@mailwoman/corpus 9.4.0 → 10.1.0

package/out/jp/tools/fetch/houjin.js

@@ -0,0 +1,106 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ *   Fetch the National Tax Agency's corporate-number register (法人番号公表サイト 全件データ): every
+ *   corporation assigned a number, about 5,000,000 rows, with the head-office address as one string
+ *   (国内所在地) plus the prefecture and municipality split out and the postcode. The nationwide CSV
+ *   in Unicode is one zip of about 255 MB.
+ *
+ *   The agency issues the corporate number, so the identifier is an `identity` assertion. The head-office
+ *   string beside it is an `observation` the corporation filed, on `registered-office` rows. The Overture
+ *   rows the JP corpus is built from assert `address` on `premise` rows. That distinction matters when the two are mixed.
+ *
+ *   The agency states the three published fields may be used freely by anyone ("どなたでも自由にご利用
+ *   いただくことができます"), with no attribution condition on the download page.
+ *
+ *   The file sits behind a form. The page hands out a per-session token. A post with that token,
+ *   `event=download` and the file number of the nationwide Unicode CSV answers the zip. The file
+ *   number is read off the page rather than pinned, because the agency re-issues the files monthly.
+ *
+ *   Invoke via `mailwoman corpus fetch houjin-jp --out-root <path>`.
+ */
+import { makeDirectories } from "@mailwoman/core/fs/writers";
+import { sha256File } from "@mailwoman/core/hash";
+import { attachmentFilename, cookieHeader, streamBodyToFile, withRetries, writeManifest } from "#tools/fetch/download";
+const SLUG = "houjin-jp";
+const PAGE_URL = "https://www.houjin-bangou.nta.go.jp/download/zenken/";
+const FORM_URL = `${PAGE_URL}index.html`;
+const TOKEN_FIELD = "jp.go.nta.houjin_bangou.framework.web.common.CNSFWTokenProcessor.request.token";
+const LICENSE = "国税庁法人番号公表サイト — the three published fields are free for anyone to use (利用規約)";
+const ATTRIBUTION = "国税庁 法人番号公表サイト (National Tax Agency, Corporate Number Publication Site)";
+const USER_AGENT = "Mozilla/5.0 (X11; Linux x86_64) mailwoman-corpus-fetch";
+/**
+ * The nationwide row of the "CSV形式・Unicode" table stores the file number in its `doDownload(N)` handler.
+ */
+function nationwideUnicodeFileNumber(html) {
+    const start = html.indexOf('id="csv-unicode"');
+    if (start === -1)
+        return undefined;
+    const section = html.slice(start, html.indexOf('id="xml-unicode"', start));
+    const row = /全国[\s\S]{0,600}?doDownload\((\d+)\)/.exec(section);
+    return row?.[1];
+}
+export async function fetchHoujinJP(options, report) {
+    const destDir = options.outRoot(SLUG);
+    await makeDirectories(destDir);
+    const page = await fetch(PAGE_URL, { headers: { "user-agent": USER_AGENT, accept: "text/html" } });
+    if (!page.ok)
+        throw new Error(`houjin-jp: the download page answered HTTP ${page.status}`);
+    const html = await page.text();
+    const token = new RegExp(`${TOKEN_FIELD.replaceAll(".", "\\.")}" value="([^"]+)"`).exec(html)?.[1];
+    const fileNumber = nationwideUnicodeFileNumber(html);
+    if (!token || !fileNumber) {
+        report?.(`  ✗ the page carried ${token ? "a token" : "no token"} and ${fileNumber ? `file ${fileNumber}` : "no nationwide Unicode CSV row"}`);
+        return { fetched: 0, skipped: 0, failed: 1, failedCodes: ["zenkoku-csv-unicode"] };
+    }
+    report?.(`=== ${SLUG}: nationwide CSV (Unicode), file ${fileNumber}`);
+    let filename = `zenkoku_all_${fileNumber}.zip`;
+    let bytes;
+    try {
+        bytes = await withRetries(async () => {
+            // A fresh page per attempt.
+            // The token is bound to the session cookie.
+            // Either may have lapsed.
+            const attemptPage = await fetch(PAGE_URL, { headers: { "user-agent": USER_AGENT, accept: "text/html" } });
+            const attemptToken = new RegExp(`${TOKEN_FIELD.replaceAll(".", "\\.")}" value="([^"]+)"`).exec(await attemptPage.text())?.[1];
+            if (!attemptToken)
+                throw new Error("the download page carried no token");
+            const res = await fetch(FORM_URL, {
+                method: "POST",
+                headers: {
+                    "user-agent": USER_AGENT,
+                    cookie: cookieHeader(attemptPage),
+                    referer: PAGE_URL,
+                    "content-type": "application/x-www-form-urlencoded",
+                },
+                body: new URLSearchParams({ [TOKEN_FIELD]: attemptToken, event: "download", selDlFileNo: fileNumber }),
+                signal: AbortSignal.timeout(1_800_000),
+            });
+            if (!res.ok || !(res.headers.get("content-type") ?? "").includes("octet-stream")) {
+                throw new Error(`HTTP ${res.status} ${res.headers.get("content-type") ?? ""}`);
+            }
+            filename = attachmentFilename(res, filename);
+            return streamBodyToFile(res, destDir(filename));
+        }, { report, label: filename });
+    }
+    catch (error) {
+        report?.(`  ✗ ${error.message}`);
+        return { fetched: 0, skipped: 0, failed: 1, failedCodes: ["zenkoku-csv-unicode"] };
+    }
+    const dest = destDir(filename);
+    const sha = await sha256File(dest);
+    report?.(`  ✓ ${filename} ${(bytes / 1024 / 1024).toFixed(1)} MB  sha256=${sha}`);
+    const manifest = {
+        source: SLUG,
+        source_url: PAGE_URL,
+        license: LICENSE,
+        attribution: ATTRIBUTION,
+        downloaded_at: new Date().toISOString(),
+        files: [{ source_url: FORM_URL, downloaded_at: new Date().toISOString(), filename, sha256: sha, bytes }],
+    };
+    await writeManifest(destDir("MANIFEST.json"), manifest);
+    return { fetched: 1, skipped: 0, failed: 0, failedCodes: [] };
+}
+//# sourceMappingURL=houjin.js.map

