matters only for the surface INDEX: a source
-	 * can match a phrase an earlier source introduced, never a later one.
-	 */
-	harvests: readonly SubVenueHarvest[]
-	/**
-	 * Provenance rows, copied off the fetch manifests by the caller.
-	 */
-	sources: readonly SubVenueLexiconSource[]
-	/**
-	 * Curation decisions. Defaults to the committed {@link SUBVENUE_PROMOTIONS}; pass an empty array to build the
-	 * pre-curation table (which is what the promotion census itself is taken against).
-	 */
-	promotions?: readonly SubVenuePromotion[]
-}
-
-/**
- * Build the lexicon table. PURE and deterministic — same inputs, byte-identical output.
- *
- * Order of operations is required in three places:
- *
- * 1. Seed surfaces are inserted before anything else, so `terminal` indexes to the `terminal` designator rather than to
- *    whichever Wikidata alias sorts first.
- * 2. Head nouns are derived AFTER Wikidata and BEFORE the harvests, because `ターミナル` has to exist as a surface before a
- *    Japanese extract can be searched for it. That ordering is the entire reason the Japan harvest finds anything — see
- *    `PROVENANCE.md`.
- * 3. Promotions are applied LAST, over the union, so a decision can promote a surface whichever source produced it.
- */
-export function buildSubVenueLexicon(input: BuildSubVenueLexiconInput): SubVenueLexiconTable {
-	const designators: SubVenueDesignator[] = SHIPPED_DESIGNATOR_SEED.map((seed) => ({
-		id: seed.id,
-		tier: LexiconTier.SubVenue,
-		modifierEligible: seed.modifierEligible,
-		shipped: true,
-		provenance: [...seed.provenance],
-	}))
-
-	const byID = new Map(designators.map((d) => [d.id, d]))
-
-	for (const proposed of PROPOSED_DESIGNATORS) {
-		const existing = byID.get(proposed.id)
-
-		if (existing) {
-			existing.provenance = [...new Set([...existing.provenance, ...proposed.provenance])]
-
-			continue
-		}
-
-		const record: SubVenueDesignator = {
-			id: proposed.id,
-			tier: proposed.tier,
-			modifierEligible: false,
-			shipped: false,
-			provenance: [...proposed.provenance],
-		}
-
-		designators.push(record)
-		byID.set(record.id, record)
-	}
-
-	// A Wikidata concept id is provenance for the designator it names, whether or not the concept
-	// contributed a usable surface.
-	for (const [id, qid] of Object.entries(CONCEPT_QIDS)) {
-		const record = byID.get(id)
-
-		if (record) {
-			record.provenance = [...new Set([...record.provenance, `wikidata:${qid}`])]
-		}
-	}
-
-	const modifiers: SubVenueModifier[] = SHIPPED_MODIFIER_SEED.map((id) => ({
-		id,
-		shipped: true,
-		provenance: ["codex:directionals"],
-	}))
-
-	const surfaces: SubVenueSurface[] = [
-		...designators.map((d): SubVenueSurface => ({
-			phrase: d.id,
-			recordID: d.id,
-			recordKind: "designator",
-			lang: "en",
-			region: "",
-			source: "seed",
-			// The English designator IS the shipped vocabulary — curated by construction.
-			curated: d.shipped,
-			observations: 0,
-			context: {},
-		})),
-		...modifiers.map((m): SubVenueSurface => ({
-			phrase: m.id,
-			recordID: m.id,
-			recordKind: "modifier",
-			lang: "en",
-			region: "",
-			source: "seed",
-			curated: true,
-			observations: 0,
-			context: {},
-		})),
-	]
-
-	if (input.wikidata) {
-		surfaces.push(...surfacesFromWikidata(input.wikidata))
-	}
-
-	surfaces.push(...deriveHeadNounSurfaces(surfaces))
-
-	const identifierShapes: IdentifierShape[] = []
-
-	for (const harvest of input.harvests) {
-		const attested = extractAttestedPhrases(harvest.rows, buildSurfaceIndex(surfaces), {
-			source: harvest.source,
-			region: harvest.region,
-		})
-
-		surfaces.push(...attested.surfaces)
-		identifierShapes.push(...attested.identifierShapes)
-	}
-
-	const promotions = [...(input.promotions ?? SUBVENUE_PROMOTIONS)]
-	const curated = applyPromotions(surfaces, promotions)
-
-	// Deterministic order everywhere. `localeCompare` matches the tie-break discipline
-	// `generate-taxonomy.ts` and `build-brands.ts` already use.
-	designators.sort((a, b) => a.id.localeCompare(b.id))
-	modifiers.sort((a, b) => a.id.localeCompare(b.id))
-
-	curated.sort(
-		(a, b) =>
-			a.phrase.localeCompare(b.phrase) ||
-			a.recordID.localeCompare(b.recordID) ||
-			a.lang.localeCompare(b.lang) ||
-			a.region.localeCompare(b.region) ||
-			a.source.localeCompare(b.source)
-	)
-
-	return {
-		version: SUBVENUE_LEXICON_VERSION,
-		sources: input.sources.toSorted((a, b) => a.id.localeCompare(b.id)),
-		designators,
-		modifiers,
-		surfaces: curated,
-		identifierShapes: identifierShapes.toSorted(
-			(a, b) =>
-				a.designatorID.localeCompare(b.designatorID) ||
-				a.region.localeCompare(b.region) ||
-				a.shape.localeCompare(b.shape)
-		),
-		promotions: promotions.toSorted(
-			(a, b) =>
-				a.designatorID.localeCompare(b.designatorID) ||
-				a.locale.localeCompare(b.locale) ||
-				a.phrase.localeCompare(b.phrase)
-		),
-	}
-}
-
-/**
- * Serialize the table the way the committed artifact stores it: pretty-printed, trailing newline. Run `oxfmt` over the
- * result before committing — repo law is that committed JSON is oxfmt-clean, which `JSON.stringify` cannot reproduce.
