feat(voices): TTS时长修复 + 提取视频配音 + AI配音标识 #1656

Merged
auto-approve-bot merged 4 commits from fix/tts-duration-and-video-extract into develop 2026-09-03 19:04:36 +08:00
8 changed files with 418 additions and 3 deletions
+22 -1
View File
@@ -2,7 +2,9 @@
from __future__ import annotations
import json
import logging
import subprocess
import tempfile
from pathlib import Path
from typing import Any, Optional
@@ -430,6 +432,8 @@ def save_tts_job_to_library(
storage_key = f"uploads/voice/tts/{job.id}.{audio_format}"
tmp_path: Path | None = None
audio_duration: float | None = None
file_size = 0
try:
with tempfile.NamedTemporaryFile(suffix=f".{audio_format}", delete=False) as tmp:
tmp_path = Path(tmp.name)
@@ -445,6 +449,23 @@ def save_tts_job_to_library(
)
file_size = tmp_path.stat().st_size
storage_service.upload_file(tmp_path, storage_key, content_type=content_type)
# 从音频文件提取时长(ffprobe),作为 job.duration 的兜底
try:
proc = subprocess.run(
[
"ffprobe", "-v", "quiet", "-print_format", "json",
"-show_format", str(tmp_path),
],
capture_output=True, text=True, timeout=10,
)
if proc.returncode == 0:
fmt = json.loads(proc.stdout).get("format", {})
dur = float(fmt.get("duration", 0))
if dur > 0:
audio_duration = dur
except Exception:
logger.warning("ffprobe 提取时长失败: job_id=%s", job.id, exc_info=True)
except HTTPException:
raise
except Exception as e:
@@ -482,7 +503,7 @@ def save_tts_job_to_library(
mime_type=content_type,
metadata=metadata_,
file_size=file_size,
duration=job.duration or None,
duration=job.duration or audio_duration or None,
status=AssetStatus.READY,
classification_status=ClassificationStatus.PENDING, # 音频不参与内容分类,保持 pending 与 ingest 链路一致
uploaded_by_user_id=user_id,
+1
View File
@@ -28,4 +28,5 @@ export {
deleteTTSJob,
getTtsVoices,
previewTts,
extractVideoVoice,
} from "./jobs"
+53
View File
@@ -70,3 +70,56 @@ export const previewTts = async (data: TTSPreviewRequest): Promise<TTSPreviewRes
const response = await apiClient.post<TTSPreviewResponse>("/tts/preview", data)
return response.data
}
/**
* 从视频中提取配音(上传视频 → 后端提取人声 → 保存到配音素材库)
* 支持 mp4/mov/webm 格式
*/
export const extractVideoVoice = async (
file: File,
onProgress?: (percent: number) => void,
): Promise<{ asset_id: string; duration: number }> => {
const formData = new FormData()
formData.append("file", file)
return new Promise((resolve, reject) => {
const xhr = new XMLHttpRequest()
xhr.open("POST", "/api/v1/tts/extract-video-voice")
// 携带认证 token(从 localStorage 获取,与 apiClient 拦截器一致)
const token = localStorage.getItem("access_token")
if (token) {
xhr.setRequestHeader("Authorization", `Bearer ${token}`)
}
xhr.timeout = 10 * 60 * 1000 // 10 分钟超时
xhr.upload.onprogress = (e) => {
if (e.lengthComputable && onProgress) {
onProgress(Math.round((e.loaded / e.total) * 100))
}
}
xhr.onload = () => {
if (xhr.status >= 200 && xhr.status < 300) {
try {
resolve(JSON.parse(xhr.responseText))
} catch {
reject(new Error("服务器返回数据解析失败"))
}
} else {
try {
const err = JSON.parse(xhr.responseText)
reject(new Error(err.detail || err.message || `提取失败: HTTP ${xhr.status}`))
} catch {
reject(new Error(`提取失败: HTTP ${xhr.status}`))
}
}
}
xhr.onerror = () => reject(new Error("网络错误,请检查网络连接"))
xhr.ontimeout = () => reject(new Error("上传超时(10分钟),请检查网络或尝试更小的文件"))
xhr.send(formData)
})
}
+40 -2
View File
@@ -16,7 +16,12 @@
*/
import React, { useCallback, useEffect, useState } from "react"
import { useSearchParams } from "react-router-dom"
import { UploadOutlined, AudioOutlined, RobotOutlined } from "@ant-design/icons"
import {
UploadOutlined,
AudioOutlined,
RobotOutlined,
VideoCameraOutlined,
} from "@ant-design/icons"
import { Button } from "@/components/ui"
import PageHead from "@/components/layout/PageHead"
import { type AssetItem } from "@/api/assets"
@@ -35,6 +40,8 @@ import { useTtsSynthesize } from "./hooks/useTtsSynthesize"
import { useVoiceUpload } from "./hooks/useVoiceUpload"
import { useMaterialDelete } from "./hooks/useMaterialDelete"
import { useMaterialBatchDelete } from "./hooks/useMaterialBatchDelete"