package/out/jp/tools/fetch/houjin.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"houjin.js","sourceRoot":"","sources":["../../../../lib/jp/tools/fetch/houjin.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;GAsBG;AAEH,OAAO,EAAE,eAAe,EAAE,MAAM,4BAA4B,CAAA;AAC5D,OAAO,EAAE,UAAU,EAAE,MAAM,sBAAsB,CAAA;AAGjD,OAAO,EAAE,kBAAkB,EAAE,YAAY,EAAE,gBAAgB,EAAE,WAAW,EAAE,aAAa,EAAE,MAAM,uBAAuB,CAAA;AAEtH,MAAM,IAAI,GAAG,WAAW,CAAA;AACxB,MAAM,QAAQ,GAAG,sDAAsD,CAAA;AACvE,MAAM,QAAQ,GAAG,GAAG,QAAQ,YAAY,CAAA;AACxC,MAAM,WAAW,GAAG,gFAAgF,CAAA;AACpG,MAAM,OAAO,GAAG,6EAA6E,CAAA;AAC7F,MAAM,WAAW,GAAG,wEAAwE,CAAA;AAC5F,MAAM,UAAU,GAAG,wDAAwD,CAAA;AAI3E;;GAEG;AACH,SAAS,2BAA2B,CAAC,IAAY;IAChD,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,kBAAkB,CAAC,CAAA;IAE9C,IAAI,KAAK,KAAK,CAAC,CAAC;QAAE,OAAO,SAAS,CAAA;IAClC,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,IAAI,CAAC,OAAO,CAAC,kBAAkB,EAAE,KAAK,CAAC,CAAC,CAAA;IAC1E,MAAM,GAAG,GAAG,qCAAqC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAA;IAE/D,OAAO,GAAG,EAAE,CAAC,CAAC,CAAC,CAAA;AAChB,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,aAAa,CAClC,OAA6B,EAC7B,MAA+B;IAE/B,MAAM,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,IAAI,CAAC,CAAA;IACrC,MAAM,eAAe,CAAC,OAAO,CAAC,CAAA;IAE9B,MAAM,IAAI,GAAG,MAAM,KAAK,CAAC,QAAQ,EAAE,EAAE,OAAO,EAAE,EAAE,YAAY,EAAE,UAAU,EAAE,MAAM,EAAE,WAAW,EAAE,EAAE,CAAC,CAAA;IAElG,IAAI,CAAC,IAAI,CAAC,EAAE;QAAE,MAAM,IAAI,KAAK,CAAC,8CAA8C,IAAI,CAAC,MAAM,EAAE,CAAC,CAAA;IAC1F,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,IAAI,EAAE,CAAA;IAC9B,MAAM,KAAK,GAAG,IAAI,MAAM,CAAC,GAAG,WAAW,CAAC,UAAU,CAAC,GAAG,EAAE,KAAK,CAAC,mBAAmB,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;IAClG,MAAM,UAAU,GAAG,2BAA2B,CAAC,IAAI,CAAC,CAAA;IAEpD,IAAI,CAAC,KAAK,IAAI,CAAC,UAAU,EAAE,CAAC;QAC3B,MAAM,EAAE,CACP,wBAAwB,KAAK,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,UAAU,QAAQ,UAAU,CAAC,CAAC,CAAC,QAAQ,UAAU,EAAE,CAAC,CAAC,CAAC,+BAA+B,EAAE,CACnI,CAAA;QAED,OAAO,EAAE,OAAO,EAAE,CAAC,EAAE,OAAO,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,WAAW,EAAE,CAAC,qBAAqB,CAAC,EAAE,CAAA;IACnF,CAAC;IAED,MAAM,EAAE,CAAC,OAAO,IAAI,oCAAoC,UAAU,EAAE,CAAC,CAAA;IAErE,IAAI,QAAQ,GAAG,eAAe,UAAU,MAAM,CAAA;IAC9C,IAAI,KAAa,CAAA;IAEjB,IAAI,CAAC;QACJ,KAAK,GAAG,MAAM,WAAW,CACxB,KAAK,IAAI,EAAE;YACV,4BAA4B;YAC5B,4CAA4C;YAC5C,0BAA0B;YAC1B,MAAM,WAAW,GAAG,MAAM,KAAK,CAAC,QAAQ,EAAE,EAAE,OAAO,EAAE,EAAE,YAAY,EAAE,UAAU,EAAE,MAAM,EAAE,WAAW,EAAE,EAAE,CAAC,CAAA;YAEzG,MAAM,YAAY,GAAG,IAAI,MAAM,CAAC,GAAG,WAAW,CAAC,UAAU,CAAC,GAAG,EAAE,KAAK,CAAC,mBAAmB,CAAC,CAAC,IAAI,CAC7F,MAAM,WAAW,CAAC,IAAI,EAAE,CACxB,EAAE,CAAC,CAAC,CAAC,CAAA;YAEN,IAAI,CAAC,YAAY;gBAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAA;YAExE,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,QAAQ,EAAE;gBACjC,MAAM,EAAE,MAAM;gBACd,OAAO,EAAE;oBACR,YAAY,EAAE,UAAU;oBACxB,MAAM,EAAE,YAAY,CAAC,WAAW,CAAC;oBACjC,OAAO,EAAE,QAAQ;oBACjB,cAAc,EAAE,mCAAmC;iBACnD;gBACD,IAAI,EAAE,IAAI,eAAe,CAAC,EAAE,CAAC,WAAW,CAAC,EAAE,YAAY,EAAE,KAAK,EAAE,UAAU,EAAE,WAAW,EAAE,UAAU,EAAE,CAAC;gBACtG,MAAM,EAAE,WAAW,CAAC,OAAO,CAAC,SAAS,CAAC;aACtC,CAAC,CAAA;YAEF,IAAI,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,IAAI,EAAE,CAAC,CAAC,QAAQ,CAAC,cAAc,CAAC,EAAE,CAAC;gBAClF,MAAM,IAAI,KAAK,CAAC,QAAQ,GAAG,CAAC,MAAM,IAAI,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,IAAI,EAAE,EAAE,CAAC,CAAA;YAC/E,CAAC;YAED,QAAQ,GAAG,kBAAkB,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAA;YAE5C,OAAO,gBAAgB,CAAC,GAAG,EAAE,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAA;QAChD,CAAC,EACD,EAAE,MAAM,EAAE,KAAK,EAAE,QAAQ,EAAE,CAC3B,CAAA;IACF,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QAChB,MAAM,EAAE,CAAC,OAAQ,KAAe,CAAC,OAAO,EAAE,CAAC,CAAA;QAE3C,OAAO,EAAE,OAAO,EAAE,CAAC,EAAE,OAAO,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,WAAW,EAAE,CAAC,qBAAqB,CAAC,EAAE,CAAA;IACnF,CAAC;IAED,MAAM,IAAI,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAA;IAC9B,MAAM,GAAG,GAAG,MAAM,UAAU,CAAC,IAAI,CAAC,CAAA;IAClC,MAAM,EAAE,CAAC,OAAO,QAAQ,IAAI,CAAC,KAAK,GAAG,IAAI,GAAG,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,eAAe,GAAG,EAAE,CAAC,CAAA;IAEjF,MAAM,QAAQ,GAA6B;QAC1C,MAAM,EAAE,IAAI;QACZ,UAAU,EAAE,QAAQ;QACpB,OAAO,EAAE,OAAO;QAChB,WAAW,EAAE,WAAW;QACxB,aAAa,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;QACvC,KAAK,EAAE,CAAC,EAAE,UAAU,EAAE,QAAQ,EAAE,aAAa,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,GAAG,EAAE,KAAK,EAAE,CAAC;KACxG,CAAA;IAED,MAAM,aAAa,CAAC,OAAO,CAAC,eAAe,CAAC,EAAE,QAAQ,CAAC,CAAA;IAEvD,OAAO,EAAE,OAAO,EAAE,CAAC,EAAE,OAAO,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,WAAW,EAAE,EAAE,EAAE,CAAA;AAC9D,CAAC"}

package/out/kr/adapters/juso/label-rows.d.ts

@@ -0,0 +1,62 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ * The Korean road-name address register (주소DB) used as alignment ground truth. The ministry assigns
+ * the road names and building numbers it publishes, so the register asserts `address` on `premise`
+ * rows where the permit registry's address field is an `observation`. The archive is four
+ * pipe-delimited CP949 text files per 시도 plus one nationwide road-code file, as the guide inside the
+ * zip states (붙임1, the 전체분 layout), and both its member names and its contents need the publisher's
+ * encoding.
+ */
+import type { PathBuilderLike } from "path-ts";
+/**
+ * The 2026 edition merged 전남광주통합특별시 while older sources still write 전라남도 and 광주광역시,
+ * so a typed string in either form aligns against the register through this map.
+ */
+export declare const REGION_ALIASES: Readonly<Record<string, string>>;
+/**
+ * One address of the register, with the pieces the road form and the lot form each write.
+ */
+export interface JusoLabelRow {
+    region: string;
+    sigungu: string;
+    /**
+     * The 읍/면 included in the road address between the 시군구 and the road (읍면동구분 `0`), else empty.
+     *
+     * A 동 is never written in the road form, it goes in the parenthetical.
+     */
+    eupmyeon: string;
+    dong: string;
+    ri: string;
+    road: string;
+    number: string;
+    postcode: string;
+    building: string;
+    lot: string;
+    underground: boolean;
+}
+/**
+ * What the official form writes in parentheses after the number: the 법정동 in a 동 area, the 리 in an 읍/면.
+ */
+export declare function parenthetical(row: JusoLabelRow): string;
+/**
+ * Every full-edition member of one kind as `[region suffix, archive name]`, sorted by region.
+ *
+ * The monthly edition's 변동 (change-only) files under the same prefixes are excluded by name.
+ */
+export declare function regionMembers(names: readonly string[], prefix: string): Array<[string, string]>;
+export interface ReadJusoOptions {
+    /**
+     * Stop after this many rows per 시도, for a fixture or a smoke run.
+     * Omit for the whole register.
+     */
+    maxRowsPerRegion?: number;
+}
+/**
+ * Stream every label row region by region, holding one 시도's lot and supplement files
+ * while its address file streams, so the join never holds the country at once.
+ */
+export declare function readJusoLabelRows(archivePath: PathBuilderLike, options?: ReadJusoOptions): AsyncGenerator<JusoLabelRow>;
+//# sourceMappingURL=label-rows.d.ts.map

