@mailwoman/corpus 10.5.0 → 10.7.0

package/out/us/tools/nad-skeletons.js

@@ -0,0 +1,262 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ * @file Samples National Address Database rows into OpenAddresses-layout CSV skeletons, measures the address
+ *   shapes each state's rows carry, and tallies county ZIP-to-city tables.
+ *
+ *   The OA-skeleton recipes read a zipped CSV with OpenAddresses' columns, so a NAD sample written in that
+ *   layout reaches them through the same reader as the OpenAddresses extracts they also read.
+ */
+import { matchLeadingDesignator } from "@mailwoman/codex/us";
+import { pathExists } from "@mailwoman/core/fs/readers/stat";
+import { writeLocalBuffer, writeLocalJSONFile, writeLocalTextFile } from "@mailwoman/core/fs/writers";
+import { sha256Hex } from "@mailwoman/core/hash";
+import { tryParsingJSON } from "@mailwoman/core/json";
+import { mulberry32 } from "@mailwoman/core/random";
+import ADMZip from "adm-zip";
+import { resolvePathBuilder } from "path-ts";
+import { TextSpliterator } from "spliterator";
+import { Globerator } from "spliterator/node/fs";
+import { composeNADLocality, decomposeNADStreet, nadHouseNumber } from "#us/adapters/usgov/nad/adapter";
+import { nadSkeletonFile, nadZipCityPath } from "#us/recipes/skeleton-sources";
+/**
+ * The OpenAddresses columns a skeleton CSV carries, in OpenAddresses' order.
+ */
+export const NAD_SKELETON_COLUMNS = [
+    "NUMBER",
+    "STREET",
+    "UNIT",
+    "CITY",
+    "DISTRICT",
+    "REGION",
+    "POSTCODE",
+    "ID",
+];
+const RURAL_ROUTE = /^(RR|R R|RURAL ROUTE|HC|HCR|HIGHWAY CONTRACT)\b/i;
+const text = (value) => (value == null ? "" : String(value).trim());
+/**
+ * The unit identifier of a NAD row without its designator, as OpenAddresses' `UNIT` column mostly states it.
+ *
+ * NAD writes `APT 6`, `UNIT D` or a bare `200`; the recipes add their own designator in
+ * front of the identifier, so the designator NAD wrote is dropped rather than doubled.
+ */
+export function nadUnitIdentifier(record) {
+    const unit = text(record.Unit) || text(record.Room);
+    if (!unit)
+        return "";
+    const designator = matchLeadingDesignator(unit);
+    const identifier = designator ? unit.slice(unit.indexOf(designator.matched) + designator.matched.length) : unit;
+    return identifier.replace(/^\s*#\s*/, "").trim();
+}
+/**
+ * Projects one NAD record onto the OpenAddresses columns, or returns null
+ * when it lacks a house number, street, locality, state or five-digit ZIP.
+ *
+ * A street whose name NAD publishes in capitals is written wholly in capitals,
+ * because NAD spells its structured types in title case (`East MAIN Street`)
+ * and a recipe would otherwise teach that mixed case.
+ */
+export function nadSkeletonRow(record) {
+    const number = nadHouseNumber(record);
+    const street = decomposeNADStreet(record);
+    const city = composeNADLocality(record);
+    const region = text(record.State).toUpperCase();
+    const postcode = text(record.Zip_Code);
+    if (!number || !street?.full || !city || !region || !/^\d{5}$/.test(postcode))
+        return null;
+    const name = street.street ?? street.full;
+    const capitalized = /[A-Z]/.test(name) && name === name.toUpperCase();
+    return {
+        NUMBER: number,
+        STREET: capitalized ? street.full.toUpperCase() : street.full,
+        UNIT: nadUnitIdentifier(record),
+        CITY: city,
+        DISTRICT: text(record.County),
+        REGION: region,
+        POSTCODE: postcode,
+        ID: text(record.UUID) || text(record.OBJECTID),
+    };
+}
+function csvCell(value) {
+    return /[",\r\n]/.test(value) ? `"${value.replaceAll('"', '""')}"` : value;
+}
+const newTally = () => ({
+    usable: 0,
+    unit: 0,
+    types: new Set(),
+    zips: new Set(),
+    prefixType: 0,
+    numericName: 0,
+    postDirectional: 0,
+    ruralRoute: 0,
+    hyphenated: 0,
+    grid: 0,
+});
+function tallyShape(tally, record, row) {
+    tally.usable++;
+    if (row.UNIT) {
+        tally.unit++;
+    }
+    const postType = text(record.St_PosTyp).toUpperCase();
+    const preType = text(record.St_PreTyp).toUpperCase();
+    if (postType) {
+        tally.types.add(postType);
+    }
+    if (preType) {
+        tally.types.add(preType);
+        tally.prefixType++;
+    }
+    tally.zips.add(row.POSTCODE);
+    if (/^\d+$/.test(text(record.St_Name))) {
+        tally.numericName++;
+    }
+    if (text(record.St_PosDir)) {
+        tally.postDirectional++;
+    }
+    if (RURAL_ROUTE.test(row.STREET)) {
+        tally.ruralRoute++;
+    }
+    if (/^\d+-\d+/.test(row.NUMBER)) {
+        tally.hyphenated++;
+    }
+    if (/^[NSEW]\d/i.test(row.NUMBER)) {
+        tally.grid++;
+    }
+}
+function shapeOf(region, t) {
+    const share = (n) => (t.usable ? n / t.usable : 0);
+    return {
+        region,
+        usable: t.usable,
+        unitShare: share(t.unit),
+        streetTypes: t.types.size,
+        zips: t.zips.size,
+        prefixTypeShare: share(t.prefixType),
+        numericNameShare: share(t.numericName),
+        postDirectionalShare: share(t.postDirectional),
+        ruralRouteShare: share(t.ruralRoute),
+        hyphenatedNumberShare: share(t.hyphenated),
+        gridNumberShare: share(t.grid),
+    };
+}
+/**
+ * Reads every NAD page once, measures every state's address shapes,
+ * keeps a uniform reservoir sample of at most `cap` rows for each requested state,
+ * and tallies the majority city of each ZIP in each requested county.
+ *
+ * The sample is written in reservoir order rather than page order, so a reader that
+ * takes the first N rows still draws across the whole state.
+ * It refuses to replace a file that already exists.
