Questions: '?' glued to the word, as in v2's training text (a space made wh-questions rise)

cleaner.js

@@ -1,7 +1,7 @@
 // Browser port of matcha_patch/polish_cleaners.py + polish_mixed_cleaners: Polish text normalisation, espeak-ng (WASM) IPA,
 // English words in {braces} phonemized with the en-us voice, symbol ids. Must match the Python cleaners character for character.
 import { initialize, setVoice, getPhonemes } from "https://cdn.jsdelivr.net/npm/espeak-phonemizer@0.1.2/src/espeak.mjs";
-import { SYMBOL_TO_ID } from "./symbols.js?v=20261005b";
+import { SYMBOL_TO_ID } from "./symbols.js?v=20261005c";
 
 const PL_WORD = "0-9a-ząćęłńóśźż";
 const esc = (s) => s.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
@@ -62,7 +62,7 @@ const unstress = (ph) => { const i = ph.indexOf(" "); const head = i < 0 ? ph :
 
 export async function textToIpa(text) {
   await initPhonemizer();
-  text = text.replace(/\s*\?/g, " ?");   // "kuchni?" -> "kuchni ?": '?' as its own token gives Matcha the rising question intonation
+  text = text.replace(/\s+\?/g, "?");   // v2 was trained on VoxCPM2 texts with '?' glued to the word: "kuchni ?" -> "kuchni?" (a space makes wh-questions rise)
   if (!text.includes("{")) return plPhon(text);
   let out = "", pos = 0;
   for (const m of text.matchAll(/\{([^{}]+)\}/g)) {

english.js

@@ -3,7 +3,7 @@
 // unlisted words; a frequency table is too big for the browser). Inflected forms whose full form is a common Polish word
 // (wina, legendy, deskę) are precomputed in english_data.json -> pl_inflected_block.
 let D = null;
-export async function initEnglish(url = new URL("./english_data.json?v=20261005b", import.meta.url)) {
+export async function initEnglish(url = new URL("./english_data.json?v=20261005c", import.meta.url)) {
   if (!D) {
     const j = await (await fetch(url)).json();
     D = { cur: new Set(j.curated), rap: new Set(j.read_as_polish), suf: j.suffixes, loan: j.loan_stems, block: new Set(j.pl_inflected_block) };

index.html

@@ -92,9 +92,9 @@ footer img{height:30px;display:block}
 <script>window.coi={coepCredentialless:()=>true}</script><script src="coi-serviceworker.js"></script>
 <script src="https://cdn.jsdelivr.net/npm/onnxruntime-web@1.30.0/dist/ort.all.min.js" crossorigin="anonymous"></script>
 <script type="module">
-import { textToIpa, ipaToIds, initPhonemizer } from "./cleaner.js?v=20261005b";
-import { initEnglish, markEnglish } from "./english.js?v=20261005b";
-import { EFFECTS, DEFAULTS, UI_DEFAULTS, applyChain } from "./effects.js?v=20261005b";
+import { textToIpa, ipaToIds, initPhonemizer } from "./cleaner.js?v=20261005c";
+import { initEnglish, markEnglish } from "./english.js?v=20261005c";
+import { EFFECTS, DEFAULTS, UI_DEFAULTS, applyChain } from "./effects.js?v=20261005c";
 const $=id=>document.getElementById(id); const SR=22050;
 const LOCAL = location.hostname==="localhost"||location.hostname==="127.0.0.1";
 const MODEL_BASE = LOCAL ? "/release-v2/" : "https://huggingface.co/machinekind/Matcha-TTS-PL-v2/resolve/main/";