@mailwoman/corpus 9.2.0 → 9.4.0

package/out/src/tools/corpus-stats.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"corpus-stats.js","sourceRoot":"","sources":["../../../src/tools/corpus-stats.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAoCG;AAEH,OAAO,EAAE,WAAW,EAAE,QAAQ,EAAE,aAAa,EAAE,MAAM,SAAS,CAAA;AAC9D,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAA;AAEhC,OAAO,EAAE,aAAa,EAAE,MAAM,6BAA6B,CAAA;AAE3D,MAAM,GAAG,GAAG,GAAG,CAAA;AACf,MAAM,gBAAgB,GAAG,CAAC,CAAA;AAQ1B,SAAS,cAAc,CAAC,SAAiB;IACxC,MAAM,IAAI,GAAG,QAAQ,CAAC,SAAS,CAAC,CAAA;IAEhC,IAAI,IAAI,CAAC,WAAW,EAAE,EAAE,CAAC;QACxB,OAAO,WAAW,CAAC,SAAS,CAAC;aAC3B,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC;aACrC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,CAAA;IACjC,CAAC;IAED,IAAI,IAAI,CAAC,MAAM,EAAE,IAAI,SAAS,CAAC,QAAQ,CAAC,UAAU,CAAC;QAAE,OAAO,CAAC,SAAS,CAAC,CAAA;IAEvE,gFAAgF;IAChF,OAAO,CAAC,SAAS,CAAC,CAAA;AACnB,CAAC;AAED;;;;GAIG;AACH,KAAK,SAAS,CAAC,CAAC,eAAe,CAC9B,SAAiB,EACjB,KAAc;IAEd,YAAY,MAAM,GAAG,MAAM,aAAa,CAAC,QAAQ,CAAyC,SAAS,CAAC,CAAA;IAEpG,IAAI,OAAO,GAAG,CAAC,CAAA;IAEf,IAAI,KAAK,EAAE,MAAM,GAAG,IAAI,MAAM,CAAC,OAAO,CAAC,QAAQ,EAAE,QAAQ,CAAC,EAAE,CAAC;QAC5D,IAAI,KAAK,KAAK,SAAS,IAAI,OAAO,IAAI,KAAK;YAAE,MAAK;QAElD,MAAM,GAAG,CAAA;QAET,OAAO,EAAE,CAAA;IACV,CAAC;AACF,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,gBAAgB,CAAC,IAAwB;IAC9D,MAAM,UAAU,GAAG,cAAc,CAAC,IAAI,CAAC,SAAS,CAAC,CAAA;IAEjD,OAAO,CAAC,KAAK,CAAC,cAAc,UAAU,CAAC,MAAM,mBAAmB,CAAC,CAAA;IAEjE,MAAM,UAAU,GAAG,IAAI,GAAG,EAA+B,CAAA;IACzD,MAAM,WAAW,GAAG,IAAI,GAAG,EAA+B,CAAA;IAC1D,IAAI,SAAS,GAAG,CAAC,CAAA;IAEjB,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;QAC/B,OAAO,CAAC,KAAK,CAAC,WAAW,IAAI,KAAK,CAAC,CAAA;QAEnC,MAAM,MAAM,GAAG,SAAS,CAAA;QAExB,IAAI,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,IAAI,eAAe,CAAC,IAAI,EAAE,IAAI,CAAC,aAAa,CAAC,EAAE,CAAC;YAClF,SAAS,EAAE,CAAA;YAEX,IAAI,MAAM,CAAC,MAAM,KAAK,MAAM,CAAC,MAAM;gBAAE,SAAQ;YAE7C,iBAAiB;YACjB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACxC,MAAM,EAAE,GAAG,MAAM,CAAC,CAAC,CAAE,CAAA;gBACrB,MAAM,EAAE,GAAG,MAAM,CAAC,CAAC,CAAE,CAAA;gBACrB,IAAI,QAAQ,GAAG,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC,CAAA;gBAEjC,IAAI,CAAC,QAAQ,EAAE,CAAC;oBACf,QAAQ,GAAG,IAAI,GAAG,EAAE,CAAA;oBACpB,UAAU,CAAC,GAAG,CAAC,EAAE,EAAE,QAAQ,CAAC,CAAA;gBAC7B,CAAC;gBAED,QAAQ,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;gBAE7C,IAAI,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC;oBAC3B,MAAM,SAAS,GAAG,EAAE,GAAG,GAAG,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAE,CAAA;oBAC3C,MAAM,WAAW,GAAG,EAAE,GAAG,GAAG,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAE,CAAA;oBAC7C,IAAI,IAAI,GAAG,WAAW,CAAC,GAAG,CAAC,SAAS,CAAC,CAAA;oBAErC,IAAI,CAAC,IAAI,EAAE,CAAC;wBACX,IAAI,GAAG,IAAI,GAAG,EAAE,CAAA;wBAChB,WAAW,CAAC,GAAG,CAAC,SAAS,EAAE,IAAI,CAAC,CAAA;oBACjC,CAAC;oBAED,IAAI,CAAC,GAAG,CAAC,WAAW,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;gBACxD,CAAC;YACF,CAAC;QACF,CAAC;QAED,OAAO,CAAC,KAAK,CACZ,KAAK,SAAS,GAAG,MAAM,0BAA0B,UAAU,CAAC,IAAI,mBAAmB,WAAW,CAAC,IAAI,iBAAiB,CACpH,CAAA;IACF,CAAC;IAED,sFAAsF;IACtF,uFAAuF;IACvF,aAAa;IACb,IAAI,aAAa,GAAG,CAAC,CAAA;IAErB,KAAK,MAAM,CAAC,CAAC,EAAE,QAAQ,CAAC,IAAI,WAAW,EAAE,CAAC;QACzC,IAAI,KAAK,GAAG,CAAC,CAAA;QAEb,KAAK,MAAM,CAAC,IAAI,QAAQ,CAAC,MAAM,EAAE,EAAE,CAAC;YACnC,KAAK,IAAI,CAAC,CAAA;QACX,CAAC;QAED,IAAI,KAAK,GAAG,gBAAgB,EAAE,CAAC;YAC9B,WAAW,CAAC,MAAM,CAAC,CAAC,CAAC,CAAA;YAErB,aAAa,EAAE,CAAA;QAChB,CAAC;IACF,CAAC;IAED,OAAO,CAAC,KAAK,CAAC,UAAU,aAAa,uBAAuB,WAAW,CAAC,IAAI,SAAS,CAAC,CAAA;IAEtF,MAAM,GAAG,GAAG;QACX,SAAS,EAAE,SAAS;QACpB,WAAW,EAAE,UAAU;QACvB,MAAM,EAAE,EAA4C;QACpD,OAAO,EAAE,EAA4C;KACrD,CAAA;IAED,KAAK,MAAM,CAAC,EAAE,EAAE,QAAQ,CAAC,IAAI,UAAU,EAAE,CAAC;QACzC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,WAAW,CAAC,QAAQ,CAAC,CAAA;IAC9C,CAAC;IAED,KAAK,MAAM,CAAC,CAAC,EAAE,QAAQ,CAAC,IAAI,WAAW,EAAE,CAAC;QACzC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,MAAM,CAAC,WAAW,CAAC,QAAQ,CAAC,CAAA;IAC9C,CAAC;IAED,aAAa,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAA;IACnD,MAAM,MAAM,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,GAAG,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAA;IAEhF,OAAO,CAAC,KAAK,CACZ,SAAS,IAAI,CAAC,UAAU,KAAK,MAAM,UAAU,SAAS,UAAU,UAAU,CAAC,IAAI,YAAY,WAAW,CAAC,IAAI,UAAU,CACrH,CAAA;AACF,CAAC"}