package/out/kr/adapters/juso/label-rows.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"label-rows.d.ts","sourceRoot":"","sources":["../../../../lib/kr/adapters/juso/label-rows.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAIH,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,SAAS,CAAA;AA0C9C;;;GAGG;AACH,eAAO,MAAM,cAAc,EAAE,QAAQ,CAAC,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAK3D,CAAA;AAED;;GAEG;AACH,MAAM,WAAW,YAAY;IAC5B,MAAM,EAAE,MAAM,CAAA;IACd,OAAO,EAAE,MAAM,CAAA;IACf;;;;OAIG;IACH,QAAQ,EAAE,MAAM,CAAA;IAChB,IAAI,EAAE,MAAM,CAAA;IACZ,EAAE,EAAE,MAAM,CAAA;IACV,IAAI,EAAE,MAAM,CAAA;IACZ,MAAM,EAAE,MAAM,CAAA;IACd,QAAQ,EAAE,MAAM,CAAA;IAChB,QAAQ,EAAE,MAAM,CAAA;IAChB,GAAG,EAAE,MAAM,CAAA;IACX,WAAW,EAAE,OAAO,CAAA;CACpB;AAED;;GAEG;AACH,wBAAgB,aAAa,CAAC,GAAG,EAAE,YAAY,GAAG,MAAM,CAEvD;AAED;;;;GAIG;AACH,wBAAgB,aAAa,CAAC,KAAK,EAAE,SAAS,MAAM,EAAE,EAAE,MAAM,EAAE,MAAM,GAAG,KAAK,CAAC,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC,CAK/F;AA6CD,MAAM,WAAW,eAAe;IAC/B;;;OAGG;IACH,gBAAgB,CAAC,EAAE,MAAM,CAAA;CACzB;AAED;;;GAGG;AACH,wBAAuB,iBAAiB,CACvC,WAAW,EAAE,eAAe,EAC5B,OAAO,GAAE,eAAoB,GAC3B,cAAc,CAAC,YAAY,CAAC,CAuE9B"}

package/out/kr/adapters/juso/label-rows.js

@@ -0,0 +1,172 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ * The Korean road-name address register (주소DB) used as alignment ground truth. The ministry assigns
+ * the road names and building numbers it publishes, so the register asserts `address` on `premise`
+ * rows where the permit registry's address field is an `observation`. The archive is four
+ * pipe-delimited CP949 text files per 시도 plus one nationwide road-code file, as the guide inside the
+ * zip states (붙임1, the 전체분 layout), and both its member names and its contents need the publisher's
+ * encoding.
+ */
+import { decodeByteStream } from "@mailwoman/core/fs/streams";
+import { listZipEntries, readZipEntry } from "@mailwoman/core/fs/zip";
+import { TextSpliterator } from "spliterator";
+/**
+ * The publisher's encoding; `cp949` rather than `euc-kr`, because the register's UHC extension appears
+ * in about one row in 48,000 and Node's whatwg `euc-kr` decodes it to a different string with no error.
+ */
+const ENCODING = "cp949";
+/**
+ * The road-code file's columns, 0-based.
+ */
+const RC = { code: 0, road: 1, serial: 3, region: 4, sigungu: 6, eupmyeondong: 8, kind: 10 };
+/**
+ * 주소 columns.
+ */
+const AD = { id: 0, roadCode: 1, serial: 2, underground: 3, main: 4, sub: 5, postcode: 6 };
+/**
+ * 지번 columns.
+ */
+const LOT = { id: 0, dong: 5, ri: 6, mountain: 7, main: 8, sub: 9, primary: 10 };
+/**
+ * 부가정보 columns.
+ */
+const SUP = { id: 0, postcode: 3, buildingRegister: 6, buildingLocal: 7, apartment: 8 };
+/**
+ * Joins 도로명코드 to 읍면동일련번호 with a NUL no code contains, since neither part has a fixed width
+ * and a printable separator could let two different pairs produce one key.
+ */
+const CODE_SEPARATOR = "\0";
+/**
+ * The road-code table's key for one 도로명코드 and 읍면동일련번호.
+ */
+function roadCodeKey(code, serial) {
+    return `${code}${CODE_SEPARATOR}${serial}`;
+}
+/**
+ * The 2026 edition merged 전남광주통합특별시 while older sources still write 전라남도 and 광주광역시,
+ * so a typed string in either form aligns against the register through this map.
+ */
+export const REGION_ALIASES = {
+    전라남도: "전남광주통합특별시",
+    광주광역시: "전남광주통합특별시",
+    강원도: "강원특별자치도",
+    전라북도: "전북특별자치도",
+};
+/**
+ * What the official form writes in parentheses after the number: the 법정동 in a 동 area, the 리 in an 읍/면.
+ */
+export function parenthetical(row) {
+    return row.eupmyeon ? row.ri : row.dong;
+}
+/**
+ * Every full-edition member of one kind as `[region suffix, archive name]`, sorted by region.
+ *
+ * The monthly edition's 변동 (change-only) files under the same prefixes are excluded by name.
+ */
+export function regionMembers(names, prefix) {
+    return names
+        .filter((name) => name.startsWith(`${prefix}_`) && name.endsWith(".txt") && !name.includes("변동"))
+        .map((name) => [name.slice(prefix.length + 1, -4), name])
+        .toSorted(([a], [b]) => a.localeCompare(b));
+}
+/**
+ * Split one CP949 pipe-delimited member into its fields, line by line, without holding the member.
+ */
+async function* fields(archivePath, member) {
+    const bytes = readZipEntry(archivePath, member, { filenameEncoding: ENCODING });
+    for await (const line of TextSpliterator.fromAsync(decodeByteStream(bytes, ENCODING))) {
+        yield line.split("|");
+    }
+}
+/**
+ * The road-code table keyed by {@link roadCodeKey}, held whole because every region's
+ * addresses join against it and read as one composite string key rather than a nested map.
+ */
+async function loadRoadCodes(archivePath, member) {
+    const codes = new Map();
+    for await (const row of fields(archivePath, member)) {
+        if (row.length <= RC.kind)
+            continue;
+        codes.set(roadCodeKey(row[RC.code], row[RC.serial]), [
+            row[RC.region],
+            row[RC.sigungu],
+            row[RC.eupmyeondong],
+            row[RC.road],
+            row[RC.kind],
+        ]);
+    }
+    return codes;
+}
+/**
+ * A house or lot number as the register writes it: leading zeros dropped, a `-<부번>` only when there is one.
+ */
+function joinNumber(main, sub) {
+    const head = main.replace(/^0+/u, "") || "0";
+    const tail = sub.replace(/^0+/u, "");
+    return tail && tail !== "0" ? `${head}-${tail}` : head;
+}
+/**
+ * Stream every label row region by region, holding one 시도's lot and supplement files
+ * while its address file streams, so the join never holds the country at once.
+ */
+export async function* readJusoLabelRows(archivePath, options = {}) {
+    const names = (await listZipEntries(archivePath, { filenameEncoding: ENCODING })).map((entry) => entry.name);
+    const roadCodeMember = names.find((name) => name.includes("도로명코드") && !name.includes("변동"));
+    if (!roadCodeMember) {
+        throw new Error(`no 도로명코드 member in ${archivePath} — the archive is not a 전체분 edition`);
+    }
+    const codes = await loadRoadCodes(archivePath, roadCodeMember);
+    const lotMembers = new Map(regionMembers(names, "지번"));
+    const supplementMembers = new Map(regionMembers(names, "부가정보"));
+    for (const [suffix, member] of regionMembers(names, "주소")) {
+        const lotMember = lotMembers.get(suffix);
+        const supplementMember = supplementMembers.get(suffix);
+        if (!lotMember || !supplementMember) {
+            throw new Error(`region ${suffix} has an address file but no 지번 or 부가정보 file`);
+        }
+        const lots = new Map();
+        for await (const row of fields(archivePath, lotMember)) {
+            if (row.length <= LOT.primary || row[LOT.primary] !== "1")
+                continue;
+            const lot = `${row[LOT.mountain] === "1" ? "산" : ""}${joinNumber(row[LOT.main], row[LOT.sub])}`;
+            lots.set(row[LOT.id], [row[LOT.dong], row[LOT.ri], lot]);
+        }
+        const supplements = new Map();
+        for await (const row of fields(archivePath, supplementMember)) {
+            if (row.length <= SUP.apartment)
+                continue;
+            supplements.set(row[SUP.id], [row[SUP.postcode], row[SUP.buildingLocal] || row[SUP.buildingRegister]]);
+        }
+        let emitted = 0;
+        for await (const row of fields(archivePath, member)) {
+            if (row.length <= AD.postcode)
+                continue;
+            const code = codes.get(roadCodeKey(row[AD.roadCode], row[AD.serial]));
+            if (!code)
+                continue;
+            const [region, sigungu, eupmyeondong, road, kind] = code;
+            const [dong = "", ri = "", lot = ""] = lots.get(row[AD.id]) ?? [];
+            const [postcode = row[AD.postcode], building = ""] = supplements.get(row[AD.id]) ?? [];
+            yield {
+                region,
+                sigungu,
+                eupmyeon: kind === "0" ? eupmyeondong : "",
+                dong,
+                ri,
+                road,
+                number: joinNumber(row[AD.main], row[AD.sub]),
+                postcode: postcode || row[AD.postcode],
+                building,
+                lot,
+                underground: row[AD.underground] === "1",
+            };
+            emitted += 1;
+            if (options.maxRowsPerRegion !== undefined && emitted >= options.maxRowsPerRegion)
+                break;
+        }
+    }
+}
+//# sourceMappingURL=label-rows.js.map