+ */
+export async function buildNADSkeletons(options) {
+    const { inputDirectory, cap, seed } = options;
+    const states = new Set(options.states.map((s) => s.toUpperCase()));
+    const outputs = [
+        ...[...states].map((region) => nadSkeletonFile(region).zip),
+        ...options.zipCityCounties.map(nadZipCityPath),
+    ];
+    for (const output of outputs) {
+        if (await pathExists(output))
+            throw new Error(`${output} already exists; a skeleton artifact is never replaced`);
+    }
+    const random = mulberry32(seed);
+    const tallies = new Map();
+    const reservoirs = new Map();
+    const zipCity = new Map();
+    const countyKey = (region, county) => `${region}|${county.toLowerCase()}`;
+    const wantedCounties = new Set(options.zipCityCounties.map((c) => countyKey(c.region, c.county)));
+    const pages = await Globerator.files("ndjson", { cwd: inputDirectory, absolute: false, recursive: false }).toSorted();
+    let pagesRead = 0;
+    for (const page of pages) {
+        for await (const line of TextSpliterator.fromAsync(resolvePathBuilder(inputDirectory, page))) {
+            if (!line)
+                continue;
+            const record = tryParsingJSON(line);
+            if (!record)
+                continue;
+            const row = nadSkeletonRow(record);
+            if (!row)
+                continue;
+            let tally = tallies.get(row.REGION);
+            if (!tally) {
+                tallies.set(row.REGION, (tally = newTally()));
+            }
+            tallyShape(tally, record, row);
+            if (states.has(row.REGION)) {
+                let reservoir = reservoirs.get(row.REGION);
+                if (!reservoir) {
+                    reservoirs.set(row.REGION, (reservoir = { seen: 0, rows: [] }));
+                }
+                reservoir.seen++;
+                if (reservoir.rows.length < cap) {
+                    reservoir.rows.push(row);
+                }
+                else {
+                    const slot = Math.floor(random() * reservoir.seen);
+                    if (slot < cap) {
+                        reservoir.rows[slot] = row;
+                    }
+                }
+            }
+            const key = countyKey(row.REGION, row.DISTRICT);
+            if (wantedCounties.has(key)) {
+                let byZip = zipCity.get(key);
+                if (!byZip) {
+                    zipCity.set(key, (byZip = new Map()));
+                }
+                let byCity = byZip.get(row.POSTCODE);
+                if (!byCity) {
+                    byZip.set(row.POSTCODE, (byCity = new Map()));
+                }
+                byCity.set(row.CITY, (byCity.get(row.CITY) ?? 0) + 1);
+            }
+        }
+        options.onPage?.(++pagesRead, pages.length);
+    }
+    const summary = {
+        pages: pages.length,
+        cap,
+        seed,
+        shapes: [...tallies].map(([region, t]) => shapeOf(region, t)).toSorted((a, b) => b.usable - a.usable),
+        samples: [],
+        zipCity: [],
+    };
+    for (const region of states) {
+        const reservoir = reservoirs.get(region);
+        if (!reservoir?.rows.length)
+            throw new Error(`NAD holds no usable row for ${region}`);
+        const { zip, csv } = nadSkeletonFile(region);
+        const body = [
+            NAD_SKELETON_COLUMNS.join(","),
+            ...reservoir.rows.map((row) => NAD_SKELETON_COLUMNS.map((column) => csvCell(row[column])).join(",")),
+            "",
+        ].join("\n");
+        const archive = new ADMZip();
+        archive.addFile(csv, Buffer.from(body, "utf8"));
+        await writeLocalBuffer(archive.toBuffer(), zip);
+        summary.samples.push({ region, rows: reservoir.rows.length, zip: String(zip), csv, csvSHA256: sha256Hex(body) });
+    }
+    for (const county of options.zipCityCounties) {
+        const byZip = zipCity.get(countyKey(county.region, county.county));
+        if (!byZip?.size)
+            throw new Error(`NAD holds no usable row for ${county.county} County, ${county.region}`);
+        const lines = ["ZIP,CITY,ROWS"];
+        for (const [zip, byCity] of [...byZip].toSorted(([a], [b]) => a.localeCompare(b))) {
+            const [city, rows] = [...byCity].toSorted((a, b) => b[1] - a[1] || a[0].localeCompare(b[0]))[0];
+            lines.push([zip, csvCell(city), String(rows)].join(","));
+        }
+        const body = `${lines.join("\n")}\n`;
+        const path = nadZipCityPath(county);
+        await writeLocalTextFile(body, path);
+        summary.zipCity.push({ ...county, zips: byZip.size, path: String(path), sha256: sha256Hex(body) });
+    }
+    return summary;
+}
+/**
+ * Writes the summary beside the skeletons it describes.
+ */
+export function writeNADSkeletonSummary(summary, path) {
+    return writeLocalJSONFile(summary, path);
+}
+//# sourceMappingURL=nad-skeletons.js.map

package/out/us/tools/nad-skeletons.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"nad-skeletons.js","sourceRoot":"","sources":["../../../lib/us/tools/nad-skeletons.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EAAE,sBAAsB,EAAE,MAAM,qBAAqB,CAAA;AAC5D,OAAO,EAAE,UAAU,EAAE,MAAM,iCAAiC,CAAA;AAC5D,OAAO,EAAE,gBAAgB,EAAE,kBAAkB,EAAE,kBAAkB,EAAE,MAAM,4BAA4B,CAAA;AACrG,OAAO,EAAE,SAAS,EAAE,MAAM,sBAAsB,CAAA;AAChD,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAA;AACrD,OAAO,EAAE,UAAU,EAAE,MAAM,wBAAwB,CAAA;AACnD,OAAO,MAAM,MAAM,SAAS,CAAA;AAC5B,OAAO,EAAE,kBAAkB,EAAwB,MAAM,SAAS,CAAA;AAClE,OAAO,EAAE,eAAe,EAAE,MAAM,aAAa,CAAA;AAC7C,OAAO,EAAE,UAAU,EAAE,MAAM,qBAAqB,CAAA;AAEhD,OAAO,EAAE,kBAAkB,EAAE,kBAAkB,EAAE,cAAc,EAAkB,MAAM,gCAAgC,CAAA;AACvH,OAAO,EAAE,eAAe,EAAE,cAAc,EAAyB,MAAM,8BAA8B,CAAA;AAErG;;GAEG;AACH,MAAM,CAAC,MAAM,oBAAoB,GAAG;IACnC,QAAQ;IACR,QAAQ;IACR,MAAM;IACN,MAAM;IACN,UAAU;IACV,QAAQ;IACR,UAAU;IACV,IAAI;CACK,CAAA;AAOV,MAAM,WAAW,GAAG,kDAAkD,CAAA;AACtE,MAAM,IAAI,GAAG,CAAC,KAAc,EAAU,EAAE,CAAC,CAAC,KAAK,IAAI,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC,CAAA;AAEpF;;;;;GAKG;AACH,MAAM,UAAU,iBAAiB,CAAC,MAAiB;IAClD,MAAM,