text-sanctifier 1.3.2 → 1.3.3

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "text-sanctifier",
-  "version": "1.3.2",
+  "version": "1.3.3",
   "type": "module",
   "description": "A brutal text normalizer and invisible trash scrubber for modern web projects.",
   "engines": {
@@ -27,6 +27,10 @@
       "types": "./dist/types/index.d.ts",
       "default": "./src/index.js"
     },
+    "./hidden": {
+      "types": "./dist/types/hiddenText.d.ts",
+      "default": "./src/hiddenText.js"
+    },
     "./browser": {
       "types": "./dist/types/index.d.ts",
       "default": "./dist/text-sanctifier.min.js"

package/src/categories.generated.js

@@ -1,223 +1,223 @@
-// src/categories.generated.js
-//
-// Generated by scripts/gen-categories.js from the table in scripts/categories.js.
-// Do not edit: change the table and run `npm run gen`. `npm run ci` fails while this file is stale.
-
-/** @typedef {import('./inspectText.js').TrashCategory} TrashCategory */
-
-/**
- * Removed by `nukeControls`. Categories: bidi_control, nul, c0_control, c1_control.
- * Global: reset `lastIndex` before calling `.test()`.
- * @type {RegExp}
- */
-export const CONTROL_CHARS_REGEX = /[\u0000-\u0008\u000B\u000C\u000E-\u001F\u007F-\u009F\u061C\u200E\u200F\u202A-\u202E\u2066-\u2069]+/g;
-
-/**
- * Removed by `purgeInvisibleChars`. Categories: bidi_control, zero_width, nonbreaking_space, bom, invisible_format, hangul_filler, variation_selector, tag.
- * Global: reset `lastIndex` before calling `.test()`.
- * @type {RegExp}
- */
-export const INVISIBLE_TRASH_REGEX = /[\u00A0\u00AD\u034F\u061C\u115F\u1160\u17B4\u17B5\u180B-\u180F\u2000-\u200F\u202A-\u202F\u205F-\u206F\u3000\u3164\uFE00-\uFE0F\uFEFF\uFFA0\uFFF0-\uFFF8\u{1BCA0}-\u{1BCA3}\u{1D173}-\u{1D17A}\u{E0000}-\u{E0FFF}]+/gu;
-
-/**
- * Category of one BMP code point (one UTF-16 code unit), or null when it is not trash.
- * Surrogates return null; the caller pairs them and uses classifyAstral.
- * @param {number} c
- * @returns {TrashCategory|null}
- */
-export function classifyCodeUnit(c) {
-    return c < 0x2000 ? classifyBelow2000(c) : classifyFrom2000(c);
-}
-
-// Two halves rather than one tree: V8 inlines a function only while its bytecode is
-// small (460 bytes by default), and the single pass is much faster when it does.
-
-/** @param {number} c @returns {TrashCategory|null} */
-function classifyBelow2000(c) {
-    if (c < 0x00AE) {
-        if (c < 0x0020) {
-            if (c < 0x000B) {
-                if (c < 0x0001) {
-                    return 'nul';
-                }
-                if (c < 0x0009) {
-                    return 'c0_control';
-                }
-                return null;
-            }
-            if (c < 0x000D) {
-                return 'c0_control';
-            }
-            if (c < 0x000E) {
-                return null;
-            }
-            return 'c0_control';
-        }
-        if (c < 0x00A0) {
-            if (c < 0x007F) {
-                return null;
-            }
-            if (c < 0x0080) {
-                return 'c0_control';
-            }
-            return 'c1_control';
-        }
-        if (c < 0x00A1) {
-            return 'nonbreaking_space';
-        }
-        if (c < 0x00AD) {
-            return null;
-        }
-        return 'invisible_format';
-    }
-    if (c < 0x1161) {
-        if (c < 0x061C) {
-            if (c < 0x034F) {
-                return null;
-            }
-            if (c < 0x0350) {
-                return 'invisible_format';
-            }
-            return null;
-        }
-        if (c < 0x061D) {
-            return 'bidi_control';
-        }
-        if (c < 0x115F) {
-            return null;
-        }
-        return 'hangul_filler';
-    }
-    if (c < 0x180B) {
-        if (c < 0x17B4) {
-            return null;
-        }
-        if (c < 0x17B6) {
-            return 'invisible_format';
-        }
-        return null;
-    }
-    if (c < 0x180F) {
-        if (c < 0x180E) {
-            return 'variation_selector';
-        }
-        return 'invisible_format';
-    }
-    if (c < 0x1810) {
-        return 'variation_selector';
-    }
-    return null;
-}
-
-/** @param {number} c @returns {TrashCategory|null} */
-function classifyFrom2000(c) {
