@ozgurguler/trnorm 0.1.2 → 0.1.3

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ozgurguler/trnorm",
-  "version": "0.1.2",
+  "version": "0.1.3",
   "description": "Turkish text normalization for speech, in JavaScript: written -> spoken form for TTS, evaluation normalization for ASR (WER/CER), and spoken -> digits. The same output as the Python trnorm.",
   "type": "module",
   "exports": {

package/src/lexicon.js

@@ -271,6 +271,13 @@ export const CAPS_WORDS = new Set([
   'YIL', 'GÜN', 'İŞ', 'EV', 'EL', 'AN', 'ÖZ',
 ]);
 
+// Titles that are also written in capitals or without the spaces ("DR. ÖĞR. ÜYESİ", "DR.ÖĞR.ÜYESİ", "Prof.Dr."):
+// those forms are read the same (normalizer.js _ABBR_FORMS). Only titles: "AV." or "SOK." in capitals could be other words.
+export const TITLE_ABBREVIATIONS = [
+  'Prof. Dr.', 'Doç. Dr.', 'Yrd. Doç. Dr.', 'Dr. Öğr. Üyesi', 'Öğr. Gör.', 'Arş. Gör.', 'Uzm. Dr.', 'Op. Dr.',
+  'Yrd. Doç.', 'Prof.', 'Doç.', 'Dr.', 'Uzm.', 'Müh.',
+];
+
 // Dotted / colon abbreviations (case-sensitive). Longest match wins.
 export const ABBREVIATIONS = dict({
   'Prof. Dr.': 'profesör doktor',

package/src/normalizer.js

@@ -491,8 +491,20 @@ function _pass_web(text) {
   return _sub(_URL_RE, url, text);
 }
 
+// ABBREVIATIONS, and each title also in capitals and without the spaces after its dots
+const _ABBR_FORMS = (() => {
+  const forms = new Map(L.ABBREVIATIONS);
+  for (const key of L.TITLE_ABBREVIATIONS) {
+    const spoken = L.ABBREVIATIONS.get(key);
+    for (const form of [key.replaceAll('. ', '.'), tr_upper(key), tr_upper(key).replaceAll('. ', '.')]) {
+      if (!forms.has(form)) forms.set(form, spoken);
+    }
+  }
+  return forms;
+})();
+
 function _build_abbr_re() {
-  const keys = [...L.ABBREVIATIONS.keys()].sort((a, b) => plen(b) - plen(a));
+  const keys = [..._ABBR_FORMS.keys()].sort((a, b) => plen(b) - plen(a));
   const alt = keys.map(re.escape).join('|');
   const first = [...new Set(keys.map(first_cp))].join('');
   // every key has a "." or a ":" ("Dr.", "No:")
@@ -509,7 +521,7 @@ const _SPACE_CAPITAL_RE = re.compile(`${S}+[${UP}]`);
 function _pass_abbreviations(text) {
   const repl = (m) => {
     const key = m[1];
-    let out = L.ABBREVIATIONS.get(key);
+    let out = _ABBR_FORMS.get(key);
     if (L.SENTENCE_FINAL_ABBR.has(key)) {
       const rest = text.slice(end(m));
       if (!strip(rest) || _SPACE_CAPITAL_RE.match(rest)) out += '.';