- */
-export function serializeSubVenueLexicon(table: SubVenueLexiconTable): string {
-	return JSON.stringify(table, null, 2) + "\n"
-}
-
-/**
- * The Wikidata fetch manifest's shape, narrowed to the fields the lexicon copies into `sources[]`.
- */
-interface WikidataFetchManifest {
-	endpoint?: string
-	license?: string
-	downloaded_at?: string
-	files?: Array<{ filename?: string; rows?: number }>
-}
-
-/**
- * One OSM extract to harvest: the JSONL path plus the ISO country its rows describe.
- */
-export interface SubVenueExtractInput {
-	path: string
-	region: string
-}
-
-export interface GenerateSubVenueLexiconOptions {
-	/**
-	 * Directory holding the `mailwoman corpus fetch wikidata-subvenue` output. Omit to build the seed-only table.
-	 */
-	wikidataDir?: string
-	/**
-	 * OSM extract JSONLs, one per region.
-	 */
-	extracts?: readonly SubVenueExtractInput[]
-	/**
-	 * Already-read Overture rows (`readOvertureSubVenues`), grouped by the caller. Kept as a parameter rather than a path
-	 * so this function stays free of a 3.9 GB database dependency — the CLI opens `poi.db`, this assembles.
-	 */
-	overtureRows?: readonly (SubVenueHarvestRow & { country: string })[]
-	/**
-	 * `poi.db`'s layer vintage, for `sources[]`. Only read when `overtureRows` is non-empty.
-	 */
-	overtureVintage?: string
-	/**
-	 * Where the table is written.
-	 */
-	outPath: string
-}
-
-/**
- * Read the fetch outputs, build the table, and write it.
- *
- * The IO half only — every decision lives in {@link buildSubVenueLexicon}, which is pure. Run `oxfmt` over `outPath`
- * afterwards; repo law is that committed JSON is oxfmt-clean.
- */
-export async function generateSubVenueLexicon(options: GenerateSubVenueLexiconOptions): Promise<SubVenueLexiconTable> {
-	const sources: SubVenueLexiconSource[] = []
-	const harvests: SubVenueHarvest[] = []
-	let wikidata: unknown = null
-
-	if (options.wikidataDir) {
-		wikidata = await readLocalJSONFile<unknown>(join(options.wikidataDir, "designator-labels.json"))
-
-		const manifest = await readLocalJSONFile<WikidataFetchManifest>(join(options.wikidataDir, "MANIFEST.json"))
-
-		const labelFile = manifest.files?.find((f) => f.filename === "designator-labels.json")
-
-		sources.push({
-			id: "wikidata",
-			origin: manifest.endpoint ?? "https://query.wikidata.org/sparql",
-			license: manifest.license ?? "CC0",
-			// The DATE only. A full ISO timestamp would make every re-fetch a diff in the committed artifact for
-			// no information a reader of a vocabulary table can act on.
-			retrieved: (manifest.downloaded_at ?? "").slice(0, 10),
-			rows: labelFile?.rows ?? 0,
-		})
-	}
-
-	for (const extract of options.extracts ?? []) {
-		const rows = await readSubVenueJSONL(extract.path)
-
-		harvests.push({ rows, source: "osm", region: extract.region })
-
-		sources.push({
-			id: `osm:${extract.region.toLowerCase()}`,
-			// The extract's NAME, never its path. `AGENTS.md` forbids re-hardcoding the lab data root
-			// anywhere, and a committed artifact carrying `/mnt/playpen/...` would do exactly that while
-			// telling a reader on another machine nothing. `great-britain` identifies the Geofabrik region,
-			// which is the fact that matters.
-			origin: `OpenStreetMap via Geofabrik (${basename(extract.path, ".jsonl")})`,
-			license: "ODbL (OpenStreetMap)",
-			// The extract's mtime — when the rows were produced. `corpus/AGENTS.md`'s standing warning that
-			// a file's mtime is not its DATA's vintage applies to a downloaded archive; this file is a build
-			// output of ours, so its mtime is exactly the right number.
-			retrieved: isoDate((await statPath(extract.path)).mtime),
-			rows: rows.length,
-		})
-	}
-
-	if (options.overtureRows?.length) {
-		// Overture rows carry their own country, so they are harvested per REGION rather than in one pass —
-		// a `region` on the surface is the axis promotion is decided on and a mixed-country bucket would
-		// make it meaningless.
-		const byCountry = new Map<string, SubVenueHarvestRow[]>()
-
-		for (const row of options.overtureRows) {
-			const bucket = byCountry.get(row.country) ?? []
-			bucket.push(row)
-			byCountry.set(row.country, bucket)
-		}
-
-		for (const [country, rows] of [...byCountry].toSorted((a, b) => a[0].localeCompare(b[0]))) {
-			harvests.push({ rows, source: "overture", region: country })
-		}
-
-		sources.push({
-			id: "overture",
-			origin: "Overture Maps Foundation places, via the poi.db spatial layer",
-			license: "CDLA-Permissive-2.0",
-			retrieved: options.overtureVintage ?? "",
-			rows: options.overtureRows.length,
-		})
-	}
-
-	const table = buildSubVenueLexicon({ wikidata, harvests, sources })
-	await writeLocalFile(serializeSubVenueLexicon(table), options.outPath)
-
-	return table
-}

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.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.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"}