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;IAEnD,IAAI,CAAC,IAAI;QAAE,OAAO,EAAE,CAAA;IAEpB,MAAM,UAAU,GAAG,sBAAsB,CAAC,IAAI,CAAC,CAAA;IAC/C,MAAM,UAAU,GAAG,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,OAAO,CAAC,GAAG,UAAU,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,CAAA;IAE/G,OAAO,UAAU,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAA;AACjD,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,cAAc,CAAC,MAAiB;IAC/C,MAAM,MAAM,GAAG,cAAc,CAAC,MAAM,CAAC,CAAA;IACrC,MAAM,MAAM,GAAG,kBAAkB,CAAC,MAAM,CAAC,CAAA;IACzC,MAAM,IAAI,GAAG,kBAAkB,CAAC,MAAM,CAAC,CAAA;IACvC,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,WAAW,EAAE,CAAA;IAC/C,MAAM,QAAQ,GAAG,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAA;IAEtC,IAAI,CAAC,MAAM,IAAI,CAAC,MAAM,EAAE,IAAI,IAAI,CAAC,IAAI,IAAI,CAAC,MAAM,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,QAAQ,CAAC;QAAE,OAAO,IAAI,CAAA;IAE1F,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,IAAI,MAAM,CAAC,IAAI,CAAA;IACzC,MAAM,WAAW,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,IAAI,KAAK,IAAI,CAAC,WAAW,EAAE,CAAA;IAErE,OAAO;QACN,MAAM,EAAE,MAAM;QACd,MAAM,EAAE,WAAW,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI;QAC7D,IAAI,EAAE,iBAAiB,CAAC,MAAM,CAAC;QAC/B,IAAI,EAAE,IAAI;QACV,QAAQ,EAAE,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC;QAC7B,MAAM,EAAE,MAAM;QACd,QAAQ,EAAE,QAAQ;QAClB,EAAE,EAAE,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC;KAC9C,CAAA;AACF,CAAC;AAED,SAAS,OAAO,CAAC,KAAa;IAC7B,OAAO,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,CAAA;AAC3E,CAAC;AA4CD,MAAM,QAAQ,GAAG,GAAe,EAAE,CAAC,CAAC;IACnC,MAAM,EAAE,CAAC;IACT,IAAI,EAAE,CAAC;IACP,KAAK,EAAE,IAAI,GAAG,EAAE;IAChB,IAAI,EAAE,IAAI,GAAG,EAAE;IACf,UAAU,EAAE,CAAC;IACb,WAAW,EAAE,CAAC;IACd,eAAe,EAAE,CAAC;IAClB,UAAU,EAAE,CAAC;IACb,UAAU,EAAE,CAAC;IACb,IAAI,EAAE,CAAC;CACP,CAAC,CAAA;AAEF,SAAS,UAAU,CAAC,KAAiB,EAAE,MAAiB,EAAE,GAAmB;IAC5E,KAAK,CAAC,MAAM,EAAE,CAAA;IAEd,IAAI,GAAG,CAAC,IAAI,EAAE,CAAC;QACd,KAAK,CAAC,IAAI,EAAE,CAAA;IACb,CAAC;IAED,MAAM,QAAQ,GAAG,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,WAAW,EAAE,CAAA;IACrD,MAAM,OAAO,GAAG,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,WAAW,EAAE,CAAA;IAEpD,IAAI,QAAQ,EAAE,CAAC;QACd,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAA;IAC1B,CAAC;IAED,IAAI,OAAO,EAAE,CAAC;QACb,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,CAAA;QAExB,KAAK,CAAC,UAAU,EAAE,CAAA;IACnB,CAAC;IAED,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAA;IAE5B,IAAI,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,EAAE,CAAC;QACxC,KAAK,CAAC,WAAW,EAAE,CAAA;IACpB,CAAC;IAED,IAAI,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,EAAE,CAAC;QAC5B,KAAK,CAAC,eAAe,EAAE,CAAA;IACxB,CAAC;IAED,IAAI,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC;QAClC,KAAK,CAAC,UAAU,EAAE,CAAA;IACnB,CAAC;IAED,IAAI,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC;QACjC,KAAK,CAAC,UAAU,EAAE,CAAA;IACnB,CAAC;IAED,IAAI,YAAY,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC;QACnC,KAAK,CAAC,IAAI,EAAE,CAAA;IACb,CAAC;AACF,CAAC;AAED,SAAS,OAAO,CAAC,MAAc,EAAE,CAAa;IAC7C,MAAM,KAAK,GAAG,CAAC,CAAS,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;IAE1D,OAAO;QACN,MAAM;QACN,MAAM,EAAE,CAAC,CAAC,MAAM;QAChB,SAAS,EAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC;QACxB,WAAW,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI;QACzB,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI;QACjB,eAAe,EAAE,KAAK,CAAC,CAAC,CAAC,UAAU,CAAC;QACpC,gBAAgB,EAAE,KAAK,CAAC,CAAC,CAAC,WAAW,CAAC;QACtC,oBAAoB,EAAE,KAAK,CAAC,CAAC,CAAC,eAAe,CAAC;QAC9C,eAAe,EAAE,KAAK,CAAC,CAAC,CAAC,UAAU,CAAC;QACpC,qBAAqB,EAAE,KAAK,CAAC,CAAC,CAAC,UAAU,CAAC;QAC1C,eAAe,EAAE,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC;KAC9B,CAAA;AACF,CAAC;AAgCD;;;;;;;;GAQG;AACH,MAAM,CAAC,KAAK,UAAU,iBAAiB,CAAC,OAAiC;IACxE,MAAM,EAAE,cAAc,EAAE,GAAG,EAAE,IAAI,EAAE,GAAG,OAAO,CAAA;IAC7C,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC,CAAA;IAElE,MAAM,OAAO,GAAG;QACf,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,eAAe,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC;QAC3D,GAAG,OAAO,CAAC,eAAe,CAAC,GAAG,CAAC,cAAc,CAAC;KAC9C,CAAA;IAED,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC9B,IAAI,MAAM,UAAU,CAAC,MAAM,CAAC;YAAE,MAAM,IAAI,KAAK,CAAC,GAAG,MAAM,wDAAwD,CAAC,CAAA;IACjH,CAAC;IAED,MAAM,MAAM,GAAG,UAAU,CAAC,IAAI,CAAC,CAAA;IAC/B,MAAM,OAAO,GAAG,IAAI,GAAG,EAAsB,CAAA;IAC7C,MAAM,UAAU,GAAG,IAAI,GAAG,EAAoD,CAAA;IAC9E,MAAM,OAAO,GAAG,IAAI,GAAG,EAA4C,CAAA;IACnE,MAAM,SAAS,GAAG,CAAC,MAAc,EAAE,MAAc,EAAE,EAAE,CAAC,GAAG,MAAM,IAAI,MAAM,CAAC,WAAW,EAAE,EAAE,CAAA;IACzF,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,eAAe,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAA;IAEjG,MAAM,KAAK,GAAG,MAAM,UAAU,CAAC,KAAK,CAAC,QAAQ,EAAE,EAAE,GAAG,EAAE,cAAc,EAAE,QAAQ,EAAE,KAAK,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAA;IAErH,IAAI,SAAS,GAAG,CAAC,CAAA;IAEjB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QAC1B,IAAI,KAAK,EAAE,MAAM,IAAI,IAAI,eAAe,CAAC,SAAS,CAAC,kBAAkB,CAAC,cAAc,EAAE,IAAI,CAAC,CAAC,EAAE,CAAC;YAC9F,IAAI,CAAC,IAAI;gBAAE,SAAQ;YACnB,MAAM,MAAM,GAAG,cAAc,CAAY,IAAI,CAAC,CAAA;YAE9C,IAAI,CAAC,MAAM;gBAAE,SAAQ;YACrB,MAAM,GAAG,GAAG,cAAc,CAAC,MAAM,CAAC,CAAA;YAElC,IAAI,CAAC,GAAG;gBAAE,SAAQ;YAElB,IAAI,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;YAEnC,IAAI,CAAC,KAAK,EAAE,CAAC;gBACZ,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC,KAAK,GAAG,QAAQ,EAAE,CAAC,CAAC,CAAA;YAC9C,CAAC;YAED,UAAU,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,CAAC,CAAA;YAE9B,IAAI,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC5B,IAAI,SAAS,GAAG,UAAU,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;gBAE1C,IAAI,CAAC,SAAS,EAAE,CAAC;oBAChB,UAAU,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC,SAAS,GAAG,EAAE,IAAI,EAAE,CAAC,EAAE,IAAI,EAAE,EAAE,EAAE,CAAC,CAAC,CAAA;gBAChE,CAAC;gBAED,SAAS,CAAC,IAAI,EAAE,CAAA;gBAEhB,IAAI,SAAS,CAAC,IAAI,CAAC,MAAM,GAAG,GAAG,EAAE,CAAC;oBACjC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;gBACzB,CAAC;qBAAM,CAAC;oBACP,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,GAAG,SAAS,CAAC,IAAI,CAAC,CAAA;oBAElD,IAAI,IAAI,GAAG,GAAG,EAAE,CAAC;wBAChB,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,GAAG,CAAA;oBAC3B,CAAC;gBACF,CAAC;YACF,CAAC;YAED,MAAM,GAAG,GAAG,SAAS,CAAC,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,QAAQ,CAAC,CAAA;YAE/C,IAAI,cAAc,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC7B,IAAI,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;gBAE5B,IAAI,CAAC,KAAK,EAAE,CAAC;oBACZ,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,KAAK,GAAG,IAAI,GAAG,EAAE,CAAC,CAAC,CAAA;gBACtC,CAAC;gBAED,IAAI,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAA;gBAEpC,IAAI,CAAC,MAAM,EAAE,CAAC;oBACb,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,QAAQ,EAAE,CAAC,MAAM,GAAG,IAAI,GAAG,EAAE,CAAC,CAAC,CAAA;gBAC9C,CAAC;gBAED,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;YACtD,CAAC;QACF,CAAC;QAED,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,SAAS,EAAE,KAAK,CAAC,MAAM,CAAC,CAAA;IAC5C,CAAC;IAED,MAAM,OAAO,GAAuB;QACnC,KAAK,EAAE,KAAK,CAAC,MAAM;QACnB,GAAG;QACH,IAAI;QACJ,MAAM,EAAE,CAAC,GAAG,OAAO,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,MAAM,CAAC;QACrG,OAAO,EAAE,EAAE;QACX,OAAO,EAAE,EAAE;KACX,CAAA;IAED,KAAK,MAAM,MAAM,IAAI,MAAM,EAAE,CAAC;QAC7B,MAAM,SAAS,GAAG,UAAU,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;QAExC,IAAI,CAAC,SAAS,EAAE,IAAI,CAAC,MAAM;YAAE,MAAM,IAAI,KAAK,CAAC,+BAA+B,MAAM,EAAE,CAAC,CAAA;QACrF,MAAM,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,eAAe,CAAC,MAAM,CAAC,CAAA;QAE5C,MAAM,IAAI,GAAG;YACZ,oBAAoB,CAAC,IAAI,CAAC,GAAG,CAAC;YAC9B,GAAG,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,oBAAoB,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACpG,EAAE;SACF,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;QAEZ,MAAM,OAAO,GAAG,IAAI,MAAM,EAAE,CAAA;QAE5B,OAAO,CAAC,OAAO,CAAC,GAAG,EAAE,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC,CAAA;QAC/C,MAAM,gBAAgB,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,GAAG,CAAC,CAAA;QAC/C,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,SAAS,CAAC,IAAI,CAAC,MAAM,EAAE,GAAG,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,GAAG,EAAE,SAAS,EAAE,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;IACjH,CAAC;IAED,KAAK,MAAM,MAAM,IAAI,OAAO,CAAC,eAAe,EAAE,CAAC;QAC9C,MAAM,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,SAAS,CAAC,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,MAAM,CAAC,CAAC,CAAA;QAElE,IAAI,CAAC,KAAK,EAAE,IAAI;YAAE,MAAM,IAAI,KAAK,CAAC,+BAA+B,MAAM,CAAC,MAAM,YAAY,MAAM,CAAC,MAAM,EAAE,CAAC,CAAA;QAC1G,MAAM,KAAK,GAAG,CAAC,eAAe,CAAC,CAAA;QAE/B,KAAK,MAAM,CAAC,GAAG,EAAE,MAAM,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YACnF,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAE,CAAA;YAEhG,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,OAAO,CAAC,IAAI,CAAC,EAAE,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAA;QACzD,CAAC;QAED,MAAM,IAAI,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAA;QACpC,MAAM,IAAI,GAAG,cAAc,CAAC,MAAM,CAAC,CAAA;QAEnC,MAAM,kBAAkB,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;QACpC,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,GAAG,MAAM,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;IACnG,CAAC;IAED,OAAO,OAAO,CAAA;AACf,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,uBAAuB,CAAC,OAA2B,EAAE,IAAqB;IACzF,OAAO,kBAAkB,CAAC,OAAO,EAAE,IAAI,CAAC,CAAA;AACzC,CAAC"}

package/out/uy/adapters/ide/adapter.d.ts

@@ -0,0 +1,64 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ * `ide-uy`: the national address points of Uruguay, `Direcciones Geográficas del Uruguay`, which the
+ * Infraestructura de Datos Espaciales (IDE) publishes on catalogodatos.gub.uy as one CSV per department.
+ *
+ * The CSVs are Windows-1252 (`ORDOÑEZ` reads `ORDO\xd1EZ`), and quote a street name that holds a
+ * nickname in doubled quotes (`"AVENIDA JOSE ""PEPE"" D' ELIA"`). Absent values read `N/A`.
+ *
+ * The adapter reads `nombre_via` as the street, `num_puerta` with `letra_puerta` as the house number,
+ * `codigo_postal` as the postcode, `localidad` as the locality, `departamento` as the region and
+ * `nombre_inmueble` as a venue. A census locality named `<DEPARTAMENTO> RURAL` names the department's rural
+ * remainder rather than a place, so such a row keeps the department and no locality. A street the register
+ * leaves unnamed carries a generated code (`UYARMQA_SN_012`) and is dropped, as is a point located only by a
+ * route kilometer, which has no door number to render.
+ *
+ * The IDE publishes the dataset under the Licencia de Datos Abiertos – Uruguay, which grants adaptation and
+ * commercial use and requires the note of origin to name the provider, the license and the dataset, and to
+ * mention changes. Every row records that license.
+ */
+import { type AdapterOptions, type CanonicalRow, type CorpusAdapter } from "#types";
+/**
+ * Registry id for this adapter, stamped into every row it emits.
+ */
+export declare const IDE_UY_ADAPTER_ID = "ide-uy";
+/**
+ * The license the IDE publishes the dataset under.
+ * It has no SPDX identifier.
+ */
+export declare const IDE_UY_LICENSE = "LicenseRef-Licencia-Datos-Abiertos-Uruguay-0.1";
+/**
+ * Every jurisdiction this adapter emits, checked against a caller's `--country`.
+ */
+export declare const IDE_UY_COUNTRIES: readonly string[];
+/**
+ * The columns the adapter reads from each department's CSV.
+ */
+export interface IDEUruguayRecord {
+    codigo_postal: string;
+    codigo_via: string;
+    nombre_via: string;
+    num_puerta: string;
+    letra_puerta: string;
+    km: string;
+    nombre_inmueble: string;
+    localidad: string;
+    departamento: string;
+    latitud: string;
+    longitud: string;
+}
+/**
+ * Maps one record onto the components the corpus asserts, or returns null when the row is refused.
+ *
+ * Each refusal increments its reason in `dropped`.
+ */
+export declare function componentsForIDERecord(record: IDEUruguayRecord, dropped?: Pick<AdapterOptions, "dropped">): CanonicalRow["components"] | null;
+export declare function createIDEUruguayAdapter(): CorpusAdapter;
+/**
+ * The configured adapter instance registered with the corpus builder.