package/out/src/tools/fetch/ban.d.ts

@@ -1,31 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   Fetch the full French BAN (Base Adresse Nationale) — all metropolitan départements (01-95, 2A,
- *   2B) plus 5 overseas DOM/TOM (971-976 excl. 975).
- *
- *   Source: https://adresse.data.gouv.fr/data/ban/adresses/latest/csv/
- *   Licence: Licence Ouverte 2.0 (attribution required — Tier B).
- *
- *   Files already present with matching sha256 are skipped (re-runnable). Downloads `.csv.gz`,
- *   decompresses to `.csv`, deletes the `.gz` artifact. One shared `MANIFEST.json` at
- *   `<outRoot>/ban/MANIFEST.json` covers all codes.
- *
- *   Invoke via `mailwoman corpus fetch ban --out-root <path>`. Built-in `fetch` with gzip/brotli
- *   decompression replaces curl; native `node:zlib` gunzip replaces the `gunzip` subprocess; no
- *   Python.
- */
-import type { BaseFetchOptions, FetchSummary } from "./download.ts";
-export interface BanManifestEntry {
-    dept_code: string;
-    filename: string;
-    source_url: string;
-    downloaded_at: string;
-    sha256: string;
-    bytes: number;
-}
-export type FetchBanOptions = BaseFetchOptions;
-export declare function fetchBan(options: FetchBanOptions, report?: (line: string) => void): Promise<FetchSummary>;
-//# sourceMappingURL=ban.d.ts.map

package/out/src/tools/fetch/ban.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"ban.d.ts","sourceRoot":"","sources":["../../../../src/tools/fetch/ban.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAWH,OAAO,KAAK,EAAE,gBAAgB,EAAE,YAAY,EAAE,MAAM,eAAe,CAAA;AAsHnE,MAAM,WAAW,gBAAgB;IAChC,SAAS,EAAE,MAAM,CAAA;IACjB,QAAQ,EAAE,MAAM,CAAA;IAChB,UAAU,EAAE,MAAM,CAAA;IAClB,aAAa,EAAE,MAAM,CAAA;IACrB,MAAM,EAAE,MAAM,CAAA;IACd,KAAK,EAAE,MAAM,CAAA;CACb;AAED,MAAM,MAAM,eAAe,GAAG,gBAAgB,CAAA;AAE9C,wBAAsB,QAAQ,CAAC,OAAO,EAAE,eAAe,EAAE,MAAM,CAAC,EAAE,CAAC,IAAI,EAAE,MAAM,KAAK,IAAI,GAAG,OAAO,CAAC,YAAY,CAAC,CAmI/G"}