import { useVideoExtract } from "./hooks/useVideoExtract"
import VideoExtractModal from "./components/VideoExtractModal"
import "./voices.css"
let toastIdSeq = 0
@@ -159,6 +166,18 @@ const VoiceLibrary: React.FC = () => {
handleUploadClose,
} = useVoiceUpload({ showToast })
// ── 提取视频配音 ──────────────────────────────────────
const {
extractOpen,
extractFile,
extractProgress,
isExtracting,
setExtractOpen,
handleFileSelect: handleExtractFileSelect,
handleExtract,
handleExtractClose,
} = useVideoExtract({ showToast })
// ── URL 参数自动打开上传弹窗 ────────────────────────────
const [searchParams, setSearchParams] = useSearchParams()
@@ -200,6 +219,14 @@ const VoiceLibrary: React.FC = () => {
>
</Button>
<Button
buttonType="primary"
buttonSize="sm"
icon={<VideoCameraOutlined />}
onClick={() => setExtractOpen(true)}
>
</Button>
<Button
buttonType="ghost"
buttonSize="sm"
@@ -300,7 +327,18 @@ const VoiceLibrary: React.FC = () => {
/>
)}
{/* ── 弹窗集合 ──────────────────────────────────── */}
{/* ── 视频提取配音弹窗 ─────────────────────────────── */}
<VideoExtractModal
open={extractOpen}
file={extractFile}
progress={extractProgress}
isExtracting={isExtracting}
onClose={handleExtractClose}
onFileSelect={handleExtractFileSelect}
onExtract={handleExtract}
/>
{/* ── 弹窗集合 ─────────────────────────────────── */}
<VoiceModals
cloneModalOpen={cloneModalOpen}
onCloneClose={() => setCloneModalOpen(false)}
@@ -136,6 +136,8 @@ export const MaterialVoiceTab: React.FC<MaterialVoiceTabProps> = ({
const material = mapAssetToMaterial(asset)
// duration 优先取顶层(后端从 metadata 提取),兜底 metadata
const cardDuration = asset.duration || material.duration || 0
// AI 生成素材标识(metadata.source === "tts_job"
const isAiMaterial = (asset.metadata as Record<string, unknown>)?.source === "tts_job"
const isPlaying = playingId === asset.id
const isSelected = selectedIds.has(asset.id)
// 播放中以 audio 真实时长为准,未播放显示卡片时长
@@ -184,6 +186,7 @@ export const MaterialVoiceTab: React.FC<MaterialVoiceTabProps> = ({
<div className="xx-voice-info vmat-info">
<div className="xx-voice-name" title={asset.name}>
{asset.name}
{isAiMaterial && <span className="vmat-ai-badge">AI</span>}
</div>
<div className="xx-voice-subtitle">
{asset.file_size ? `${formatFileSize(asset.file_size)}` : "--"}
@@ -0,0 +1,207 @@
import React, { useRef } from "react"
import { Modal } from "antd"
import { InboxOutlined, CloseOutlined } from "@ant-design/icons"
interface VideoExtractModalProps {
open: boolean
file: File | null
progress: number | null
isExtracting: boolean
onClose: () => void
onFileSelect: (file: File | null) => void
onExtract: () => void
}
const ACCEPT_TYPES = ".mp4,.mov,.webm"
const VideoExtractModal: React.FC<VideoExtractModalProps> = ({
open,
file,
progress,
isExtracting,
onClose,
onFileSelect,
onExtract,
}) => {
const inputRef = useRef<HTMLInputElement>(null)
return (
<Modal
title={<span style={{ fontSize: 16, fontWeight: 600 }}></span>}
open={open}
onCancel={() => {
if (isExtracting) return
onClose()
}}
footer={null}
width={480}
maskClosable={!isExtracting}
>
{!file ? (
<div
className="vmat-upload-dropzone"
onClick={() => inputRef.current?.click()}
style={{
border: "2px dashed #d9d9d9",
borderRadius: 8,
padding: "40px 20px",
textAlign: "center",
cursor: "pointer",
transition: "border-color 0.3s",
}}
onMouseEnter={(e) => (e.currentTarget.style.borderColor = "#7c3aed")}
onMouseLeave={(e) => (e.currentTarget.style.borderColor = "#d9d9d9")}
>
<InboxOutlined style={{ fontSize: 32, color: "#7c3aed", marginBottom: 12 }} />
<p style={{ margin: "0 0 8px", fontSize: 14, color: "#333" }}></p>
<span style={{ fontSize: 12, color: "#999" }}> MP4MOVWebM </span>
<input
ref={inputRef}
type="file"
accept={ACCEPT_TYPES}
style={{ display: "none" }}
onChange={(e) => {
const f = e.target.files?.[0]
if (f) onFileSelect(f)
}}
/>
</div>
) : (
<div>
<div
style={{
display: "flex",
alignItems: "center",
justifyContent: "space-between",
padding: "12px 16px",
background: "#fafafa",
borderRadius: 8,
marginBottom: 16,
}}
>
<span
style={{