+ */
+export declare const ideUruguayAdapter: CorpusAdapter;
+//# sourceMappingURL=adapter.d.ts.map

package/out/uy/adapters/ide/adapter.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"adapter.d.ts","sourceRoot":"","sources":["../../../../lib/uy/adapters/ide/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;GAqBG;AAeH,OAAO,EAEN,KAAK,cAAc,EACnB,KAAK,YAAY,EACjB,KAAK,aAAa,EAGlB,MAAM,QAAQ,CAAA;AAEf;;GAEG;AACH,eAAO,MAAM,iBAAiB,WAAW,CAAA;AAEzC;;;GAGG;AACH,eAAO,MAAM,cAAc,mDAAmD,CAAA;AAE9E;;GAEG;AACH,eAAO,MAAM,gBAAgB,EAAE,SAAS,MAAM,EAAW,CAAA;AAEzD;;GAEG;AACH,MAAM,WAAW,gBAAgB;IAChC,aAAa,EAAE,MAAM,CAAA;IACrB,UAAU,EAAE,MAAM,CAAA;IAClB,UAAU,EAAE,MAAM,CAAA;IAClB,UAAU,EAAE,MAAM,CAAA;IAClB,YAAY,EAAE,MAAM,CAAA;IACpB,EAAE,EAAE,MAAM,CAAA;IACV,eAAe,EAAE,MAAM,CAAA;IACvB,SAAS,EAAE,MAAM,CAAA;IACjB,YAAY,EAAE,MAAM,CAAA;IACpB,OAAO,EAAE,MAAM,CAAA;IACf,QAAQ,EAAE,MAAM,CAAA;CAChB;AA2BD;;;;GAIG;AACH,wBAAgB,sBAAsB,CACrC,MAAM,EAAE,gBAAgB,EACxB,OAAO,CAAC,EAAE,IAAI,CAAC,cAAc,EAAE,SAAS,CAAC,GACvC,YAAY,CAAC,YAAY,CAAC,GAAG,IAAI,CA6EnC;AAeD,wBAAgB,uBAAuB,IAAI,aAAa,CAkEvD;AAED;;GAEG;AACH,eAAO,MAAM,iBAAiB,eAA4B,CAAA"}

package/out/uy/adapters/ide/adapter.js

@@ -0,0 +1,206 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ * `ide-uy`: the national address points of Uruguay, `Direcciones Geográficas del Uruguay`, which the
+ * Infraestructura de Datos Espaciales (IDE) publishes on catalogodatos.gub.uy as one CSV per department.
+ *
+ * The CSVs are Windows-1252 (`ORDOÑEZ` reads `ORDO\xd1EZ`), and quote a street name that holds a
+ * nickname in doubled quotes (`"AVENIDA JOSE ""PEPE"" D' ELIA"`). Absent values read `N/A`.
+ *
+ * The adapter reads `nombre_via` as the street, `num_puerta` with `letra_puerta` as the house number,
+ * `codigo_postal` as the postcode, `localidad` as the locality, `departamento` as the region and
+ * `nombre_inmueble` as a venue. A census locality named `<DEPARTAMENTO> RURAL` names the department's rural
+ * remainder rather than a place, so such a row keeps the department and no locality. A street the register
+ * leaves unnamed carries a generated code (`UYARMQA_SN_012`) and is dropped, as is a point located only by a
+ * route kilometer, which has no door number to render.
+ *
+ * The IDE publishes the dataset under the Licencia de Datos Abiertos – Uruguay, which grants adaptation and
+ * commercial use and requires the note of origin to name the provider, the license and the dataset, and to
+ * mention changes. Every row records that license.
+ */
+import { formatAddressRow } from "@mailwoman/codex/address/format";
+import { matchCountryPostcode } from "@mailwoman/codex/postcode/country-formats";
+import { readLocalBuffer } from "@mailwoman/core/fs/readers";
+import { tryStat } from "@mailwoman/core/fs/readers/stat";
+import { decodeBytes } from "@mailwoman/core/fs/streams";
+import { resolvePath } from "path-ts";
+import { CSVSpliterator } from "spliterator";
+import { Globerator } from "spliterator/node/fs";
+import { UnsupportedCountryError } from "#adapters/errors";
+import { stableSourceIDFromParts } from "#adapters/source-id";
+import { composeHouseNumber } from "#adapters/street-line";
+import { SourceRegister } from "#registers";
+import { AddressRole, countDropped, SurfaceOrigin, } from "#types";
+/**
+ * Registry id for this adapter, stamped into every row it emits.
+ */
+export const IDE_UY_ADAPTER_ID = "ide-uy";
+/**
+ * The license the IDE publishes the dataset under.
+ * It has no SPDX identifier.
+ */
+export const IDE_UY_LICENSE = "LicenseRef-Licencia-Datos-Abiertos-Uruguay-0.1";
+/**
+ * Every jurisdiction this adapter emits, checked against a caller's `--country`.
+ */
+export const IDE_UY_COUNTRIES = ["UY"];
+const ABSENT = "N/A";
+/**
+ * A value, or null when the IDE writes it as absent.
+ */
+function present(value) {
+    const trimmed = value?.trim() ?? "";
+    return trimmed && trimmed !== ABSENT ? trimmed : null;
+}
+/**
+ * The house-number suffixes `letra_puerta` holds that a door number is written with:
+ * `BIS` and a single letter.
+ *
+ * The two-letter values (`DU`, `CC`, `TR`) are codes the dataset's metadata
+ * does not define, so they are left off.
+ */
+const LETTER_SUFFIX = /^(?:BIS|[A-Z])$/u;
+/**
+ * A street the register leaves unnamed carries a generated code, such as `UYARMQA_SN_012`.
+ */
+const GENERATED_STREET = /_SN_|^[A-Z0-9]+_[A-Z0-9_]+$/u;
+/**
+ * Maps one record onto the components the corpus asserts, or returns null when the row is refused.
+ *
+ * Each refusal increments its reason in `dropped`.
+ */
+export function componentsForIDERecord(record, dropped) {
+    const tally = dropped ?? {};
+    const street = present(record.nombre_via);
+    if (!street) {
+        countDropped(tally, "row:no-street");
+        return null;
+    }
+    if (GENERATED_STREET.test(street)) {
+        countDropped(tally, "row:street-generated-code");
+        return null;
+    }
+    const number = present(record.num_puerta);
+    if (!number) {
+        countDropped(tally, present(record.km) ? "row:route-kilometer-without-door-number" : "row:no-door-number");
+        return null;
+    }
+    const letter = present(record.letra_puerta);
+    const suffix = letter && LETTER_SUFFIX.test(letter) ? letter : "";
+    if (letter && !suffix) {
+        countDropped(tally, "component:letra_puerta:code");
+    }
+    const components = {
+        house_number: composeHouseNumber(number, suffix, " "),
+        street,
+    };
+    const venue = present(record.nombre_inmueble);
+    // A post office's distribution agent is listed under the agent's own name (`Agente` and a person's names),
+    // and a code such as `CH109` names no place, so neither becomes a venue.
+    if (venue && /^AGENTE\b/iu.test(venue)) {
+        countDropped(tally, "component:venue:names-a-person");
+    }
+    else if (venue && !/\p{L}{3}/u.test(venue)) {
+        countDropped(tally, "component:venue:code");
+    }
+    else if (venue) {
+        components.venue = venue;
+    }
+    const postcode = present(record.codigo_postal);
+    const matched = postcode ? matchCountryPostcode(postcode, "UY") : null;
+    if (matched) {
+        components.postcode = matched;
+    }
+    else if (postcode) {
+        countDropped(tally, "component:postcode:not-country-format");
+    }
+    const locality = present(record.localidad);
+    const region = present(record.departamento);
+    if (locality && !locality.endsWith(" RURAL")) {
+        components.locality = locality;
+    }
+    else if (locality) {
+        countDropped(tally, "component:locality:rural-remainder");
+    }
+    if (region && region !== components.locality) {
+        components.region = region;
+    }
+    if (!components.locality && !components.postcode) {
+        countDropped(tally, "row:no-locality-or-postcode");
+        return null;
+    }
+    return components;
+}
+/**
+ * The department CSVs an input names: the file itself, or every `.csv` in a directory, in name order.