package/out/src/tools/fetch/ban.js

@@ -1,239 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   Fetch the full French BAN (Base Adresse Nationale) — all metropolitan départements (01-95, 2A,
- *   2B) plus 5 overseas DOM/TOM (971-976 excl. 975).
- *
- *   Source: https://adresse.data.gouv.fr/data/ban/adresses/latest/csv/
- *   Licence: Licence Ouverte 2.0 (attribution required — Tier B).
- *
- *   Files already present with matching sha256 are skipped (re-runnable). Downloads `.csv.gz`,
- *   decompresses to `.csv`, deletes the `.gz` artifact. One shared `MANIFEST.json` at
- *   `<outRoot>/ban/MANIFEST.json` covers all codes.
- *
- *   Invoke via `mailwoman corpus fetch ban --out-root <path>`. Built-in `fetch` with gzip/brotli
- *   decompression replaces curl; native `node:zlib` gunzip replaces the `gunzip` subprocess; no
- *   Python.
- */
-import { existsSync, mkdirSync, statSync } from "node:fs";
-import { readFile, unlink, writeFile } from "node:fs/promises";
-import { join } from "node:path";
-import { setTimeout as sleep } from "node:timers/promises";
-import { gunzipSync } from "node:zlib";
-import { BYTES_PER_KIB, ByteFormatter } from "@mailwoman/core/fs/utils";
-import { sha256File } from "@mailwoman/core/utils";
-import { downloadToFile, loadManifestEntries, writeManifest } from "./download.js";
-/**
- * Bytes per KiB — the divisor for human-readable sizes, and the floor below which a "download" is an error page rather
- * than data.
- */
-const BASE_URL = "https://adresse.data.gouv.fr/data/ban/adresses/latest/csv";
-/**
- * All département codes — metropolitan 01-95 (with 2A/2B for Corsica instead of 20) plus overseas DOM/TOM. Codes do not
- * change.
- */
-const DEPT_CODES = [
-    "01",
-    "02",
-    "03",
-    "04",
-    "05",
-    "06",
-    "07",
-    "08",
-    "09",
-    "10",
-    "11",
-    "12",
-    "13",
-    "14",
-    "15",
-    "16",
-    "17",
-    "18",
-    "19",
-    "21",
-    "22",
-    "23",
-    "24",
-    "25",
-    "26",
-    "27",
-    "28",
-    "29",
-    "2A",
-    "2B",
-    "30",
-    "31",
-    "32",
-    "33",
-    "34",
-    "35",
-    "36",
-    "37",
-    "38",
-    "39",
-    "40",
-    "41",
-    "42",
-    "43",
-    "44",
-    "45",
-    "46",
-    "47",
-    "48",
-    "49",
-    "50",
-    "51",
-    "52",
-    "53",
-    "54",
-    "55",
-    "56",
-    "57",
-    "58",
-    "59",
-    "60",
-    "61",
-    "62",
-    "63",
-    "64",
-    "65",
-    "66",
-    "67",
-    "68",
-    "69",
-    "70",
-    "71",
-    "72",
-    "73",
-    "74",
-    "75",
-    "76",
-    "77",
-    "78",
-    "79",
-    "80",
-    "81",
-    "82",
-    "83",
-    "84",
-    "85",
-    "86",
-    "87",
-    "88",
-    "89",
-    "90",
-    "91",
-    "92",
-    "93",
-    "94",
-    "95",
-    "971",
-    "972",
-    "973",
-    "974",
-    "976",
-];
-export async function fetchBan(options, report) {
-    const banDir = join(options.outRoot, "ban");
-    const manifestPath = join(banDir, "MANIFEST.json");
-    mkdirSync(banDir, { recursive: true });
-    // Load existing entries (code -> entry): skip detection + preservation of untouched codes.
-    const entries = await loadManifestEntries(manifestPath, (entry) => entry.dept_code);
-    let fetched = 0;
-    let skipped = 0;
-    let failed = 0;
-    const failedCodes = [];
-    for (const code of DEPT_CODES) {
-        const filename = `adresses-${code}.csv`;
-        const gzFile = join(banDir, `${filename}.gz`);
-        const csvFile = join(banDir, filename);
-        const url = `${BASE_URL}/adresses-${code}.csv.gz`;
-        report?.(`=== dept ${code}`);
-        // If the CSV already exists, compare its sha256 against the manifest.
-        if (existsSync(csvFile)) {
-            const existingSha = await sha256File(csvFile);
-            const recordedSha = entries.get(code)?.sha256;
-            if (recordedSha && existingSha === recordedSha) {
-                report?.(`  → already present + sha matches — skipping`);
-                skipped++;
-                continue;
-            }
-            report?.(`  → present but sha mismatch or no manifest entry — re-fetching`);
-            await unlink(csvFile);
-        }
-        // Download the gzipped CSV.
-        try {
-            await downloadToFile({
-                url,
-                dest: gzFile,
-                timeoutMs: 600_000,
-                headers: { "Accept-Encoding": "gzip, br" },
-                report,
-            });
-        }
-        catch (error) {
-            report?.(`  ✗ download failed: ${url} (${error.message})`);
-            failed++;
-            failedCodes.push(code);
-            continue;
-        }
-        // Guard against truncated 404/error pages.
-        const gzSize = statSync(gzFile).size;
-        if (gzSize < BYTES_PER_KIB) {
-            report?.(`  ✗ response too small (${gzSize} bytes) — probable 404 / error page`);
-            await unlink(gzFile);
-            failed++;
-            failedCodes.push(code);
-            continue;
-        }
-        // Decompress in-place; delete the .gz.
-        try {
-            await writeFile(csvFile, gunzipSync(await readFile(gzFile)));
-        }
-        catch (error) {
-            report?.(`  ✗ decompress failed: ${error.message}`);
-            await unlink(gzFile);
-            failed++;
-            failedCodes.push(code);
-            continue;
-        }
-        await unlink(gzFile);
-        if (!existsSync(csvFile)) {
-            report?.(`  ✗ decompressed file not found at ${csvFile}`);
-            failed++;
-            failedCodes.push(code);
-            continue;
-        }
-        const bytes = statSync(csvFile).size;
-        const sha = await sha256File(csvFile);
-        entries.set(code, {
-            dept_code: code,
-            filename,
-            source_url: url,
-            downloaded_at: new Date().toISOString(),
-            sha256: sha,
-            bytes,
-        });
-        report?.(`  ✓ ${ByteFormatter.formatIEC(bytes)}  sha256=${sha}`);
-        fetched++;
-        // Be a polite citizen — short pause between requests.
-        await sleep(200);
-    }
-    // Write the consolidated MANIFEST.json (entries sorted by dept_code, codepoint order).
-    const sorted = [...entries.values()].toSorted((a, b) => a.dept_code < b.dept_code ? -1 : a.dept_code > b.dept_code ? 1 : 0);
-    await writeManifest(manifestPath, sorted);
-    report?.(`Wrote ${manifestPath} with ${sorted.length} entries.`);
-    report?.(`=== summary ===`);
-    report?.(`fetched:  ${fetched}`);
-    report?.(`skipped:  ${skipped} (already present + sha matched)`);
-    report?.(`failed:   ${failed}`);
-    if (failedCodes.length) {
-        report?.(`failed codes: ${failedCodes.join(" ")}`);
-    }
-    return { fetched, skipped, failed, failedCodes };
-}
-//# sourceMappingURL=ban.js.map