package/out/kr/adapters/juso/label-rows.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"label-rows.js","sourceRoot":"","sources":["../../../../lib/kr/adapters/juso/label-rows.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,OAAO,EAAE,gBAAgB,EAAE,MAAM,4BAA4B,CAAA;AAC7D,OAAO,EAAE,cAAc,EAAE,YAAY,EAAE,MAAM,wBAAwB,CAAA;AAErE,OAAO,EAAE,eAAe,EAAE,MAAM,aAAa,CAAA;AAE7C;;;GAGG;AACH,MAAM,QAAQ,GAAG,OAAO,CAAA;AAExB;;GAEG;AACH,MAAM,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,OAAO,EAAE,CAAC,EAAE,YAAY,EAAE,CAAC,EAAE,IAAI,EAAE,EAAE,EAAW,CAAA;AAErG;;GAEG;AACH,MAAM,EAAE,GAAG,EAAE,EAAE,EAAE,CAAC,EAAE,QAAQ,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,WAAW,EAAE,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,QAAQ,EAAE,CAAC,EAAW,CAAA;AAEnG;;GAEG;AACH,MAAM,GAAG,GAAG,EAAE,EAAE,EAAE,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,QAAQ,EAAE,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,OAAO,EAAE,EAAE,EAAW,CAAA;AAEzF;;GAEG;AACH,MAAM,GAAG,GAAG,EAAE,EAAE,EAAE,CAAC,EAAE,QAAQ,EAAE,CAAC,EAAE,gBAAgB,EAAE,CAAC,EAAE,aAAa,EAAE,CAAC,EAAE,SAAS,EAAE,CAAC,EAAW,CAAA;AAEhG;;;GAGG;AACH,MAAM,cAAc,GAAG,IAAI,CAAA;AAE3B;;GAEG;AACH,SAAS,WAAW,CAAC,IAAY,EAAE,MAAc;IAChD,OAAO,GAAG,IAAI,GAAG,cAAc,GAAG,MAAM,EAAE,CAAA;AAC3C,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,MAAM,cAAc,GAAqC;IAC/D,IAAI,EAAE,WAAW;IACjB,KAAK,EAAE,WAAW;IAClB,GAAG,EAAE,SAAS;IACd,IAAI,EAAE,SAAS;CACf,CAAA;AAwBD;;GAEG;AACH,MAAM,UAAU,aAAa,CAAC,GAAiB;IAC9C,OAAO,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,CAAA;AACxC,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,aAAa,CAAC,KAAwB,EAAE,MAAc;IACrE,OAAO,KAAK;SACV,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,MAAM,GAAG,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;SAChG,GAAG,CAAC,CAAC,IAAI,EAAoB,EAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;SAC1E,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,CAAA;AAC7C,CAAC;AAED;;GAEG;AACH,KAAK,SAAS,CAAC,CAAC,MAAM,CAAC,WAA4B,EAAE,MAAc;IAClE,MAAM,KAAK,GAAG,YAAY,CAAC,WAAW,EAAE,MAAM,EAAE,EAAE,gBAAgB,EAAE,QAAQ,EAAE,CAAC,CAAA;IAE/E,IAAI,KAAK,EAAE,MAAM,IAAI,IAAI,eAAe,CAAC,SAAS,CAAC,gBAAgB,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAC,EAAE,CAAC;QACvF,MAAM,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAA;IACtB,CAAC;AACF,CAAC;AAED;;;GAGG;AACH,KAAK,UAAU,aAAa,CAAC,WAA4B,EAAE,MAAc;IACxE,MAAM,KAAK,GAAG,IAAI,GAAG,EAAoB,CAAA;IAEzC,IAAI,KAAK,EAAE,MAAM,GAAG,IAAI,MAAM,CAAC,WAAW,EAAE,MAAM,CAAC,EAAE,CAAC;QACrD,IAAI,GAAG,CAAC,MAAM,IAAI,EAAE,CAAC,IAAI;YAAE,SAAQ;QAEnC,KAAK,CAAC,GAAG,CAAC,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAE,EAAE,GAAG,CAAC,EAAE,CAAC,MAAM,CAAE,CAAC,EAAE;YACtD,GAAG,CAAC,EAAE,CAAC,MAAM,CAAE;YACf,GAAG,CAAC,EAAE,CAAC,OAAO,CAAE;YAChB,GAAG,CAAC,EAAE,CAAC,YAAY,CAAE;YACrB,GAAG,CAAC,EAAE,CAAC,IAAI,CAAE;YACb,GAAG,CAAC,EAAE,CAAC,IAAI,CAAE;SACb,CAAC,CAAA;IACH,CAAC;IAED,OAAO,KAAK,CAAA;AACb,CAAC;AAED;;GAEG;AACH,SAAS,UAAU,CAAC,IAAY,EAAE,GAAW;IAC5C,MAAM,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,IAAI,GAAG,CAAA;IAC5C,MAAM,IAAI,GAAG,GAAG,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAA;IAEpC,OAAO,IAAI,IAAI,IAAI,KAAK,GAAG,CAAC,CAAC,CAAC,GAAG,IAAI,IAAI,IAAI,EAAE,CAAC,CAAC,CAAC,IAAI,CAAA;AACvD,CAAC;AAUD;;;GAGG;AACH,MAAM,CAAC,KAAK,SAAS,CAAC,CAAC,iBAAiB,CACvC,WAA4B,EAC5B,OAAO,GAAoB,EAAE;IAE7B,MAAM,KAAK,GAAG,CAAC,MAAM,cAAc,CAAC,WAAW,EAAE,EAAE,gBAAgB,EAAE,QAAQ,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;IAE5G,MAAM,cAAc,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAA;IAE3F,IAAI,CAAC,cAAc,EAAE,CAAC;QACrB,MAAM,IAAI,KAAK,CAAC,sBAAsB,WAAW,qCAAqC,CAAC,CAAA;IACxF,CAAC;IAED,MAAM,KAAK,GAAG,MAAM,aAAa,CAAC,WAAW,EAAE,cAAc,CAAC,CAAA;IAC9D,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,aAAa,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,CAAA;IACtD,MAAM,iBAAiB,GAAG,IAAI,GAAG,CAAC,aAAa,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,CAAA;IAE/D,KAAK,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,IAAI,aAAa,CAAC,KAAK,EAAE,IAAI,CAAC,EAAE,CAAC;QAC3D,MAAM,SAAS,GAAG,UAAU,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;QACxC,MAAM,gBAAgB,GAAG,iBAAiB,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;QAEtD,IAAI,CAAC,SAAS,IAAI,CAAC,gBAAgB,EAAE,CAAC;YACrC,MAAM,IAAI,KAAK,CAAC,UAAU,MAAM,6CAA6C,CAAC,CAAA;QAC/E,CAAC;QAED,MAAM,IAAI,GAAG,IAAI,GAAG,EAAoC,CAAA;QAExD,IAAI,KAAK,EAAE,MAAM,GAAG,IAAI,MAAM,CAAC,WAAW,EAAE,SAAS,CAAC,EAAE,CAAC;YACxD,IAAI,GAAG,CAAC,MAAM,IAAI,GAAG,CAAC,OAAO,IAAI,GAAG,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,GAAG;gBAAE,SAAQ;YAEnE,MAAM,GAAG,GAAG,GAAG,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,GAAG,UAAU,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAE,EAAE,GAAG,CAAC,GAAG,CAAC,GAAG,CAAE,CAAC,EAAE,CAAA;YAEjG,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAE,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAE,EAAE,GAAG,CAAC,GAAG,CAAC,EAAE,CAAE,EAAE,GAAG,CAAC,CAAC,CAAA;QAC5D,CAAC;QAED,MAAM,WAAW,GAAG,IAAI,GAAG,EAA4B,CAAA;QAEvD,IAAI,KAAK,EAAE,MAAM,GAAG,IAAI,MAAM,CAAC,WAAW,EAAE,gBAAgB,CAAC,EAAE,CAAC;YAC/D,IAAI,GAAG,CAAC,MAAM,IAAI,GAAG,CAAC,SAAS;gBAAE,SAAQ;YAEzC,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAE,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAE,EAAE,GAAG,CAAC,GAAG,CAAC,aAAa,CAAC,IAAI,GAAG,CAAC,GAAG,CAAC,gBAAgB,CAAE,CAAC,CAAC,CAAA;QAC1G,CAAC;QAED,IAAI,OAAO,GAAG,CAAC,CAAA;QAEf,IAAI,KAAK,EAAE,MAAM,GAAG,IAAI,MAAM,CAAC,WAAW,EAAE,MAAM,CAAC,EAAE,CAAC;YACrD,IAAI,GAAG,CAAC,MAAM,IAAI,EAAE,CAAC,QAAQ;gBAAE,SAAQ;YAEvC,MAAM,IAAI,GAAG,KAAK,CAAC,GAAG,CAAC,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC,QAAQ,CAAE,EAAE,GAAG,CAAC,EAAE,CAAC,MAAM,CAAE,CAAC,CAAC,CAAA;YAEvE,IAAI,CAAC,IAAI;gBAAE,SAAQ;YAEnB,MAAM,CAAC,MAAM,EAAE,OAAO,EAAE,YAAY,EAAE,IAAI,EAAE,IAAI,CAAC,GAAG,IAAgD,CAAA;YACpG,MAAM,CAAC,IAAI,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,GAAG,GAAG,EAAE,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,CAAE,CAAC,IAAI,EAAE,CAAA;YAClE,MAAM,CAAC,QAAQ,GAAG,GAAG,CAAC,EAAE,CAAC,QAAQ,CAAE,EAAE,QAAQ,GAAG,EAAE,CAAC,GAAG,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,CAAE,CAAC,IAAI,EAAE,CAAA;YAExF,MAAM;gBACL,MAAM;gBACN,OAAO;gBACP,QAAQ,EAAE,IAAI,KAAK,GAAG,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,EAAE;gBAC1C,IAAI;gBACJ,EAAE;gBACF,IAAI;gBACJ,MAAM,EAAE,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAE,EAAE,GAAG,CAAC,EAAE,CAAC,GAAG,CAAE,CAAC;gBAC/C,QAAQ,EAAE,QAAQ,IAAI,GAAG,CAAC,EAAE,CAAC,QAAQ,CAAE;gBACvC,QAAQ;gBACR,GAAG;gBACH,WAAW,EAAE,GAAG,CAAC,EAAE,CAAC,WAAW,CAAC,KAAK,GAAG;aACxC,CAAA;YAED,OAAO,IAAI,CAAC,CAAA;YAEZ,IAAI,OAAO,CAAC,gBAAgB,KAAK,SAAS,IAAI,OAAO,IAAI,OAAO,CAAC,gBAAgB;gBAAE,MAAK;QACzF,CAAC;IACF,CAAC;AACF,CAAC"}