+ */
+async function departmentFiles(inputPath) {
+    const stat = await tryStat(inputPath);
+    if (!stat?.isDirectory())
+        return [inputPath];
+    const names = await Globerator.from("*.csv", { cwd: inputPath, absolute: false }).toSorted();
+    return names.map((name) => `${inputPath}/${name}`);
+}
+export function createIDEUruguayAdapter() {
+    return {
+        id: IDE_UY_ADAPTER_ID,
+        defaultLicense: IDE_UY_LICENSE,
+        addressRole: AddressRole.Premise,
+        register: SourceRegister.UruguayIDEDirecciones,
+        surface: SurfaceOrigin.Rendered,
+        description: "Direcciones Geográficas del Uruguay (IDE Uruguay): national address points, one CSV per department.",
+        async *rows(opts) {
+            if (opts.country && !IDE_UY_COUNTRIES.includes(opts.country)) {
+                throw new UnsupportedCountryError(IDE_UY_ADAPTER_ID, IDE_UY_COUNTRIES, opts.country);
+            }
+            let emitted = 0;
+            for (const file of await departmentFiles(resolvePath(opts.inputPath))) {
+                // The largest department file is 65 MB, and the publisher fixes the partition, so each is decoded whole.
+                const text = decodeBytes(await readLocalBuffer(file), "windows-1252");
+                const records = CSVSpliterator.from(Buffer.from(text, "utf8"), { normalizeKeys: false });
+                for (const record of records) {
+                    if (opts.signal?.aborted)
+                        return;
+                    if (opts.limit !== undefined && emitted >= opts.limit)
+                        return;
+                    const components = componentsForIDERecord(record, opts);
+                    if (!components)
+                        continue;
+                    const rendered = formatAddressRow(components, "UY", { singleLine: true });
+                    if (!rendered) {
+                        countDropped(opts, "row:render-failed");
+                        continue;
+                    }
+                    for (const tag of rendered.unplaced) {
+                        countDropped(opts, `component:${tag}:not-rendered`);
+                    }
+                    const { raw, components: aligned } = rendered;
+                    yield {
+                        raw,
+                        components: aligned,
+                        country: "UY",
+                        locale: "es-UY",
+                        source: IDE_UY_ADAPTER_ID,
+                        // The point's coordinates separate two doors the rendered components cannot,
+                        // such as the several access points one `num_puerta` has on a corner lot.
+                        source_id: stableSourceIDFromParts(IDE_UY_ADAPTER_ID, {
+                            ...aligned,
+                            lat: record.latitud ?? "",
+                            lon: record.longitud ?? "",
+                        }),
+                        corpus_version: "",
+                        license: IDE_UY_LICENSE,
+                    };
+                    emitted++;
+                }
+            }
+        },
+    };
+}
+/**
+ * The configured adapter instance registered with the corpus builder.
+ */
+export const ideUruguayAdapter = createIDEUruguayAdapter();
+//# sourceMappingURL=adapter.js.map

package/out/uy/adapters/ide/adapter.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"adapter.js","sourceRoot":"","sources":["../../../../lib/uy/adapters/ide/adapter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;GAqBG;AAEH,OAAO,EAAE,gBAAgB,EAAE,MAAM,iCAAiC,CAAA;AAClE,OAAO,EAAE,oBAAoB,EAAE,MAAM,2CAA2C,CAAA;AAChF,OAAO,EAAE,eAAe,EAAE,MAAM,4BAA4B,CAAA;AAC5D,OAAO,EAAE,OAAO,EAAE,MAAM,iCAAiC,CAAA;AACzD,OAAO,EAAE,WAAW,EAAE,MAAM,4BAA4B,CAAA;AACxD,OAAO,EAAE,WAAW,EAAE,MAAM,SAAS,CAAA;AACrC,OAAO,EAAE,cAAc,EAAE,MAAM,aAAa,CAAA;AAC5C,OAAO,EAAE,UAAU,EAAE,MAAM,qBAAqB,CAAA;AAEhD,OAAO,EAAE,uBAAuB,EAAE,MAAM,kBAAkB,CAAA;AAC1D,OAAO,EAAE,uBAAuB,EAAE,MAAM,qBAAqB,CAAA;AAC7D,OAAO,EAAE,kBAAkB,EAAE,MAAM,uBAAuB,CAAA;AAC1D,OAAO,EAAE,cAAc,EAAE,MAAM,YAAY,CAAA;AAC3C,OAAO,EACN,WAAW,EAIX,YAAY,EACZ,aAAa,GACb,MAAM,QAAQ,CAAA;AAEf;;GAEG;AACH,MAAM,CAAC,MAAM,iBAAiB,GAAG,QAAQ,CAAA;AAEzC;;;GAGG;AACH,MAAM,CAAC,MAAM,cAAc,GAAG,gDAAgD,CAAA;AAE9E;;GAEG;AACH,MAAM,CAAC,MAAM,gBAAgB,GAAsB,CAAC,IAAI,CAAC,CAAA;AAmBzD,MAAM,MAAM,GAAG,KAAK,CAAA;AAEpB;;GAEG;AACH,SAAS,OAAO,CAAC,KAAyB;IACzC,MAAM,OAAO,GAAG,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,CAAA;IAEnC,OAAO,OAAO,IAAI,OAAO,KAAK,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAA;AACtD,CAAC;AAED;;;;;;GAMG;AACH,MAAM,aAAa,GAAG,kBAAkB,CAAA;AAExC;;GAEG;AACH,MAAM,gBAAgB,GAAG,8BAA8B,CAAA;AAEvD;;;;GAIG;AACH,MAAM,UAAU,sBAAsB,CACrC,MAAwB,EACxB,OAAyC;IAEzC,MAAM,KAAK,GAAG,OAAO,IAAI,EAAE,CAAA;IAC3B,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,UAAU,CAAC,CAAA;IAEzC,IAAI,CAAC,MAAM,EAAE,CAAC;QACb,YAAY,CAAC,KAAK,EAAE,eAAe,CAAC,CAAA;QAEpC,OAAO,IAAI,CAAA;IACZ,CAAC;IAED,IAAI,gBAAgB,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;QACnC,YAAY,CAAC,KAAK,EAAE,2BAA2B,CAAC,CAAA;QAEhD,OAAO,IAAI,CAAA;IACZ,CAAC;IAED,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,UAAU,CAAC,CAAA;IAEzC,IAAI,CAAC,MAAM,EAAE,CAAC;QACb,YAAY,CAAC,KAAK,EAAE,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,yCAAyC,CAAC,CAAC,CAAC,oBAAoB,CAAC,CAAA;QAE1G,OAAO,IAAI,CAAA;IACZ,CAAC;IAED,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,YAAY,CAAC,CAAA;IAC3C,MAAM,MAAM,GAAG,MAAM,IAAI,aAAa,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAA;IAEjE,IAAI,MAAM,IAAI,CAAC,MAAM,EAAE,CAAC;QACvB,YAAY,CAAC,KAAK,EAAE,6BAA6B,CAAC,CAAA;IACnD,CAAC;IAED,MAAM,UAAU,GAA+B;QAC9C,YAAY,EAAE,kBAAkB,CAAC,MAAM,EAAE,MAAM,EAAE,GAAG,CAAC;QACrD,MAAM;KACN,CAAA;IAED,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAC,eAAe,CAAC,CAAA;IAE7C,2GAA2G;IAC3G,yEAAyE;IACzE,IAAI,KAAK,IAAI,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACxC,YAAY,CAAC,KAAK,EAAE,gCAAgC,CAAC,CAAA;IACtD,CAAC;SAAM,IAAI,KAAK,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QAC9C,YAAY,CAAC,KAAK,EAAE,sBAAsB,CAAC,CAAA;IAC5C,CAAC;SAAM,IAAI,KAAK,EAAE,CAAC;QAClB,UAAU,CAAC,KAAK,GAAG,KAAK,CAAA;IACzB,CAAC;IAED,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,aAAa,CAAC,CAAA;IAC9C,MAAM,OAAO,GAAG,QAAQ,CAAC,CAAC,CAAC,oBAAoB,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAA;IAEtE,IAAI,OAAO,EAAE,CAAC;QACb,UAAU,CAAC,QAAQ,GAAG,OAAO,CAAA;IAC9B,CAAC;SAAM,IAAI,QAAQ,EAAE,CAAC;QACrB,YAAY,CAAC,KAAK,EAAE,uCAAuC,CAAC,CAAA;IAC7D,CAAC;IAED,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,SAAS,CAAC,CAAA;IAC1C,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,YAAY,CAAC,CAAA;IAE3C,IAAI,QAAQ,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC9C,UAAU,CAAC,QAAQ,GAAG,QAAQ,CAAA;IAC/B,CAAC;SAAM,IAAI,QAAQ,EAAE,CAAC;QACrB,YAAY,CAAC,KAAK,EAAE,oCAAoC,CAAC,CAAA;IAC1D,CAAC;IAED,IAAI,MAAM,IAAI,MAAM,KAAK,UAAU,CAAC,QAAQ,EAAE,CAAC;QAC9C,UAAU,CAAC,MAAM,GAAG,MAAM,CAAA;IAC3B,CAAC;IAED,IAAI,CAAC,UAAU,CAAC,QAAQ,IAAI,CAAC,UAAU,CAAC,QAAQ,EAAE,CAAC;QAClD,YAAY,CAAC,KAAK,EAAE,6BAA6B,CAAC,CAAA;QAElD,OAAO,IAAI,CAAA;IACZ,CAAC;IAED,OAAO,UAAU,CAAA;AAClB,CAAC;AAED;;GAEG;AACH,KAAK,UAAU,eAAe,CAAC,SAAiB;IAC/C,MAAM,IAAI,GAAG,MAAM,OAAO,CAAC,SAAS,CAAC,CAAA;IAErC,IAAI,CAAC,IAAI,EAAE,WAAW,EAAE;QAAE,OAAO,CAAC,SAAS,CAAC,CAAA;IAE5C,MAAM,KAAK,GAAG,MAAM,UAAU,CAAC,IAAI,CAAC,OAAO,EAAE,EAAE,GAAG,EAAE,SAAS,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAA;IAE5F,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,GAAG,SAAS,IAAI,IAAI,EAAE,CAAC,CAAA;AACnD,CAAC;AAED,MAAM,UAAU,uBAAuB;IACtC,OAAO;QACN,EAAE,EAAE,iBAAiB;QACrB,cAAc,EAAE,cAAc;QAC9B,WAAW,EAAE,WAAW,CAAC,OAAO;QAChC,QAAQ,EAAE,cAAc,CAAC,qBAAqB;QAC9C,OAAO,EAAE,aAAa,CAAC,QAAQ;QAC/B,WAAW,EAAE,qGAAqG;QAElH,KAAK,CAAC,CAAC,IAAI,CAAC,IAAoB;YAC/B,IAAI,IAAI,CAAC,OAAO,IAAI,CAAC,gBAAgB,CAAC,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC;gBAC9D,MAAM,IAAI,uBAAuB,CAAC,iBAAiB,EAAE,gBAAgB,EAAE,IAAI,CAAC,OAAO,CAAC,CAAA;YACrF,CAAC;YAED,IAAI,OAAO,GAAG,CAAC,CAAA;YAEf,KAAK,MAAM,IAAI,IAAI,MAAM,eAAe,CAAC,WAAW,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC;gBACvE,yGAAyG;gBACzG,MAAM,IAAI,GAAG,WAAW,CAAC,MAAM,eAAe,CAAC,IAAI,CAAC,EAAE,cAAc,CAAC,CAAA;gBACrE,MAAM,OAAO,GAAG,cAAc,CAAC,IAAI,CAAmB,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE,MAAM,CAAC,EAAE,EAAE,aAAa,EAAE,KAAK,EAAE,CAAC,CAAA;gBAE1G,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;oBAC9B,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,UAAU,GAAG,sBAAsB,CAAC,MAAM,EAAE,IAAI,CAAC,CAAA;oBAEvD,IAAI,CAAC,UAAU;wBAAE,SAAQ;oBAEzB,MAAM,QAAQ,GAAG,gBAAgB,CAAC,UAAU,EAAE,IAAI,EAAE,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC,CAAA;oBAEzE,IAAI,CAAC,QAAQ,EAAE,CAAC;wBACf,YAAY,CAAC,IAAI,EAAE,mBAAmB,CAAC,CAAA;wBAEvC,SAAQ;oBACT,CAAC;oBAED,KAAK,MAAM,GAAG,IAAI,QAAQ,CAAC,QAAQ,EAAE,CAAC;wBACrC,YAAY,CAAC,IAAI,EAAE,aAAa,GAAG,eAAe,CAAC,CAAA;oBACpD,CAAC;oBAED,MAAM,EAAE,GAAG,EAAE,UAAU,EAAE,OAAO,EAAE,GAAG,QAAQ,CAAA;oBAE7C,MAAM;wBACL,GAAG;wBACH,UAAU,EAAE,OAAO;wBACnB,OAAO,EAAE,IAAI;wBACb,MAAM,EAAE,OAAO;wBACf,MAAM,EAAE,iBAAiB;wBACzB,6EAA6E;wBAC7E,0EAA0E;wBAC1E,SAAS,EAAE,uBAAuB,CAAC,iBAAiB,EAAE;4BACrD,GAAG,OAAO;4BACV,GAAG,EAAE,MAAM,CAAC,OAAO,IAAI,EAAE;4BACzB,GAAG,EAAE,MAAM,CAAC,QAAQ,IAAI,EAAE;yBAC1B,CAAC;wBACF,cAAc,EAAE,EAAE;wBAClB,OAAO,EAAE,cAAc;qBACvB,CAAA;oBAED,OAAO,EAAE,CAAA;gBACV,CAAC;YACF,CAAC;QACF,CAAC;KACD,CAAA;AACF,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,iBAAiB,GAAG,uBAAuB,EAAE,CAAA"}

package/out/ve/recipes/reviewed-postcode-tail.d.ts

@@ -3,7 +3,10 @@
  * @license AGPL-3.0
  * @author Teffen Ellis, et al.
  *
- *   Renders four hand-reviewed Venezuelan postcode facts in the `locality postcode, region` order.
+ *   Renders hand-reviewed Venezuelan postcode facts in the `locality postcode, region` order.
+ *
+ *   A fact is kept in the data file only when its citation reaches a register row whose elected decision permits
+ *   transform and train. None does today, so the file holds no tuple and the recipe refuses to run.
  */
 import type { PathBuilderLike } from "path-ts";
 import { type CorpusRecipe } from "#recipes/scaffold";
@@ -47,7 +50,8 @@ export declare function defaultReviewedPostcodeTuplePath(): string;
 /**
  * Reads and validates the reviewed tuples.
  *
- * @throws When the file does not hold exactly four complete, unique, reviewed tuples.
+ * @throws When the file does not hold as many complete, unique, reviewed
+ * tuples as its `tupleCount` declares.