package/out/src/tools/fetch/ban.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"ban.js","sourceRoot":"","sources":["../../../../src/tools/fetch/ban.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAEH,OAAO,EAAE,UAAU,EAAE,SAAS,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAA;AACzD,OAAO,EAAE,QAAQ,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAA;AAC9D,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAA;AAChC,OAAO,EAAE,UAAU,IAAI,KAAK,EAAE,MAAM,sBAAsB,CAAA;AAC1D,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAA;AAEtC,OAAO,EAAE,aAAa,EAAE,aAAa,EAAE,MAAM,0BAA0B,CAAA;AACvE,OAAO,EAAE,UAAU,EAAE,MAAM,uBAAuB,CAAA;AAGlD,OAAO,EAAE,cAAc,EAAE,mBAAmB,EAAE,aAAa,EAAE,MAAM,eAAe,CAAA;AAElF;;;GAGG;AAEH,MAAM,QAAQ,GAAG,2DAA2D,CAAA;AAE5E;;;GAGG;AACH,MAAM,UAAU,GAAG;IAClB,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,KAAK;IACL,KAAK;IACL,KAAK;IACL,KAAK;IACL,KAAK;CACL,CAAA;AAaD,MAAM,CAAC,KAAK,UAAU,QAAQ,CAAC,OAAwB,EAAE,MAA+B;IACvF,MAAM,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,OAAO,EAAE,KAAK,CAAC,CAAA;IAC3C,MAAM,YAAY,GAAG,IAAI,CAAC,MAAM,EAAE,eAAe,CAAC,CAAA;IAClD,SAAS,CAAC,MAAM,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAA;IAEtC,2FAA2F;IAC3F,MAAM,OAAO,GAAG,MAAM,mBAAmB,CAAmB,YAAY,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,SAAS,CAAC,CAAA;IAErG,IAAI,OAAO,GAAG,CAAC,CAAA;IACf,IAAI,OAAO,GAAG,CAAC,CAAA;IACf,IAAI,MAAM,GAAG,CAAC,CAAA;IACd,MAAM,WAAW,GAAa,EAAE,CAAA;IAEhC,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;QAC/B,MAAM,QAAQ,GAAG,YAAY,IAAI,MAAM,CAAA;QACvC,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,EAAE,GAAG,QAAQ,KAAK,CAAC,CAAA;QAC7C,MAAM,OAAO,GAAG,IAAI,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAA;QACtC,MAAM,GAAG,GAAG,GAAG,QAAQ,aAAa,IAAI,SAAS,CAAA;QAEjD,MAAM,EAAE,CAAC,YAAY,IAAI,EAAE,CAAC,CAAA;QAE5B,sEAAsE;QACtE,IAAI,UAAU,CAAC,OAAO,CAAC,EAAE,CAAC;YACzB,MAAM,WAAW,GAAG,MAAM,UAAU,CAAC,OAAO,CAAC,CAAA;YAC7C,MAAM,WAAW,GAAG,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,MAAM,CAAA;YAE7C,IAAI,WAAW,IAAI,WAAW,KAAK,WAAW,EAAE,CAAC;gBAChD,MAAM,EAAE,CAAC,8CAA8C,CAAC,CAAA;gBAExD,OAAO,EAAE,CAAA;gBAET,SAAQ;YACT,CAAC;YAED,MAAM,EAAE,CAAC,iEAAiE,CAAC,CAAA;YAC3E,MAAM,MAAM,CAAC,OAAO,CAAC,CAAA;QACtB,CAAC;QAED,4BAA4B;QAC5B,IAAI,CAAC;YACJ,MAAM,cAAc,CAAC;gBACpB,GAAG;gBACH,IAAI,EAAE,MAAM;gBACZ,SAAS,EAAE,OAAO;gBAClB,OAAO,EAAE,EAAE,iBAAiB,EAAE,UAAU,EAAE;gBAC1C,MAAM;aACN,CAAC,CAAA;QACH,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YAChB,MAAM,EAAE,CAAC,wBAAwB,GAAG,KAAM,KAAe,CAAC,OAAO,GAAG,CAAC,CAAA;YAErE,MAAM,EAAE,CAAA;YACR,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YAEtB,SAAQ;QACT,CAAC;QAED,2CAA2C;QAC3C,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,IAAI,CAAA;QAEpC,IAAI,MAAM,GAAG,aAAa,EAAE,CAAC;YAC5B,MAAM,EAAE,CAAC,2BAA2B,MAAM,qCAAqC,CAAC,CAAA;YAChF,MAAM,MAAM,CAAC,MAAM,CAAC,CAAA;YAEpB,MAAM,EAAE,CAAA;YACR,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YAEtB,SAAQ;QACT,CAAC;QAED,uCAAuC;QACvC,IAAI,CAAC;YACJ,MAAM,SAAS,CAAC,OAAO,EAAE,UAAU,CAAC,MAAM,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,CAAA;QAC7D,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YAChB,MAAM,EAAE,CAAC,0BAA2B,KAAe,CAAC,OAAO,EAAE,CAAC,CAAA;YAC9D,MAAM,MAAM,CAAC,MAAM,CAAC,CAAA;YAEpB,MAAM,EAAE,CAAA;YACR,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YAEtB,SAAQ;QACT,CAAC;QAED,MAAM,MAAM,CAAC,MAAM,CAAC,CAAA;QAEpB,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,EAAE,CAAC;YAC1B,MAAM,EAAE,CAAC,sCAAsC,OAAO,EAAE,CAAC,CAAA;YAEzD,MAAM,EAAE,CAAA;YACR,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YAEtB,SAAQ;QACT,CAAC;QAED,MAAM,KAAK,GAAG,QAAQ,CAAC,OAAO,CAAC,CAAC,IAAI,CAAA;QACpC,MAAM,GAAG,GAAG,MAAM,UAAU,CAAC,OAAO,CAAC,CAAA;QAErC,OAAO,CAAC,GAAG,CAAC,IAAI,EAAE;YACjB,SAAS,EAAE,IAAI;YACf,QAAQ;YACR,UAAU,EAAE,GAAG;YACf,aAAa,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;YACvC,MAAM,EAAE,GAAG;YACX,KAAK;SACL,CAAC,CAAA;QAEF,MAAM,EAAE,CAAC,OAAO,aAAa,CAAC,SAAS,CAAC,KAAK,CAAC,YAAY,GAAG,EAAE,CAAC,CAAA;QAEhE,OAAO,EAAE,CAAA;QAET,sDAAsD;QACtD,MAAM,KAAK,CAAC,GAAG,CAAC,CAAA;IACjB,CAAC;IAED,uFAAuF;IACvF,MAAM,MAAM,GAAG,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CACtD,CAAC,CAAC,SAAS,GAAG,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,GAAG,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAClE,CAAA;IAED,MAAM,aAAa,CAAC,YAAY,EAAE,MAAM,CAAC,CAAA;IACzC,MAAM,EAAE,CAAC,SAAS,YAAY,SAAS,MAAM,CAAC,MAAM,WAAW,CAAC,CAAA;IAEhE,MAAM,EAAE,CAAC,iBAAiB,CAAC,CAAA;IAC3B,MAAM,EAAE,CAAC,aAAa,OAAO,EAAE,CAAC,CAAA;IAChC,MAAM,EAAE,CAAC,aAAa,OAAO,kCAAkC,CAAC,CAAA;IAChE,MAAM,EAAE,CAAC,aAAa,MAAM,EAAE,CAAC,CAAA;IAE/B,IAAI,WAAW,CAAC,MAAM,EAAE,CAAC;QACxB,MAAM,EAAE,CAAC,iBAAiB,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAA;IACnD,CAAC;IAED,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,WAAW,EAAE,CAAA;AACjD,CAAC"}