package/out/kr/adapters/localdata/align.d.ts

@@ -0,0 +1,31 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ * Align one permit-registry address string to the address register's key, or answer null. Alignment
+ * is exact against {@link KeyIndex} rather than fuzzy, so a string that satisfies the whole key becomes
+ * a training row of matched spans and one that does not becomes a board row.
+ */
+import { type KeyIndex } from "#kr/adapters/localdata/key-index";
+/**
+ * One aligned string: the spans that matched and the key they matched under.
+ */
+export interface Aligned {
+    raw: string;
+    spanStarts: number[];
+    spanEnds: number[];
+    spanTags: string[];
+    register: "registry_road" | "registry_lot";
+    region: string;
+    sigungu: string;
+}
+/**
+ * Align one road-name string, or answer null.
+ */
+export declare function alignRoadAddress(text: string, index: KeyIndex): Aligned | null;
+/**
+ * Align one lot-number string, or answer null.
+ */
+export declare function alignLotAddress(text: string, index: KeyIndex): Aligned | null;
+//# sourceMappingURL=align.d.ts.map

package/out/kr/adapters/localdata/align.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"align.d.ts","sourceRoot":"","sources":["../../../../lib/kr/adapters/localdata/align.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,EAAE,KAAK,QAAQ,EAAwB,MAAM,kCAAkC,CAAA;AAuBtF;;GAEG;AACH,MAAM,WAAW,OAAO;IACvB,GAAG,EAAE,MAAM,CAAA;IACX,UAAU,EAAE,MAAM,EAAE,CAAA;IACpB,QAAQ,EAAE,MAAM,EAAE,CAAA;IAClB,QAAQ,EAAE,MAAM,EAAE,CAAA;IAClB,QAAQ,EAAE,eAAe,GAAG,cAAc,CAAA;IAC1C,MAAM,EAAE,MAAM,CAAA;IACd,OAAO,EAAE,MAAM,CAAA;CACf;AA6DD;;GAEG;AACH,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,QAAQ,GAAG,OAAO,GAAG,IAAI,CA6G9E;AAED;;GAEG;AACH,wBAAgB,eAAe,CAAC,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,QAAQ,GAAG,OAAO,GAAG,IAAI,CAiE7E"}

