한국어 우선 MeloTTS 웹 서비스 추가 (GPU, 속도/피치, 로그인 없음)

- FastAPI 백엔드: /api/voices, /api/tts, /api/health
- 한국어 메인 + 영어(5종 악센트), 중국어/일본어 UI 제거
- 언어별 목소리 모델 선택, 속도(0.5~2.0x)/피치(±12반음) 조절
- 로그인/사용제한/과금 유도 없음 (무제한 테스트)
- NVIDIA GPU 자동 사용 (torch cu128, Blackwell sm_120)
- 유저 친화 한국어 UI (frontend)
- Docker/compose (GPU 예약), 8788 포트
- MeloTTS(순수 파이썬) 벤더링 + 검증된 버전 고정(constraints)
This commit is contained in:
claude
2026-08-23 21:57:48 +09:00
parent a6f3bac584
commit f197177ae1
77 changed files with 11621 additions and 1 deletions

144
frontend/app.js Normal file
View File

@@ -0,0 +1,144 @@
const $ = (id) => document.getElementById(id);
const els = {
text: $("text"),
charCount: $("charCount"),
language: $("language"),
speaker: $("speaker"),
speed: $("speed"),
speedVal: $("speedVal"),
pitch: $("pitch"),
pitchVal: $("pitchVal"),
generate: $("generate"),
btnText: document.querySelector(".btn-text"),
spinner: document.querySelector(".spinner"),
result: $("result"),
audio: $("audio"),
download: $("download"),
error: $("error"),
deviceChip: $("deviceChip"),
};
let voices = [];
let lastUrl = null;
function updateCharCount() {
els.charCount.textContent = els.text.value.length;
}
function populateSpeakers() {
const code = els.language.value;
const lang = voices.find((v) => v.code === code);
els.speaker.innerHTML = "";
if (!lang) return;
lang.speakers.forEach((s) => {
const opt = document.createElement("option");
opt.value = s.id;
opt.textContent = s.label;
els.speaker.appendChild(opt);
});
}
async function loadVoices() {
try {
const res = await fetch("/api/voices");
const data = await res.json();
voices = data.languages || [];
els.language.innerHTML = "";
voices.forEach((v) => {
const opt = document.createElement("option");
opt.value = v.code;
opt.textContent = v.label;
els.language.appendChild(opt);
});
populateSpeakers();
} catch (e) {
showError("목소리 목록을 불러오지 못했습니다.");
}
}
async function loadHealth() {
try {
const res = await fetch("/api/health");
const d = await res.json();
if (d.cuda) {
els.deviceChip.textContent = `GPU 가속: ${d.gpu}`;
els.deviceChip.classList.add("gpu");
} else {
els.deviceChip.textContent = "CPU 모드로 동작 중";
}
} catch (e) {
els.deviceChip.textContent = "";
}
}
function showError(msg) {
els.error.textContent = msg;
els.error.hidden = false;
}
function clearError() {
els.error.hidden = true;
}
function setLoading(on) {
els.generate.disabled = on;
els.spinner.hidden = !on;
els.btnText.textContent = on ? "생성 중…" : "음성 생성";
}
async function generate() {
clearError();
const text = els.text.value.trim();
if (!text) {
showError("읽을 내용을 입력해 주세요.");
return;
}
setLoading(true);
try {
const res = await fetch("/api/tts", {
method: "POST",
headers: { "Content-Type": "application/json" },
body: JSON.stringify({
text,
language: els.language.value,
speaker: els.speaker.value,
speed: parseFloat(els.speed.value),
pitch: parseFloat(els.pitch.value),
}),
});
if (!res.ok) {
let detail = "음성 생성에 실패했습니다.";
try {
const err = await res.json();
if (err.detail) detail = err.detail;
} catch (_) {}
throw new Error(detail);
}
const blob = await res.blob();
if (lastUrl) URL.revokeObjectURL(lastUrl);
lastUrl = URL.createObjectURL(blob);
els.audio.src = lastUrl;
els.download.href = lastUrl;
els.result.hidden = false;
els.audio.play().catch(() => {});
} catch (e) {
showError(e.message || "오류가 발생했습니다.");
} finally {
setLoading(false);
}
}
els.text.addEventListener("input", updateCharCount);
els.language.addEventListener("change", populateSpeakers);
els.speed.addEventListener("input", () => {
els.speedVal.textContent = parseFloat(els.speed.value).toFixed(2) + "x";
});
els.pitch.addEventListener("input", () => {
const v = parseInt(els.pitch.value, 10);
els.pitchVal.textContent = (v > 0 ? "+" + v : v) + " 반음";
});
els.generate.addEventListener("click", generate);
updateCharCount();
loadVoices();
loadHealth();