package/out/src/tools/fetch/download.d.ts

@@ -1,78 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   Download and manifest utilities for `mailwoman corpus fetch <source>`.
- */
-/**
- * The option base every `mailwoman corpus fetch <source>` module extends.
- */
-export interface BaseFetchOptions {
-    /**
-     * Destination root for downloaded source data. Each source writes its own subdirectory.
-     */
-    outRoot: string;
-}
-/**
- * The per-run result every fetch module returns; the command maps `failed > 0` to exit code 1.
- */
-export interface FetchSummary {
-    fetched: number;
-    skipped: number;
-    failed: number;
-    failedCodes: string[];
-}
-/**
- * A status worth retrying: rate limiting or a server-side failure.
- */
-/**
- * An HTTP failure that CARRIES its status, so callers branch on `error.status` rather than on message prose. The prose
- * route shipped a real flake: a caller classified "not published upstream" with `message.includes("404")`, and the
- * message contains the URL — an ephemeral test-server port such as `:40453` satisfies it while the actual status is
- * 500. Roughly 1–2% of ephemeral ports contain the substring, which is exactly the kind of sometimes-failure that burns
- * a CI run and vanishes locally.
- */
-export declare class HTTPStatusError extends Error {
-    readonly status: number;
-    constructor(status: number, message: string);
-}
-export declare function isTransientStatus(status: number): boolean;
-export interface DownloadOptions {
-    url: string;
-    dest: string;
-    /**
-     * Per-attempt timeout. Default 10 minutes — these are multi-GB government dumps.
-     */
-    timeoutMs?: number;
-    /**
-     * Extra attempts after the first, taken only on transient statuses or network errors. Default 0.
-     */
-    retries?: number;
-    /**
-     * Delay between attempts. Default 5s.
-     */
-    retryDelayMs?: number;
-    headers?: Record<string, string>;
-    report?: (line: string) => void;
-}
-/**
- * Download `url` to `dest` with per-attempt timeout and transient-status retry. Throws on a non-transient HTTP status
- * or once retries are exhausted. Returns the byte count written.
- */
-export declare function downloadToFile(options: DownloadOptions): Promise<{
-    bytes: number;
-}>;
-/**
- * Read a MANIFEST.json; `null` when missing or corrupt (callers re-fetch from scratch).
- */
-export declare function readManifest<T>(path: string): Promise<T | null>;
-/**
- * Load manifest entries into a map so untouched keys survive a partial re-fetch.
- */
-export declare function loadManifestEntries<T>(path: string, key: (entry: T) => string): Promise<Map<string, T>>;
-/**
- * Write a MANIFEST.json in the house shape: pretty-printed, trailing newline.
- */
-export declare function writeManifest(path: string, manifest: unknown): Promise<void>;
-//# sourceMappingURL=download.d.ts.map

package/out/src/tools/fetch/download.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"download.d.ts","sourceRoot":"","sources":["../../../../src/tools/fetch/download.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAuBH;;GAEG;AACH,MAAM,WAAW,gBAAgB;IAChC;;OAEG;IACH,OAAO,EAAE,MAAM,CAAA;CACf;AAED;;GAEG;AACH,MAAM,WAAW,YAAY;IAC5B,OAAO,EAAE,MAAM,CAAA;IACf,OAAO,EAAE,MAAM,CAAA;IACf,MAAM,EAAE,MAAM,CAAA;IACd,WAAW,EAAE,MAAM,EAAE,CAAA;CACrB;AAED;;GAEG;AACH;;;;;;GAMG;AACH,qBAAa,eAAgB,SAAQ,KAAK;IACzC,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAA;gBAEX,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM;CAK3C;AAED,wBAAgB,iBAAiB,CAAC,MAAM,EAAE,MAAM,GAAG,OAAO,CAEzD;AAED,MAAM,WAAW,eAAe;IAC/B,GAAG,EAAE,MAAM,CAAA;IACX,IAAI,EAAE,MAAM,CAAA;IACZ;;OAEG;IACH,SAAS,CAAC,EAAE,MAAM,CAAA;IAClB;;OAEG;IACH,OAAO,CAAC,EAAE,MAAM,CAAA;IAChB;;OAEG;IACH,YAAY,CAAC,EAAE,MAAM,CAAA;IACrB,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;IAChC,MAAM,CAAC,EAAE,CAAC,IAAI,EAAE,MAAM,KAAK,IAAI,CAAA;CAC/B;AAED;;;GAGG;AACH,wBAAsB,cAAc,CAAC,OAAO,EAAE,eAAe,GAAG,OAAO,CAAC;IAAE,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC,CA6CzF;AAED;;GAEG;AACH,wBAAsB,YAAY,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,CAAC,GAAG,IAAI,CAAC,CAQrE;AAED;;GAEG;AACH,wBAAsB,mBAAmB,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC,KAAK,EAAE,CAAC,KAAK,MAAM,GAAG,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,CAS7G;AAED;;GAEG;AACH,wBAAsB,aAAa,CAAC,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,OAAO,GAAG,OAAO,CAAC,IAAI,CAAC,CAElF"}