package/out/kr/adapters/localdata/align.js

@@ -0,0 +1,212 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ *
+ * Align one permit-registry address string to the address register's key, or answer null. Alignment
+ * is exact against {@link KeyIndex} rather than fuzzy, so a string that satisfies the whole key becomes
+ * a training row of matched spans and one that does not becomes a board row.
+ */
+import { sigunguSpan, unitKey } from "#kr/adapters/localdata/key-index";
+/**
+ * A building number: `233`, `43-1`, `지하 1476`.
+ */
+const BUILDING_NUMBER = /^(?:지하\s?)?\d+(?:-\d+)?$/u;
+/**
+ * A lot number: `43-1`, `산12`, `43번지`.
+ */
+const LOT_NUMBER = /^산?\d+(?:-\d+)?(?:번지)?$/u;
+/**
+ * A floor or unit the clerk appended: `2층`, `일부호`, `B1`, `지하1층`, `101동`.
+ */
+const UNIT_TOKEN = /^(?:지하\s?)?(?:B?\d+(?:~\d+)?(?:층|호)|\d+층|B\d+|지하\d*층?|\d+동|[가-힣]?\d*호)(?:,)?$/u;
+/**
+ * The shortest either form can be — 시도, 시군구 and the road or the 법정동.
+ * No shorter value can contain the key.
+ */
+const MINIMUM_TOKENS = 3;
+/**
+ * Record a span unless it is empty or blank.
+ *
+ * A blank span from a clerk's double space or a zero-width one at a token's own
+ * end would train the model on no component.
+ */
+function put(spans, text, start, end, tag) {
+    if (end > start && text.slice(start, end).trim()) {
+        spans.push({ start, end, tag });
+    }
+}
+/**
+ * Where each token begins in the original text, walked left to right so a repeated
+ * token takes its own position rather than the first one's.
+ */
+function tokenPositions(text, tokens, from = 0) {
+    const positions = [];
+    let cursor = from;
+    for (const token of tokens) {
+        const at = text.indexOf(token, cursor);
+        positions.push(at);
+        cursor = at + token.length;
+    }
+    return positions;
+}
+/**
+ * The index of the first unit-shaped token, or the length when there is none.
+ */
+function firstUnitIndex(tokens) {
+    const found = tokens.findIndex((token) => UNIT_TOKEN.test(token));
+    return found === -1 ? tokens.length : found;
+}
+function finish(text, spans, register, region, sigungu) {
+    const sorted = spans.toSorted((a, b) => a.start - b.start || a.end - b.end || a.tag.localeCompare(b.tag));
+    return {
+        raw: text,
+        spanStarts: sorted.map((span) => span.start),
+        spanEnds: sorted.map((span) => span.end),
+        spanTags: sorted.map((span) => span.tag),
+        register,
+        region,
+        sigungu,
+    };
+}
+/**
+ * Align one road-name string, or answer null.
+ */
+export function alignRoadAddress(text, index) {
+    const openAt = text.indexOf("(");
+    const head = openAt === -1 ? text : text.slice(0, openAt);
+    const tail = openAt === -1 ? "" : text.slice(openAt + 1);
+    const parenthetical = tail.endsWith(")") ? tail.slice(0, -1).trim() : "";
+    const commaAt = head.indexOf(",");
+    const core = commaAt === -1 ? head : head.slice(0, commaAt);
+    const detail = commaAt === -1 ? "" : head.slice(commaAt + 1);
+    const tokens = core.split(/\s+/u).filter((token) => token.length);
+    const positions = tokenPositions(text, tokens);
+    if (tokens.length < MINIMUM_TOKENS || !index.regions.has(tokens[0]))
+        return null;
+    const region = tokens[0];
+    const width = sigunguSpan(index, region, tokens, 1);
+    // A region with no 시군구 level (세종특별자치시) lists the empty string, so its strings go region → road.
+    if (!width && !index.sigunguByRegion.get(region)?.has(""))
+        return null;
+    const sigungu = tokens.slice(1, 1 + width).join(" ");
+    const unit = unitKey(region, sigungu);
+    const roads = index.roadsByUnit.get(unit) ?? new Set();
+    const dongs = index.dongsByUnit.get(unit) ?? new Set();
+    let roadAt = 1 + width;
+    let eupmyeonAt = null;
+    // In an 읍/면 area the road form includes the 읍/면 between the 시군구 and the road
+    // (`기장군 기장읍 기장해안로 205`); the register lists those names beside the 동 of the same unit.
+    if (roadAt + 2 < tokens.length && dongs.has(tokens[roadAt]) && roads.has(tokens[roadAt + 1])) {
+        eupmyeonAt = roadAt;
+        roadAt += 1;
+    }
+    // A road name is one token.
+    // A numbered branch (`대학로8길`) is part of that token in the register.
+    if (roadAt + 1 >= tokens.length || !roads.has(tokens[roadAt]))
+        return null;
+    let numberAtToken = roadAt + 1;
+    // `달구벌대로 지하 1476`: the underground marker stands as its own token before the number, outside every span.
+    if (tokens[numberAtToken] === "지하" && numberAtToken + 1 < tokens.length) {
+        numberAtToken += 1;
+    }
+    const number = tokens[numberAtToken];
+    if (!BUILDING_NUMBER.test(number))
+        return null;
+    const spans = [];
+    put(spans, text, positions[0], positions[0] + region.length, "region");
+    for (let offset = 0; offset < width; offset += 1) {
+        const token = tokens[1 + offset];
+        put(spans, text, positions[1 + offset], positions[1 + offset] + token.length, "subregion");
+    }
+    if (eupmyeonAt !== null) {
+        put(spans, text, positions[eupmyeonAt], positions[eupmyeonAt] + tokens[eupmyeonAt].length, "dependent_locality");
+    }
+    put(spans, text, positions[roadAt], positions[roadAt] + tokens[roadAt].length, "street");
+    const numberAt = positions[numberAtToken];
+    put(spans, text, numberAt, numberAt + number.length, "house_number");
+    // What follows the number, with or without a comma, reads as venue then unit, the same as the lot form.
+    const restTokens = [...tokens.slice(numberAtToken + 1), ...detail.split(/\s+/u).filter((token) => token.length)];
+    if (restTokens.length) {
+        const restPositions = tokenPositions(text, restTokens, numberAt + number.length);
+        const firstUnit = firstUnitIndex(restTokens);
+        if (firstUnit) {
+            put(spans, text, restPositions[0], restPositions[firstUnit - 1] + restTokens[firstUnit - 1].length, "venue");
+        }
+        if (firstUnit < restTokens.length) {
+            put(spans, text, restPositions[firstUnit], restPositions.at(-1) + restTokens.at(-1).length, "unit");
+        }
+    }
+    if (parenthetical) {
+        const parenAt = text.indexOf(parenthetical, head.length);
+        const innerComma = parenthetical.indexOf(",");
+        const dong = (innerComma === -1 ? parenthetical : parenthetical.slice(0, innerComma)).trim();
+        const building = (innerComma === -1 ? "" : parenthetical.slice(innerComma + 1)).trim();
+        if (dongs.has(dong)) {
+            const dongAt = text.indexOf(dong, parenAt);
+            put(spans, text, dongAt, dongAt + dong.length, "dependent_locality");
+            if (building) {
+                const buildingAt = text.indexOf(building, dongAt + dong.length);
+                put(spans, text, buildingAt, buildingAt + building.length, "venue");
+            }
+        }
+        else {
+            put(spans, text, parenAt, parenAt + parenthetical.length, "venue");
+        }
+    }
+    return finish(text, spans, "registry_road", region, sigungu);
+}
+/**
+ * Align one lot-number string, or answer null.
+ */
+export function alignLotAddress(text, index) {
+    const tokens = text.split(/\s+/u).filter((token) => token.length);
+    const positions = tokenPositions(text, tokens);
+    if (tokens.length < MINIMUM_TOKENS || !index.regions.has(tokens[0]))
+        return null;
+    const region = tokens[0];
+    const width = sigunguSpan(index, region, tokens, 1);
+    if (!width && !index.sigunguByRegion.get(region)?.has(""))
+        return null;
+    const sigungu = tokens.slice(1, 1 + width).join(" ");
+    const dongAt = 1 + width;
+    const dongs = index.dongsByUnit.get(unitKey(region, sigungu)) ?? new Set();
+    if (dongAt >= tokens.length || !dongs.has(tokens[dongAt]))
+        return null;
+    const dong = tokens[dongAt];
+    let lotAt = dongAt + 1;
+    let ri = null;
+    if (lotAt < tokens.length && index.risByDong.get(unitKey(region, sigungu, dong))?.has(tokens[lotAt])) {
+        ri = tokens[lotAt];
+        lotAt += 1;
+    }
+    if (lotAt >= tokens.length || !LOT_NUMBER.test(tokens[lotAt]))
+        return null;
+    const lot = tokens[lotAt];
+    const spans = [];
+    put(spans, text, positions[0], positions[0] + region.length, "region");
+    for (let offset = 0; offset < width; offset += 1) {
+        const token = tokens[1 + offset];
+        put(spans, text, positions[1 + offset], positions[1 + offset] + token.length, "subregion");
+    }
+    put(spans, text, positions[dongAt], positions[dongAt] + dong.length, "dependent_locality");
+    if (ri) {
+        put(spans, text, positions[lotAt - 1], positions[lotAt - 1] + ri.length, "dependent_locality");
+    }
+    put(spans, text, positions[lotAt], positions[lotAt] + lot.length, "house_number");
+    const rest = tokens.slice(lotAt + 1);
+    if (rest.length) {
+        // The clerk writes the building name first and the floor/unit after it
+        // (`교보생명빌딩 2층`, `지강빌딩 1층 일부호`): the venue is the run of tokens before the first
+        // unit-shaped one, the unit everything from there to the end.
+        const firstUnit = firstUnitIndex(rest);
+        if (firstUnit) {
+            put(spans, text, positions[lotAt + 1], positions[lotAt + firstUnit] + rest[firstUnit - 1].length, "venue");
+        }
+        if (firstUnit < rest.length) {
+            put(spans, text, positions[lotAt + 1 + firstUnit], positions[lotAt + rest.length] + rest.at(-1).length, "unit");
+        }
+    }
+    return finish(text, spans, "registry_lot", region, sigungu);
+}
+//# sourceMappingURL=align.js.map