-    if (c < 0x2070) {
-        if (c < 0x2030) {
-            if (c < 0x2010) {
-                if (c < 0x200B) {
-                    return 'nonbreaking_space';
-                }
-                if (c < 0x200E) {
-                    return 'zero_width';
-                }
-                return 'bidi_control';
-            }
-            if (c < 0x202A) {
-                return null;
-            }
-            if (c < 0x202F) {
-                return 'bidi_control';
-            }
-            return 'nonbreaking_space';
-        }
-        if (c < 0x2061) {
-            if (c < 0x205F) {
-                return null;
-            }
-            if (c < 0x2060) {
-                return 'nonbreaking_space';
-            }
-            return 'zero_width';
-        }
-        if (c < 0x2066) {
-            return 'invisible_format';
-        }
-        if (c < 0x206A) {
-            return 'bidi_control';
-        }
-        return 'invisible_format';
-    }
-    if (c < 0xFE10) {
-        if (c < 0x3164) {
-            if (c < 0x3000) {
-                return null;
-            }
-            if (c < 0x3001) {
-                return 'nonbreaking_space';
-            }
-            return null;
-        }
-        if (c < 0x3165) {
-            return 'hangul_filler';
-        }
-        if (c < 0xFE00) {
-            return null;
-        }
-        return 'variation_selector';
-    }
-    if (c < 0xFFA0) {
-        if (c < 0xFEFF) {
-            return null;
-        }
-        if (c < 0xFF00) {
-            return 'bom';
-        }
-        return null;
-    }
-    if (c < 0xFFF0) {
-        if (c < 0xFFA1) {
-            return 'hangul_filler';
-        }
-        return null;
-    }
-    if (c < 0xFFF9) {
-        return 'invisible_format';
-    }
-    return null;
-}
-
-/**
- * Category of a code point above U+FFFF, or null when it is not trash.
- * @param {number} c
- * @returns {TrashCategory|null}
- */
-export function classifyAstral(c) {
-    if (c < 0xE0000) {
-        if (c < 0x1BCA4) {
-            if (c < 0x1BCA0) {
-                return null;
-            }
-            return 'invisible_format';
-        }
-        if (c < 0x1D173) {
-            return null;
-        }
-        if (c < 0x1D17B) {
-            return 'invisible_format';
-        }
-        return null;
-    }
-    if (c < 0xE0100) {
-        if (c < 0xE0080) {
-            return 'tag';
-        }
-        return 'invisible_format';
-    }
-    if (c < 0xE01F0) {
-        return 'variation_selector';
-    }
-    if (c < 0xE1000) {
-        return 'invisible_format';
-    }
-    return null;
-}
+// src/categories.generated.js
+//
+// Generated by scripts/gen-categories.js from the table in scripts/categories.js.
+// Do not edit: change the table and run `npm run gen`. `npm run ci` fails while this file is stale.
+
+/** @typedef {import('./inspectText.js').TrashCategory} TrashCategory */
+
+/**
+ * Removed by `nukeControls`. Categories: bidi_control, nul, c0_control, c1_control.
+ * Global: reset `lastIndex` before calling `.test()`.
+ * @type {RegExp}
+ */
+export const CONTROL_CHARS_REGEX = /[\u0000-\u0008\u000B\u000C\u000E-\u001F\u007F-\u009F\u061C\u200E\u200F\u202A-\u202E\u2066-\u2069]+/g;
+
+/**
+ * Removed by `purgeInvisibleChars`. Categories: bidi_control, zero_width, nonbreaking_space, bom, invisible_format, hangul_filler, variation_selector, tag.
+ * Global: reset `lastIndex` before calling `.test()`.
+ * @type {RegExp}
+ */
+export const INVISIBLE_TRASH_REGEX = /[\u00A0\u00AD\u034F\u061C\u115F\u1160\u17B4\u17B5\u180B-\u180F\u2000-\u200F\u202A-\u202F\u205F-\u206F\u3000\u3164\uFE00-\uFE0F\uFEFF\uFFA0\uFFF0-\uFFF8\u{1BCA0}-\u{1BCA3}\u{1D173}-\u{1D17A}\u{E0000}-\u{E0FFF}]+/gu;
+
+/**
+ * Category of one BMP code point (one UTF-16 code unit), or null when it is not trash.
+ * Surrogates return null; the caller pairs them and uses classifyAstral.
+ * @param {number} c
+ * @returns {TrashCategory|null}
+ */
+export function classifyCodeUnit(c) {
+    return c < 0x2000 ? classifyBelow2000(c) : classifyFrom2000(c);
+}
+
+// Two halves rather than one tree: V8 inlines a function only while its bytecode is
+// small (460 bytes by default), and the single pass is much faster when it does.