package/out/src/tools/fetch/download.js

@@ -1,116 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   Download and manifest utilities for `mailwoman corpus fetch <source>`.
- */
-import { existsSync } from "node:fs";
-import { readFile, writeFile } from "node:fs/promises";
-import { setTimeout as sleep } from "node:timers/promises";
-import { tryParsingJSON } from "@mailwoman/core/objects";
-/**
- * Rate limited — retryable, the server is asking us to back off.
- */
-const HTTP_TOO_MANY_REQUESTS = 429;
-/**
- * Lowest 5xx status. Server-side failures are retryable; 4xx are not.
- */
-const HTTP_SERVER_ERROR_MIN = 500;
-/**
- * Highest 5xx status.
- */
-const HTTP_SERVER_ERROR_MAX = 599;
-/**
- * A status worth retrying: rate limiting or a server-side failure.
- */
-/**
- * An HTTP failure that CARRIES its status, so callers branch on `error.status` rather than on message prose. The prose
- * route shipped a real flake: a caller classified "not published upstream" with `message.includes("404")`, and the
- * message contains the URL — an ephemeral test-server port such as `:40453` satisfies it while the actual status is
- * 500. Roughly 1–2% of ephemeral ports contain the substring, which is exactly the kind of sometimes-failure that burns
- * a CI run and vanishes locally.
- */
-export class HTTPStatusError extends Error {
-    status;
-    constructor(status, message) {
-        super(message);
-        this.name = "HTTPStatusError";
-        this.status = status;
-    }
-}
-export function isTransientStatus(status) {
-    return status === HTTP_TOO_MANY_REQUESTS || (status >= HTTP_SERVER_ERROR_MIN && status <= HTTP_SERVER_ERROR_MAX);
-}
-/**
- * Download `url` to `dest` with per-attempt timeout and transient-status retry. Throws on a non-transient HTTP status
- * or once retries are exhausted. Returns the byte count written.
- */
-export async function downloadToFile(options) {
-    const { url, dest, timeoutMs = 600_000, retries = 0, retryDelayMs = 5000, headers, report } = options;
-    let lastError;
-    for (let attempt = 0; attempt <= retries; attempt++) {
-        if (attempt > 0) {
-            report?.(`retry ${attempt}/${retries} after ${retryDelayMs}ms — ${url}`);
-            await sleep(retryDelayMs);
-        }
-        let res;
-        try {
-            // Raw `fetch`, deliberately: this is the shared FILE downloader and the body is piped to disk below.
-            // `APIClient` is the repo default for API requests — small bodies, repeated calls — and buffers a
-            // non-stream response in memory, which is the one thing a multi-gigabyte transfer must not do.
-            res = await fetch(url, { headers, signal: AbortSignal.timeout(timeoutMs) });
-        }
-        catch (error) {
-            // AbortSignal timeouts and network-level failures are retryable.
-            lastError = error;
-            continue;
-        }
-        if (!res.ok) {
-            const error = new HTTPStatusError(res.status, `HTTP ${res.status} ${res.statusText} — ${url}`);
-            if (!isTransientStatus(res.status))
-                throw error;
-            lastError = error;
-            continue;
-        }
-        try {
-            const buffer = Buffer.from(await res.arrayBuffer());
-            await writeFile(dest, buffer);
-            return { bytes: buffer.byteLength };
-        }
-        catch (error) {
-            // A mid-stream abort while reading the body is retryable too.
-            lastError = error;
-        }
-    }
-    throw lastError instanceof Error ? lastError : new Error(String(lastError));
-}
-/**
- * Read a MANIFEST.json; `null` when missing or corrupt (callers re-fetch from scratch).
- */
-export async function readManifest(path) {
-    if (!existsSync(path))
-        return null;
-    // A read failure (e.g. the file vanished after the existsSync probe) maps to null like corrupt
-    // JSON does; tryParsingJSON returns null for the non-string sentinel.
-    const text = await readFile(path, "utf8").catch(() => null);
-    return tryParsingJSON(text);
-}
-/**
- * Load manifest entries into a map so untouched keys survive a partial re-fetch.
- */
-export async function loadManifestEntries(path, key) {
-    const entries = new Map();
-    const parsed = await readManifest(path);
-    for (const entry of parsed ?? []) {
-        entries.set(key(entry), entry);
-    }
-    return entries;
-}
-/**
- * Write a MANIFEST.json in the house shape: pretty-printed, trailing newline.
- */
-export async function writeManifest(path, manifest) {
-    await writeFile(path, JSON.stringify(manifest, null, 2) + "\n");
-}
-//# sourceMappingURL=download.js.map

