dwml-ts 0.5.1 → 0.5.3

package/README.md

@@ -117,12 +117,20 @@ astral-plane Mathematical Alphanumeric Symbols are handled correctly):
    `U+1D400–U+1D7FF` → ASCII (`𝐴→A`, `𝑥→x`), Greek (`π→\pi`, both plain `U+03B1..` and
    math-italic `U+1D6FC..`), relation/operator symbols (`≤→\leq`, `≥→\geq`, `≠→\ne`,
    `·→\cdot`, `×→\times`, `∞→\infty`, `→→\rightarrow`, …).
-2. **pylatexenc's builtin map** (`src/uni2latex.generated.ts`, internalized from
-   pylatexenc 2.x, MIT) — everything else with a known LaTeX escape.
+2. **Verbatim Unicode** — everything else is kept as-is (`světlo` → `světlo`,
+   `€` → `€`). OMML is Unicode-native and math renderers (MathLive, KaTeX,
+   unicode-math) accept Unicode identifiers/symbols in math mode, whereas
+   text-mode escapes (`\v{e}`, `\ss`, `\textendash`, `\ensuremath{…}`) are invalid
+   inside an equation and lossy on the way back to OMML. A literal `\` becomes
+   `{\backslash}`.
 
 LaTeX-special characters in runs (`%`, `&`, `_`, `{`, `}`, `#`, `$`, `~`, `^`) are escaped
 via dwml's `escape_latex` (`a%` → `a\%`).
 
+`unicodeToLatex()` (the pylatexenc-derived *text-mode* encoder,
+`src/uni2latex.generated.ts`) is still exported for consumers that need it, but is
+no longer used for run text.
+
 ## Examples
 
 | OMML input | LaTeX output |

package/dist/index.cjs

@@ -35,9 +35,11 @@ __webpack_require__.d(__webpack_exports__, {
     ommlNodeToLatex: ()=>ommlNodeToLatex,
     ommlToLatex: ()=>ommlToLatex,
     textContent: ()=>textContent,
-    unicodeToLatex: ()=>unicodeToLatex
+    unicodeToLatex: ()=>unicodeToLatex,
+    unicodeToMathLatex: ()=>unicodeToMathLatex
 });
 const txml_namespaceObject = require("txml/txml");
+const external_entities_namespaceObject = require("entities");
 const UNI2LATEX = {
     34: "''",
     35: "\\#",
@@ -1562,6 +1564,18 @@ function unicodeToLatex(input) {
     }
     return out;
 }
+const ASCII_MATH = {
+    0x5c: '{\\backslash}'
+};
+function unicodeToMathLatex(input) {
+    const s = String(input).normalize('NFC');
+    let out = '';
+    for (const ch of s){
+        const cp = ch.codePointAt(0);
+        out += ASCII_MATH[cp] ?? ch;
+    }
+    return out;
+}
 const CHARS = [
     '{',
     '}',
@@ -1731,6 +1745,7 @@ const T = {
     '∗': '\\ast ',
     '∘': '\\circ ',
     '∙': '\\bullet ',
+    '•': '\\bullet ',
     '−': '-',
     '<': '<',
     '>': '>',
@@ -1985,7 +2000,7 @@ function isElement(node) {
 }
 function textContent(node) {
     let out = '';
-    for (const child of node.children)out += 'string' == typeof child ? child : textContent(child);
+    for (const child of node.children)out += 'string' == typeof child ? (0, external_entities_namespaceObject.decodeXML)(child) : textContent(child);
     return out;
 }
 function pyFormat(template, positional = [], named = {}) {
@@ -2399,7 +2414,7 @@ class Converter {
             const mapped = T[ch];
             if (void 0 !== mapped) out += mapped;
             else if (CHARS.includes(ch)) out += ch;
-            else out += unicodeToLatex(ch);
+            else out += unicodeToMathLatex(ch);
         }
         return out;
     }
@@ -2490,6 +2505,7 @@ exports.ommlNodeToLatex = __webpack_exports__.ommlNodeToLatex;
 exports.ommlToLatex = __webpack_exports__.ommlToLatex;
 exports.textContent = __webpack_exports__.textContent;
 exports.unicodeToLatex = __webpack_exports__.unicodeToLatex;
+exports.unicodeToMathLatex = __webpack_exports__.unicodeToMathLatex;
 for(var __rspack_i in __webpack_exports__)if (-1 === [
     "elementChildren",
     "escapeLatex",
@@ -2498,7 +2514,8 @@ for(var __rspack_i in __webpack_exports__)if (-1 === [
     "ommlNodeToLatex",
     "ommlToLatex",
     "textContent",
-    "unicodeToLatex"
+    "unicodeToLatex",
+    "unicodeToMathLatex"
 ].indexOf(__rspack_i)) exports[__rspack_i] = __webpack_exports__[__rspack_i];
 Object.defineProperty(exports, '__esModule', {
     value: true