  */
 export declare function readReviewedPostcodeTuples(path?: PathBuilderLike): Promise<ReviewedPostcodeTuple[]>;
 /**

package/out/ve/recipes/reviewed-postcode-tail.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"reviewed-postcode-tail.d.ts","sourceRoot":"","sources":["../../../lib/ve/recipes/reviewed-postcode-tail.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAKH,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,SAAS,CAAA;AAE9C,OAAO,EAAiB,KAAK,YAAY,EAAkB,MAAM,mBAAmB,CAAA;AAKpF;;;;GAIG;AACH,eAAO,MAAM,6BAA6B,QAAsD,CAAA;AAIhG,UAAU,uBAAuB;IAChC,SAAS,EAAE,MAAM,CAAA;IACjB,GAAG,EAAE,MAAM,CAAA;IACX,WAAW,EAAE,MAAM,CAAA;IACnB,YAAY,EAAE,UAAU,CAAA;IACxB,iBAAiB,EAAE,MAAM,CAAA;CACzB;AAED;;GAEG;AACH,MAAM,WAAW,qBAAqB;IACrC,EAAE,EAAE,MAAM,CAAA;IACV,QAAQ,EAAE,MAAM,CAAA;IAChB,QAAQ,EAAE,MAAM,CAAA;IAChB,MAAM,EAAE,MAAM,CAAA;IACd,OAAO,EAAE,WAAW,CAAA;IACpB,EAAE,EAAE,IAAI,CAAA;IACR,MAAM,EAAE,OAAO,CAAA;IACf,iBAAiB,EAAE,gBAAgB,CAAA;IACnC,UAAU,EAAE,uBAAuB,CAAA;CACnC;AAQD,KAAK,SAAS,GACX,mBAAmB,GACnB,sBAAsB,GACtB,YAAY,GACZ,WAAW,GACX,cAAc,GACd,eAAe,CAAA;AAElB,UAAU,OAAO;IAChB,EAAE,EAAE,SAAS,CAAA;IACb,GAAG,EAAE,MAAM,CAAA;IACX,UAAU,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;CAClC;AAID;;GAEG;AACH,wBAAgB,gCAAgC,IAAI,MAAM,CAEzD;AAED;;;;GAIG;AACH,wBAAsB,0BAA0B,CAC/C,IAAI,GAAE,eAAoD,GACxD,OAAO,CAAC,qBAAqB,EAAE,CAAC,CAsClC;AAID;;;;;GAKG;AACH,wBAAgB,4BAA4B,CAAC,KAAK,EAAE,qBAAqB,GAAG,OAAO,EAAE,CAkCpF;AAED;;GAEG;AACH,eAAO,MAAM,0BAA0B,EAAE,YA2CxC,CAAA"}
+{"version":3,"file":"reviewed-postcode-tail.d.ts","sourceRoot":"","sources":["../../../lib/ve/recipes/reviewed-postcode-tail.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAKH,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,SAAS,CAAA;AAE9C,OAAO,EAAiB,KAAK,YAAY,EAAkB,MAAM,mBAAmB,CAAA;AAKpF;;;;GAIG;AACH,eAAO,MAAM,6BAA6B,QAAsD,CAAA;AAEhG,UAAU,uBAAuB;IAChC,SAAS,EAAE,MAAM,CAAA;IACjB,GAAG,EAAE,MAAM,CAAA;IACX,WAAW,EAAE,MAAM,CAAA;IACnB,YAAY,EAAE,UAAU,CAAA;IACxB,iBAAiB,EAAE,MAAM,CAAA;CACzB;AAED;;GAEG;AACH,MAAM,WAAW,qBAAqB;IACrC,EAAE,EAAE,MAAM,CAAA;IACV,QAAQ,EAAE,MAAM,CAAA;IAChB,QAAQ,EAAE,MAAM,CAAA;IAChB,MAAM,EAAE,MAAM,CAAA;IACd,OAAO,EAAE,WAAW,CAAA;IACpB,EAAE,EAAE,IAAI,CAAA;IACR,MAAM,EAAE,OAAO,CAAA;IACf,iBAAiB,EAAE,gBAAgB,CAAA;IACnC,UAAU,EAAE,uBAAuB,CAAA;CACnC;AAaD,KAAK,SAAS,GACX,mBAAmB,GACnB,sBAAsB,GACtB,YAAY,GACZ,WAAW,GACX,cAAc,GACd,eAAe,CAAA;AAElB,UAAU,OAAO;IAChB,EAAE,EAAE,SAAS,CAAA;IACb,GAAG,EAAE,MAAM,CAAA;IACX,UAAU,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;CAClC;AAID;;GAEG;AACH,wBAAgB,gCAAgC,IAAI,MAAM,CAEzD;AAED;;;;;GAKG;AACH,wBAAsB,0BAA0B,CAC/C,IAAI,GAAE,eAAoD,GACxD,OAAO,CAAC,qBAAqB,EAAE,CAAC,CA8ClC;AAID;;;;;GAKG;AACH,wBAAgB,4BAA4B,CAAC,KAAK,EAAE,qBAAqB,GAAG,OAAO,EAAE,CAkCpF;AAED;;GAEG;AACH,eAAO,MAAM,0BAA0B,EAAE,YAmDxC,CAAA"}

package/out/ve/recipes/reviewed-postcode-tail.js

@@ -3,7 +3,10 @@
  * @license AGPL-3.0
  * @author Teffen Ellis, et al.
  *
- *   Renders four hand-reviewed Venezuelan postcode facts in the `locality postcode, region` order.
+ *   Renders hand-reviewed Venezuelan postcode facts in the `locality postcode, region` order.
+ *
+ *   A fact is kept in the data file only when its citation reaches a register row whose elected decision permits
+ *   transform and train. None does today, so the file holds no tuple and the recipe refuses to run.
  */
 import { readLocalJSONFile } from "@mailwoman/core/fs/readers";
 import { resolveModulePath } from "@mailwoman/core/module/resolvers";
@@ -18,7 +21,6 @@ import { SurfaceOrigin } from "#types";
  * The rows get their own sampler bucket so an eval can measure them apart from other postcode placements.
  */
 export const REVIEWED_POSTCODE_TAIL_SOURCE = defaultRecipeSource("synth-reviewed-postcode-tail");
-const REVIEWED_TUPLE_COUNT = 4;
 const ISO_DATE = /^\d{4}-\d{2}-\d{2}$/;
 /**
  * Returns the path of the committed reviewed-tuple data file.
@@ -29,12 +31,17 @@ export function defaultReviewedPostcodeTuplePath() {
 /**
  * Reads and validates the reviewed tuples.
  *
- * @throws When the file does not hold exactly four complete, unique, reviewed tuples.
+ * @throws When the file does not hold as many complete, unique, reviewed
+ * tuples as its `tupleCount` declares.
  */
 export async function readReviewedPostcodeTuples(path = defaultReviewedPostcodeTuplePath()) {
     const document = await readLocalJSONFile(path);
-    if (!document.version || !ISO_DATE.test(document.reviewedAt) || document.tuples.length !== REVIEWED_TUPLE_COUNT) {
-        throw new Error("reviewed postcode tuple file must declare a version, review date, and exactly four tuples");
+    if (!document.version ||
+        !ISO_DATE.test(document.reviewedAt) ||
+        !Number.isInteger(document.tupleCount) ||
+        document.tuples.length !== document.tupleCount) {
+        throw new Error(`reviewed postcode tuple file must declare a version, a review date, and a tupleCount matching its ` +
+            `${document.tuples?.length ?? 0} tuples`);
     }
     const ids = new Set();
     const facts = new Set();
@@ -110,6 +117,10 @@ export const reviewedPostcodeTailRecipe = {
     mode: "generate",
     async run(opts, write) {
         const tuples = await readReviewedPostcodeTuples(opts.input);
+        if (!tuples.length) {
+            throw new Error("reviewed-postcode-tail has no tuple to render: no reviewed fact's citation reaches a register row " +
+                "permitting transform and train");
+        }
         let emitted = 0;
         let skipped = 0;
         for (const tuple of tuples) {