package/out/src/tools/fetch/download.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"download.js","sourceRoot":"","sources":["../../../../src/tools/fetch/download.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAA;AACpC,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAA;AACtD,OAAO,EAAE,UAAU,IAAI,KAAK,EAAE,MAAM,sBAAsB,CAAA;AAE1D,OAAO,EAAE,cAAc,EAAE,MAAM,yBAAyB,CAAA;AAExD;;GAEG;AACH,MAAM,sBAAsB,GAAG,GAAG,CAAA;AAElC;;GAEG;AACH,MAAM,qBAAqB,GAAG,GAAG,CAAA;AAEjC;;GAEG;AACH,MAAM,qBAAqB,GAAG,GAAG,CAAA;AAsBjC;;GAEG;AACH;;;;;;GAMG;AACH,MAAM,OAAO,eAAgB,SAAQ,KAAK;IAChC,MAAM,CAAQ;IAEvB,YAAY,MAAc,EAAE,OAAe;QAC1C,KAAK,CAAC,OAAO,CAAC,CAAA;QACd,IAAI,CAAC,IAAI,GAAG,iBAAiB,CAAA;QAC7B,IAAI,CAAC,MAAM,GAAG,MAAM,CAAA;IACrB,CAAC;CACD;AAED,MAAM,UAAU,iBAAiB,CAAC,MAAc;IAC/C,OAAO,MAAM,KAAK,sBAAsB,IAAI,CAAC,MAAM,IAAI,qBAAqB,IAAI,MAAM,IAAI,qBAAqB,CAAC,CAAA;AACjH,CAAC;AAqBD;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,cAAc,CAAC,OAAwB;IAC5D,MAAM,EAAE,GAAG,EAAE,IAAI,EAAE,SAAS,GAAG,OAAO,EAAE,OAAO,GAAG,CAAC,EAAE,YAAY,GAAG,IAAI,EAAE,OAAO,EAAE,MAAM,EAAE,GAAG,OAAO,CAAA;IACrG,IAAI,SAAkB,CAAA;IAEtB,KAAK,IAAI,OAAO,GAAG,CAAC,EAAE,OAAO,IAAI,OAAO,EAAE,OAAO,EAAE,EAAE,CAAC;QACrD,IAAI,OAAO,GAAG,CAAC,EAAE,CAAC;YACjB,MAAM,EAAE,CAAC,SAAS,OAAO,IAAI,OAAO,UAAU,YAAY,QAAQ,GAAG,EAAE,CAAC,CAAA;YACxE,MAAM,KAAK,CAAC,YAAY,CAAC,CAAA;QAC1B,CAAC;QAED,IAAI,GAAa,CAAA;QAEjB,IAAI,CAAC;YACJ,qGAAqG;YACrG,kGAAkG;YAClG,+FAA+F;YAC/F,GAAG,GAAG,MAAM,KAAK,CAAC,GAAG,EAAE,EAAE,OAAO,EAAE,MAAM,EAAE,WAAW,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,CAAC,CAAA;QAC5E,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YAChB,iEAAiE;YACjE,SAAS,GAAG,KAAK,CAAA;YAEjB,SAAQ;QACT,CAAC;QAED,IAAI,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC;YACb,MAAM,KAAK,GAAG,IAAI,eAAe,CAAC,GAAG,CAAC,MAAM,EAAE,QAAQ,GAAG,CAAC,MAAM,IAAI,GAAG,CAAC,UAAU,MAAM,GAAG,EAAE,CAAC,CAAA;YAE9F,IAAI,CAAC,iBAAiB,CAAC,GAAG,CAAC,MAAM,CAAC;gBAAE,MAAM,KAAK,CAAA;YAC/C,SAAS,GAAG,KAAK,CAAA;YAEjB,SAAQ;QACT,CAAC;QAED,IAAI,CAAC;YACJ,MAAM,MAAM,GAAG,MAAM,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,WAAW,EAAE,CAAC,CAAA;YACnD,MAAM,SAAS,CAAC,IAAI,EAAE,MAAM,CAAC,CAAA;YAE7B,OAAO,EAAE,KAAK,EAAE,MAAM,CAAC,UAAU,EAAE,CAAA;QACpC,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YAChB,8DAA8D;YAC9D,SAAS,GAAG,KAAK,CAAA;QAClB,CAAC;IACF,CAAC;IAED,MAAM,SAAS,YAAY,KAAK,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,CAAA;AAC5E,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,YAAY,CAAI,IAAY;IACjD,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC;QAAE,OAAO,IAAI,CAAA;IAElC,+FAA+F;IAC/F,sEAAsE;IACtE,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,CAAA;IAE3D,OAAO,cAAc,CAAI,IAAI,CAAC,CAAA;AAC/B,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,mBAAmB,CAAI,IAAY,EAAE,GAAyB;IACnF,MAAM,OAAO,GAAG,IAAI,GAAG,EAAa,CAAA;IACpC,MAAM,MAAM,GAAG,MAAM,YAAY,CAAM,IAAI,CAAC,CAAA;IAE5C,KAAK,MAAM,KAAK,IAAI,MAAM,IAAI,EAAE,EAAE,CAAC;QAClC,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,KAAK,CAAC,CAAA;IAC/B,CAAC;IAED,OAAO,OAAO,CAAA;AACf,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,aAAa,CAAC,IAAY,EAAE,QAAiB;IAClE,MAAM,SAAS,CAAC,IAAI,EAAE,IAAI,CAAC,SAAS,CAAC,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,CAAA;AAChE,CAAC"}