+
+/** @param {number} c @returns {TrashCategory|null} */
+function classifyBelow2000(c) {
+    if (c < 0x00AE) {
+        if (c < 0x0020) {
+            if (c < 0x000B) {
+                if (c < 0x0001) {
+                    return 'nul';
+                }
+                if (c < 0x0009) {
+                    return 'c0_control';
+                }
+                return null;
+            }
+            if (c < 0x000D) {
+                return 'c0_control';
+            }
+            if (c < 0x000E) {
+                return null;
+            }
+            return 'c0_control';
+        }
+        if (c < 0x00A0) {
+            if (c < 0x007F) {
+                return null;
+            }
+            if (c < 0x0080) {
+                return 'c0_control';
+            }
+            return 'c1_control';
+        }
+        if (c < 0x00A1) {
+            return 'nonbreaking_space';
+        }
+        if (c < 0x00AD) {
+            return null;
+        }
+        return 'invisible_format';
+    }
+    if (c < 0x1161) {
+        if (c < 0x061C) {
+            if (c < 0x034F) {
+                return null;
+            }
+            if (c < 0x0350) {
+                return 'invisible_format';
+            }
+            return null;
+        }
+        if (c < 0x061D) {
+            return 'bidi_control';
+        }
+        if (c < 0x115F) {
+            return null;
+        }
+        return 'hangul_filler';
+    }
+    if (c < 0x180B) {
+        if (c < 0x17B4) {
+            return null;
+        }
+        if (c < 0x17B6) {
+            return 'invisible_format';
+        }
+        return null;
+    }
+    if (c < 0x180F) {
+        if (c < 0x180E) {
+            return 'variation_selector';
+        }
+        return 'invisible_format';
+    }
+    if (c < 0x1810) {
+        return 'variation_selector';
+    }
+    return null;
+}
+
+/** @param {number} c @returns {TrashCategory|null} */
+function classifyFrom2000(c) {
+    if (c < 0x2070) {
+        if (c < 0x2030) {
+            if (c < 0x2010) {
+                if (c < 0x200B) {
+                    return 'nonbreaking_space';
+                }
+                if (c < 0x200E) {
+                    return 'zero_width';
+                }
+                return 'bidi_control';
+            }
+            if (c < 0x202A) {
+                return null;
+            }
+            if (c < 0x202F) {
+                return 'bidi_control';
+            }
+            return 'nonbreaking_space';
+        }
+        if (c < 0x2061) {
+            if (c < 0x205F) {
+                return null;
+            }
+            if (c < 0x2060) {
+                return 'nonbreaking_space';
+            }
+            return 'zero_width';
+        }
+        if (c < 0x2066) {
+            return 'invisible_format';
+        }
+        if (c < 0x206A) {
+            return 'bidi_control';
+        }
+        return 'invisible_format';
+    }
+    if (c < 0xFE10) {
+        if (c < 0x3164) {
+            if (c < 0x3000) {
+                return null;
+            }
+            if (c < 0x3001) {
+                return 'nonbreaking_space';
+            }
+            return null;
+        }
+        if (c < 0x3165) {
+            return 'hangul_filler';
+        }
+        if (c < 0xFE00) {
+            return null;
+        }
+        return 'variation_selector';
+    }
+    if (c < 0xFFA0) {
+        if (c < 0xFEFF) {
+            return null;
+        }
+        if (c < 0xFF00) {
+            return 'bom';
+        }
+        return null;
+    }
+    if (c < 0xFFF0) {
+        if (c < 0xFFA1) {
+            return 'hangul_filler';
+        }
+        return null;
+    }
+    if (c < 0xFFF9) {
+        return 'invisible_format';
+    }
+    return null;
+}
+
+/**
+ * Category of a code point above U+FFFF, or null when it is not trash.
+ * @param {number} c
+ * @returns {TrashCategory|null}
+ */
+export function classifyAstral(c) {
+    if (c < 0xE0000) {
+        if (c < 0x1BCA4) {
+            if (c < 0x1BCA0) {
+                return null;
+            }
+            return 'invisible_format';
+        }
+        if (c < 0x1D173) {
+            return null;
+        }
+        if (c < 0x1D17B) {
+            return 'invisible_format';
+        }
+        return null;
+    }
+    if (c < 0xE0100) {
+        if (c < 0xE0080) {
+            return 'tag';
+        }
+        return 'invisible_format';
+    }
+    if (c < 0xE01F0) {
+        return 'variation_selector';
+    }
+    if (c < 0xE1000) {
+        return 'invisible_format';
+    }
+    return null;
+}