@tkeron/html-parser 1.1.2 → 1.3.0

package/src/encoding/constants.ts

@@ -0,0 +1,8 @@
+export const encodingAliases: Record<string, string> = {
+  "iso-8859-1": "windows-1252",
+  "iso8859-1": "windows-1252",
+  "iso-8859-2": "iso-8859-2",
+  "iso8859-2": "iso-8859-2",
+  "utf-8": "utf-8",
+  utf8: "utf-8",
+};

package/src/encoding/detect-encoding.ts

@@ -0,0 +1,21 @@
+import { normalizeEncoding } from "./normalize-encoding.js";
+
+export const detectEncoding = (html: string): string | null => {
+  const prefix = html.substring(0, 1024);
+
+  const charsetMatch = prefix.match(
+    /<meta[^>]*charset\s*=\s*["']?([^"'\s>]+)["']?/i,
+  );
+  if (charsetMatch) {
+    return normalizeEncoding(charsetMatch[1]);
+  }
+
+  const contentTypeMatch = prefix.match(
+    /<meta[^>]*http-equiv\s*=\s*["']?\s*content-type\s*["']?[^>]*content\s*=\s*["']?\s*text\/html;\s*charset\s*=\s*([^"'\s>]+)["']?/i,
+  );
+  if (contentTypeMatch) {
+    return normalizeEncoding(contentTypeMatch[1]);
+  }
+
+  return "windows-1252";
+};

package/src/encoding/index.ts

@@ -0,0 +1 @@
+export { detectEncoding } from "./detect-encoding.ts";

package/src/encoding/normalize-encoding.ts

@@ -0,0 +1,6 @@
+import { encodingAliases } from "./constants.js";
+
+export const normalizeEncoding = (name: string): string | null => {
+  const lower = name.toLowerCase().replace(/[^a-z0-9-]/g, "");
+  return encodingAliases[lower] || lower;
+};