package/out/src/tools/fetch/geonames-dump.d.ts

@@ -1,87 +0,0 @@
-/**
- * @copyright Sister Software
- * @license AGPL-3.0
- * @author Teffen Ellis, et al.
- *
- *   Fetch GeoNames per-country GAZETTEER dumps — the 19-column `<CC>.txt` files under
- *   `https://download.geonames.org/export/dump/` (NOT the postal exports; those are `export/zip/` and
- *   `geonames-postal.ts`'s job). The dumps carry feature classes and codes (column 8: `PPLC` national capital,
- *   `PPLA` first-order administrative seat), which is what the capitals reference build consumes (#1880).
- *
- *   The catalog question is answered by the SOURCE, not by an ISO list: `countryInfo.txt` in the same directory
- *   enumerates every country GeoNames publishes, one row per ISO alpha-2 code, and also names each country's
- *   capital — the cross-check the capitals build grades its `PPLC` extraction against. Fetch that first; derive
- *   the country set from it; then a dump absent from disk is a measured gap against the source's own catalog
- *   rather than a silent hole. The dump directory may hold files this tool did not fetch: present files are never
- *   overwritten, and a present `<CC>.txt` that is NOT a 19-column gazetteer dump (GeoNames' postal exports share
- *   the basename) is reported as `wrong_format_present`, never counted as coverage.
- */
-import type { BaseFetchOptions, FetchSummary } from "./download.ts";
-export interface FetchGeonamesDumpOptions extends BaseFetchOptions {
-    /**
-     * ISO alpha-2 codes, any casing. Absent → every country `countryInfo.txt` enumerates.
-     */
-    countries?: readonly string[];
-    /**
-     * Dump directory to read from. Defaults to GeoNames' own; exists so the 404 and coverage behaviour can be exercised
-     * against a local server.
-     */
-    baseURL?: string;
-    /**
-     * Refetch a dump whose `<CC>.txt` already exists. Default false — the tool fills gaps.
-     */
-    force?: boolean;
-}
-interface GeonamesDumpFileEntry {
-    country: string;
-    filename: string;
-    source_url: string;
-    sha256: string;
-    bytes: number;
-}
-export interface GeonamesDumpManifest {
-    source: string;
-    base_url: string;
-    license: string;
-    attribution: string;
-    downloaded_at: string;
-    files: GeonamesDumpFileEntry[];
-    /**
-     * `<CC>.txt` files already on disk and left alone — the hand-fetched population this tool extends.
-     */
-    skipped_present: string[];
-    /**
-     * Countries in the source catalog that the source's dump directory nonetheless 404s — a fact about the source,
-     * recorded so a later reader does not spend the fetch to rediscover it.
-     */
-    unavailable: string[];
-    /**
-     * Present `<CC>.txt` files that are NOT 19-column gazetteer dumps — GeoNames' postal exports share the same basename,
-     * and seven tier-1 postal files sat at these paths reading as "present" until the capitals build found them
-     * capital-less. Left in place (this tool never overwrites data it did not fetch); the fix is to move the file to its
-     * own home and rerun.
-     */
-    wrong_format_present: string[];
-}
-/**
- * True when the first non-empty line carries the gazetteer dump's 19 tab-separated columns. Accepts a partial head read
- * — the first line is the whole question, so callers need not hand it a resident 350 MB dump.
- */
-export declare function looksLikeGazetteerDump(text: string): boolean;
-/**
- * Parse the ISO codes (column 1) and capital names (column 6) out of `countryInfo.txt` — `#`-prefixed lines are the
- * file's own commentary.
- */
-export declare function parseCountryInfo(text: string): Array<{
-    country: string;
-    capital: string;
-}>;
-/**
- * Download `countryInfo.txt` plus every missing `<CC>.zip`, extracting each to `<outRoot>/<CC>.txt` beside the
- * hand-fetched dumps, with a `MANIFEST.json` naming fetched, skipped-present, and source-unavailable countries.
- */
-export declare function fetchGeonamesDumps(options: FetchGeonamesDumpOptions, report?: (line: string) => void): Promise<FetchSummary & {
-    skippedPresent: string[];
-}>;
-export {};
-//# sourceMappingURL=geonames-dump.d.ts.map

package/out/src/tools/fetch/geonames-dump.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"geonames-dump.d.ts","sourceRoot":"","sources":["../../../../src/tools/fetch/geonames-dump.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AASH,OAAO,KAAK,EAAE,gBAAgB,EAAE,YAAY,EAAE,MAAM,eAAe,CAAA;AAgBnE,MAAM,WAAW,wBAAyB,SAAQ,gBAAgB;IACjE;;OAEG;IACH,SAAS,CAAC,EAAE,SAAS,MAAM,EAAE,CAAA;IAC7B;;;OAGG;IACH,OAAO,CAAC,EAAE,MAAM,CAAA;IAChB;;OAEG;IACH,KAAK,CAAC,EAAE,OAAO,CAAA;CACf;AAED,UAAU,qBAAqB;IAC9B,OAAO,EAAE,MAAM,CAAA;IACf,QAAQ,EAAE,MAAM,CAAA;IAChB,UAAU,EAAE,MAAM,CAAA;IAClB,MAAM,EAAE,MAAM,CAAA;IACd,KAAK,EAAE,MAAM,CAAA;CACb;AAED,MAAM,WAAW,oBAAoB;IACpC,MAAM,EAAE,MAAM,CAAA;IACd,QAAQ,EAAE,MAAM,CAAA;IAChB,OAAO,EAAE,MAAM,CAAA;IACf,WAAW,EAAE,MAAM,CAAA;IACnB,aAAa,EAAE,MAAM,CAAA;IACrB,KAAK,EAAE,qBAAqB,EAAE,CAAA;IAC9B;;OAEG;IACH,eAAe,EAAE,MAAM,EAAE,CAAA;IACzB;;;OAGG;IACH,WAAW,EAAE,MAAM,EAAE,CAAA;IACrB;;;;;OAKG;IACH,oBAAoB,EAAE,MAAM,EAAE,CAAA;CAC9B;AAQD;;;GAGG;AACH,wBAAgB,sBAAsB,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAuB5D;AAED;;;GAGG;AACH,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,MAAM,GAAG,KAAK,CAAC;IAAE,OAAO,EAAE,MAAM,CAAC;IAAC,OAAO,EAAE,MAAM,CAAA;CAAE,CAAC,CAiB1F;AAqBD;;;GAGG;AACH,wBAAsB,kBAAkB,CACvC,OAAO,EAAE,wBAAwB,EACjC,MAAM,CAAC,EAAE,CAAC,IAAI,EAAE,MAAM,KAAK,IAAI,GAC7B,OAAO,CAAC,YAAY,GAAG;IAAE,cAAc,EAAE,MAAM,EAAE,CAAA;CAAE,CAAC,CA2FtD"}