package/out/kr/adapters/localdata/align.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"align.js","sourceRoot":"","sources":["../../../../lib/kr/adapters/localdata/align.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,EAAiB,WAAW,EAAE,OAAO,EAAE,MAAM,kCAAkC,CAAA;AAEtF;;GAEG;AACH,MAAM,eAAe,GAAG,2BAA2B,CAAA;AAEnD;;GAEG;AACH,MAAM,UAAU,GAAG,0BAA0B,CAAA;AAE7C;;GAEG;AACH,MAAM,UAAU,GAAG,gFAAgF,CAAA;AAEnG;;;GAGG;AACH,MAAM,cAAc,GAAG,CAAC,CAAA;AAqBxB;;;;;GAKG;AACH,SAAS,GAAG,CAAC,KAAa,EAAE,IAAY,EAAE,KAAa,EAAE,GAAW,EAAE,GAAW;IAChF,IAAI,GAAG,GAAG,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC;QAClD,KAAK,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,GAAG,EAAE,CAAC,CAAA;IAChC,CAAC;AACF,CAAC;AAED;;;GAGG;AACH,SAAS,cAAc,CAAC,IAAY,EAAE,MAAyB,EAAE,IAAI,GAAG,CAAC;IACxE,MAAM,SAAS,GAAa,EAAE,CAAA;IAC9B,IAAI,MAAM,GAAG,IAAI,CAAA;IAEjB,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC5B,MAAM,EAAE,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,MAAM,CAAC,CAAA;QAEtC,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;QAClB,MAAM,GAAG,EAAE,GAAG,KAAK,CAAC,MAAM,CAAA;IAC3B,CAAC;IAED,OAAO,SAAS,CAAA;AACjB,CAAC;AAED;;GAEG;AACH,SAAS,cAAc,CAAC,MAAyB;IAChD,MAAM,KAAK,GAAG,MAAM,CAAC,SAAS,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAA;IAEjE,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAA;AAC5C,CAAC;AAED,SAAS,MAAM,CAAC,IAAY,EAAE,KAAa,EAAE,QAA6B,EAAE,MAAc,EAAE,OAAe;IAC1G,MAAM,MAAM,GAAG,KAAK,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;IAEzG,OAAO;QACN,GAAG,EAAE,IAAI;QACT,UAAU,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC;QAC5C,QAAQ,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC;QACxC,QAAQ,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC;QACxC,QAAQ;QACR,MAAM;QACN,OAAO;KACP,CAAA;AACF,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,gBAAgB,CAAC,IAAY,EAAE,KAAe;IAC7D,MAAM,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAA;IAChC,MAAM,IAAI,GAAG,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,MAAM,CAAC,CAAA;IACzD,MAAM,IAAI,GAAG,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;IACxD,MAAM,aAAa,GAAG,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAA;IAExE,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAA;IACjC,MAAM,IAAI,GAAG,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,CAAA;IAC3D,MAAM,MAAM,GAAG,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,GAAG,CAAC,CAAC,CAAA;IAE5D,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,CAAA;IACjE,MAAM,SAAS,GAAG,cAAc,CAAC,IAAI,EAAE,MAAM,CAAC,CAAA;IAE9C,IAAI,MAAM,CAAC,MAAM,GAAG,cAAc,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC;QAAE,OAAO,IAAI,CAAA;IAEjF,MAAM,MAAM,GAAG,MAAM,CAAC,CAAC,CAAE,CAAA;IACzB,MAAM,KAAK,GAAG,WAAW,CAAC,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC,CAAC,CAAA;IAEnD,gGAAgG;IAChG,IAAI,CAAC,KAAK,IAAI,CAAC,KAAK,CAAC,eAAe,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,IAAI,CAAA;IAEtE,MAAM,OAAO,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;IACpD,MAAM,IAAI,GAAG,OAAO,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;IACrC,MAAM,KAAK,GAAG,KAAK,CAAC,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,IAAI,GAAG,EAAU,CAAA;IAC9D,MAAM,KAAK,GAAG,KAAK,CAAC,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,IAAI,GAAG,EAAU,CAAA;IAE9D,IAAI,MAAM,GAAG,CAAC,GAAG,KAAK,CAAA;IACtB,IAAI,UAAU,GAAkB,IAAI,CAAA;IAEpC,6EAA6E;IAC7E,uFAAuF;IACvF,IAAI,MAAM,GAAG,CAAC,GAAG,MAAM,CAAC,MAAM,IAAI,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,CAAE,CAAC,IAAI,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAE,CAAC,EAAE,CAAC;QAChG,UAAU,GAAG,MAAM,CAAA;QACnB,MAAM,IAAI,CAAC,CAAA;IACZ,CAAC;IAED,4BAA4B;IAC5B,qEAAqE;IACrE,IAAI,MAAM,GAAG,CAAC,IAAI,MAAM,CAAC,MAAM,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,CAAE,CAAC;QAAE,OAAO,IAAI,CAAA;IAE3E,IAAI,aAAa,GAAG,MAAM,GAAG,CAAC,CAAA;IAE9B,yGAAyG;IACzG,IAAI,MAAM,CAAC,aAAa,CAAC,KAAK,IAAI,IAAI,aAAa,GAAG,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC;QACzE,aAAa,IAAI,CAAC,CAAA;IACnB,CAAC;IAED,MAAM,MAAM,GAAG,MAAM,CAAC,aAAa,CAAE,CAAA;IAErC,IAAI,CAAC,eAAe,CAAC,IAAI,CAAC,MAAM,CAAC;QAAE,OAAO,IAAI,CAAA;IAE9C,MAAM,KAAK,GAAW,EAAE,CAAA;IAExB,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,SAAS,CAAC,CAAC,CAAE,EAAE,SAAS,CAAC,CAAC,CAAE,GAAG,MAAM,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAA;IAExE,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,KAAK,EAAE,MAAM,IAAI,CAAC,EAAE,CAAC;QAClD,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,GAAG,MAAM,CAAE,CAAA;QAEjC,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,SAAS,CAAC,CAAC,GAAG,MAAM,CAAE,EAAE,SAAS,CAAC,CAAC,GAAG,MAAM,CAAE,GAAG,KAAK,CAAC,MAAM,EAAE,WAAW,CAAC,CAAA;IAC7F,CAAC;IAED,IAAI,UAAU,KAAK,IAAI,EAAE,CAAC;QACzB,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,SAAS,CAAC,UAAU,CAAE,EAAE,SAAS,CAAC,UAAU,CAAE,GAAG,MAAM,CAAC,UAAU,CAAE,CAAC,MAAM,EAAE,oBAAoB,CAAC,CAAA;IACpH,CAAC;IAED,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,SAAS,CAAC,MAAM,CAAE,EAAE,SAAS,CAAC,MAAM,CAAE,GAAG,MAAM,CAAC,MAAM,CAAE,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAA;IAE3F,MAAM,QAAQ,GAAG,SAAS,CAAC,aAAa,CAAE,CAAA;IAE1C,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,QAAQ,EAAE,QAAQ,GAAG,MAAM,CAAC,MAAM,EAAE,cAAc,CAAC,CAAA;IAEpE,wGAAwG;IACxG,MAAM,UAAU,GAAG,CAAC,GAAG,MAAM,CAAC,KAAK,CAAC,aAAa,GAAG,CAAC,CAAC,EAAE,GAAG,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAA;IAEhH,IAAI,UAAU,CAAC,MAAM,EAAE,CAAC;QACvB,MAAM,aAAa,GAAG,cAAc,CAAC,IAAI,EAAE,UAAU,EAAE,QAAQ,GAAG,MAAM,CAAC,MAAM,CAAC,CAAA;QAChF,MAAM,SAAS,GAAG,cAAc,CAAC,UAAU,CAAC,CAAA;QAE5C,IAAI,SAAS,EAAE,CAAC;YACf,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,aAAa,CAAC,CAAC,CAAE,EAAE,aAAa,CAAC,SAAS,GAAG,CAAC,CAAE,GAAG,UAAU,CAAC,SAAS,GAAG,CAAC,CAAE,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;QAChH,CAAC;QAED,IAAI,SAAS,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC;YACnC,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,aAAa,CAAC,SAAS,CAAE,EAAE,aAAa,CAAC,EAAE,CAAC,CAAC,CAAC,CAAE,GAAG,UAAU,CAAC,EAAE,CAAC,CAAC,CAAC,CAAE,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;QACvG,CAAC;IACF,CAAC;IAED,IAAI,aAAa,EAAE,CAAC;QACnB,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,aAAa,EAAE,IAAI,CAAC,MAAM,CAAC,CAAA;QACxD,MAAM,UAAU,GAAG,aAAa,CAAC,OAAO,CAAC,GAAG,CAAC,CAAA;QAC7C,MAAM,IAAI,GAAG,CAAC,UAAU,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,aAAa,CAAC,KAAK,CAAC,CAAC,EAAE,UAAU,CAAC,CAAC,CAAC,IAAI,EAAE,CAAA;QAC5F,MAAM,QAAQ,GAAG,CAAC,UAAU,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,aAAa,CAAC,KAAK,CAAC,UAAU,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAA;QAEtF,IAAI,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;YACrB,MAAM,MAAM,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,OAAO,CAAC,CAAA;YAE1C,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC,MAAM,EAAE,oBAAoB,CAAC,CAAA;YAEpE,IAAI,QAAQ,EAAE,CAAC;gBACd,MAAM,UAAU,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC,CAAA;gBAE/D,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,UAAU,EAAE,UAAU,GAAG,QAAQ,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;YACpE,CAAC;QACF,CAAC;aAAM,CAAC;YACP,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,GAAG,aAAa,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;QACnE,CAAC;IACF,CAAC;IAED,OAAO,MAAM,CAAC,IAAI,EAAE,KAAK,EAAE,eAAe,EAAE,MAAM,EAAE,OAAO,CAAC,CAAA;AAC7D,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,eAAe,CAAC,IAAY,EAAE,KAAe;IAC5D,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,CAAA;IACjE,MAAM,SAAS,GAAG,cAAc,CAAC,IAAI,EAAE,MAAM,CAAC,CAAA;IAE9C,IAAI,MAAM,CAAC,MAAM,GAAG,cAAc,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAE,CAAC;QAAE,OAAO,IAAI,CAAA;IAEjF,MAAM,MAAM,GAAG,MAAM,CAAC,CAAC,CAAE,CAAA;IACzB,MAAM,KAAK,GAAG,WAAW,CAAC,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC,CAAC,CAAA;IAEnD,IAAI,CAAC,KAAK,IAAI,CAAC,KAAK,CAAC,eAAe,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC;QAAE,OAAO,IAAI,CAAA;IAEtE,MAAM,OAAO,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;IACpD,MAAM,MAAM,GAAG,CAAC,GAAG,KAAK,CAAA;IACxB,MAAM,KAAK,GAAG,KAAK,CAAC,WAAW,CAAC,GAAG,CAAC,OAAO,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,IAAI,IAAI,GAAG,EAAU,CAAA;IAElF,IAAI,MAAM,IAAI,MAAM,CAAC,MAAM,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,CAAE,CAAC;QAAE,OAAO,IAAI,CAAA;IAEvE,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAE,CAAA;IAC5B,IAAI,KAAK,GAAG,MAAM,GAAG,CAAC,CAAA;IACtB,IAAI,EAAE,GAAkB,IAAI,CAAA;IAE5B,IAAI,KAAK,GAAG,MAAM,CAAC,MAAM,IAAI,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,OAAO,CAAC,MAAM,EAAE,OAAO,EAAE,IAAI,CAAC,CAAC,EAAE,GAAG,CAAC,MAAM,CAAC,KAAK,CAAE,CAAC,EAAE,CAAC;QACvG,EAAE,GAAG,MAAM,CAAC,KAAK,CAAE,CAAA;QACnB,KAAK,IAAI,CAAC,CAAA;IACX,CAAC;IAED,IAAI,KAAK,IAAI,MAAM,CAAC,MAAM,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAE,CAAC;QAAE,OAAO,IAAI,CAAA;IAE3E,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAE,CAAA;IAC1B,MAAM,KAAK,GAAW,EAAE,CAAA;IAExB,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,SAAS,CAAC,CAAC,CAAE,EAAE,SAAS,CAAC,CAAC,CAAE,GAAG,MAAM,CAAC,MAAM,EAAE,QAAQ,CAAC,CAAA;IAExE,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,KAAK,EAAE,MAAM,IAAI,CAAC,EAAE,CAAC;QAClD,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,GAAG,MAAM,CAAE,CAAA;QAEjC,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,SAAS,CAAC,CAAC,GAAG,MAAM,CAAE,EAAE,SAAS,CAAC,CAAC,GAAG,MAAM,CAAE,GAAG,KAAK,CAAC,MAAM,EAAE,WAAW,CAAC,CAAA;IAC7F,CAAC;IAED,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,SAAS,CAAC,MAAM,CAAE,EAAE,SAAS,CAAC,MAAM,CAAE,GAAG,IAAI,CAAC,MAAM,EAAE,oBAAoB,CAAC,CAAA;IAE5F,IAAI,EAAE,EAAE,CAAC;QACR,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,SAAS,CAAC,KAAK,GAAG,CAAC,CAAE,EAAE,SAAS,CAAC,KAAK,GAAG,CAAC,CAAE,GAAG,EAAE,CAAC,MAAM,EAAE,oBAAoB,CAAC,CAAA;IACjG,CAAC;IAED,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,SAAS,CAAC,KAAK,CAAE,EAAE,SAAS,CAAC,KAAK,CAAE,GAAG,GAAG,CAAC,MAAM,EAAE,cAAc,CAAC,CAAA;IAEnF,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAA;IAEpC,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;QACjB,uEAAuE;QACvE,gFAAgF;QAChF,8DAA8D;QAC9D,MAAM,SAAS,GAAG,cAAc,CAAC,IAAI,CAAC,CAAA;QAEtC,IAAI,SAAS,EAAE,CAAC;YACf,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,SAAS,CAAC,KAAK,GAAG,CAAC,CAAE,EAAE,SAAS,CAAC,KAAK,GAAG,SAAS,CAAE,GAAG,IAAI,CAAC,SAAS,GAAG,CAAC,CAAE,CAAC,MAAM,EAAE,OAAO,CAAC,CAAA;QAC9G,CAAC;QAED,IAAI,SAAS,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;YAC7B,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,SAAS,CAAC,KAAK,GAAG,CAAC,GAAG,SAAS,CAAE,EAAE,SAAS,CAAC,KAAK,GAAG,IAAI,CAAC,MAAM,CAAE,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAE,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;QACnH,CAAC;IACF,CAAC;IAED,OAAO,MAAM,CAAC,IAAI,EAAE,KAAK,EAAE,cAAc,EAAE,MAAM,EAAE,OAAO,CAAC,CAAA;AAC5D,CAAC"}