fix(encoding): ISO-2022-JPのページで文字化けが発生する問題を修正

package.json

@@ -27,6 +27,7 @@
     "@swc/jest": "^0.2.37",
     "@types/cheerio": "0.22.35",
     "@types/debug": "4.1.12",
+    "@types/encoding-japanese": "^2.2.1",
     "@types/escape-regexp": "^0.0.3",
     "@types/node": "22.9.0",
     "@typescript-eslint/eslint-plugin": "^7.17.0",
@@ -42,6 +43,7 @@
   },
   "dependencies": {
     "cheerio": "1.0.0",
+    "encoding-japanese": "^2.2.0",
     "escape-regexp": "0.0.1",
     "got": "^14.4.4",
     "html-entities": "2.5.2",

src/utils/encoding.ts

@@ -1,14 +1,47 @@
 import iconv from 'iconv-lite';
+import Encoding from 'encoding-japanese';
 import jschardet from 'jschardet';
 
+import type { Response } from 'got';
+
 const regCharset = new RegExp(/charset\s*=\s*["']?([\w-]+)/, 'i');
 
+export const ENCODING_JAPANESE_ENCODING_PREFIX = '__EJ__';
+
+const ENCODING_JAPANESE_SUPPORTED_ENCODING: string[] = [
+	'UTF32',
+	'UTF16',
+	'UTF16BE',
+	'UTF16LE',
+	'BINARY',
+	'ASCII',
+	'JIS',
+	'UTF8',
+	'EUCJP',
+	'SJIS',
+	'UNICODE',
+	'AUTO',
+] satisfies Encoding.Encoding[];
+
 /**
  * Detect HTML encoding
  * @param body Body in Buffer
  * @returns encoding
  */
-export function detectEncoding(body: Buffer): string {
+export function detectEncoding(res: Response): string {
+	// From header
+	const contentType = res.headers['content-type'];
+	if (contentType) {
+		const matchHeader = contentType.match(regCharset);
+		if (matchHeader) {
+			const candicate = matchHeader[1];
+			const encoding = toEncoding(candicate);
+			if (encoding != null) return encoding;
+		}
+	}
+
+	const body = res.rawBody;
+
 	// By detection
 	const detected = jschardet.detect(body, { minimumThreshold: 0.99 });
 	// eslint-disable-next-line @typescript-eslint/no-unnecessary-condition
@@ -30,10 +63,25 @@ export function detectEncoding(body: Buffer): string {
 }
 
 export function toUtf8(body: Buffer, encoding: string): string {
+	if (encoding.startsWith(ENCODING_JAPANESE_ENCODING_PREFIX)) {
+		const _encoding = encoding.slice(ENCODING_JAPANESE_ENCODING_PREFIX.length);
+
+		function assertEncoding(enc: string): enc is Encoding.Encoding {
+			return ENCODING_JAPANESE_SUPPORTED_ENCODING.includes(enc);
+		}
+
+		if (assertEncoding(_encoding)) {
+			return Encoding.codeToString(Encoding.convert(body, 'UNICODE', _encoding));
+		}
+	}
 	return iconv.decode(body, encoding);
 }
 
 function toEncoding(candicate: string): string | null {
+	// iconvで処理できない
+	// https://github.com/ashtuchkin/iconv-lite/issues/60
+	if (candicate.toLowerCase() === 'iso-2022-jp') return '__EJ__JIS';
+
 	if (iconv.encodingExists(candicate)) {
 		if (['shift_jis', 'shift-jis', 'windows-31j', 'x-sjis'].includes(candicate.toLowerCase())) return 'cp932';
 		return candicate;

src/utils/got.ts

@@ -77,7 +77,7 @@ export async function scpaping(
 		method: 'GET',
 	});
 
-	const encoding = detectEncoding(response.rawBody);
+	const encoding = detectEncoding(response);
 	const body = toUtf8(response.rawBody, encoding);
 	const $ = cheerio.load(body);