flex: 1,
overflow: "hidden",
textOverflow: "ellipsis",
whiteSpace: "nowrap",
fontSize: 14,
fontWeight: 500,
}}
title={file.name}
>
{file.name}
</span>
<span style={{ fontSize: 12, color: "#999", marginLeft: 8, flexShrink: 0 }}>
{(file.size / (1024 * 1024)).toFixed(1)} MB
</span>
{!isExtracting && (
<button
type="button"
onClick={() => {
if (inputRef.current) inputRef.current.value = ""
onFileSelect(null)
}}
style={{
border: "none",
background: "none",
cursor: "pointer",
color: "#999",
marginLeft: 8,
fontSize: 14,
}}
aria-label="移除文件"
>
<CloseOutlined />
</button>
)}
</div>
{progress !== null && (
<div style={{ marginBottom: 12 }}>
<div
style={{
height: 6,
background: "#f0f0f0",
borderRadius: 3,
overflow: "hidden",
}}
>
<div
style={{
height: "100%",
width: `${progress}%`,
background: "linear-gradient(90deg, #7c3aed, #a78bfa)",
borderRadius: 3,
transition: "width 0.3s",
}}
/>
</div>
<div
style={{
textAlign: "right",
fontSize: 12,
color: "#999",
marginTop: 4,
}}
>
{progress}%
</div>
</div>
)}
{isExtracting && (
<p style={{ textAlign: "center", fontSize: 13, color: "#7c3aed", margin: "12px 0 0" }}>
{progress === 100 ? "正在提取人声,请稍候..." : "正在上传视频..."}
</p>
)}
</div>
)}
<div
style={{
display: "flex",
justifyContent: "flex-end",
gap: 8,
marginTop: 24,
}}
>
<button
type="button"
onClick={onClose}
disabled={isExtracting}
style={{
padding: "6px 16px",
borderRadius: 6,
border: "1px solid #d9d9d9",
background: "#fff",
cursor: isExtracting ? "not-allowed" : "pointer",
fontSize: 14,
opacity: isExtracting ? 0.5 : 1,
}}
>
</button>
<button
type="button"
onClick={onExtract}
disabled={!file || isExtracting}
style={{
padding: "6px 16px",
borderRadius: 6,
border: "none",
background: !file || isExtracting ? "#d9d9d9" : "#7c3aed",
color: "#fff",
cursor: !file || isExtracting ? "not-allowed" : "pointer",
fontSize: 14,
fontWeight: 500,
}}
>
{isExtracting ? "提取中..." : "开始提取"}
</button>
</div>
</Modal>
)
}
export default VideoExtractModal
@@ -0,0 +1,74 @@
import { useState, useCallback } from "react"
import { useQueryClient } from "@tanstack/react-query"
import { extractVideoVoice } from "@/api/tts"
/**
* 视频提取配音 Hook
* 封装视频上传弹窗状态、提取进度、提取 mutation 逻辑
*/
interface UseVideoExtractProps {
showToast: (message: string, type: "success" | "error") => void
}
export function useVideoExtract({ showToast }: UseVideoExtractProps) {
const queryClient = useQueryClient()
const [extractOpen, setExtractOpen] = useState(false)
const [extractFile, setExtractFile] = useState<File | null>(null)
const [extractProgress, setExtractProgress] = useState<number | null>(null)
const [isExtracting, setIsExtracting] = useState(false)
const handleExtractClose = useCallback(() => {
setExtractOpen(false)
setExtractFile(null)
setExtractProgress(null)
setIsExtracting(false)
}, [])
const handleExtract = useCallback(async () => {
if (!extractFile) return
setIsExtracting(true)
setExtractProgress(0)
try {
await extractVideoVoice(extractFile, (p) => setExtractProgress(p))
// 刷新素材列表
queryClient.invalidateQueries({ queryKey: ["assets", "voice"] })
queryClient.invalidateQueries({ queryKey: ["voice-materials"] })
showToast("视频配音提取成功", "success")
handleExtractClose()
} catch (err: unknown) {
const msg = err instanceof Error ? err.message : "提取失败,请重试"
showToast(msg, "error")
} finally {
setIsExtracting(false)
setExtractProgress(null)
}
}, [extractFile, queryClient, showToast, handleExtractClose])
const handleFileSelect = useCallback(
(file: File | null) => {
if (!file) {
setExtractFile(null)
return
}
const validTypes = ["video/mp4", "video/quicktime", "video/webm"]
if (!validTypes.includes(file.type)) {
showToast("仅支持 MP4、MOV、WebM 格式的视频文件", "error")
return
}
setExtractFile(file)
},
[showToast],
)
return {
extractOpen,
setExtractOpen,
extractFile,
extractProgress,
isExtracting,
handleFileSelect,
handleExtract,
handleExtractClose,
}
}
+18
View File
@@ -193,6 +193,24 @@
overflow: hidden;
text-overflow: ellipsis;
flex: 1;
display: flex;
align-items: center;
}
/* AI 配音标识 */
.vmat-ai-badge {
display: inline-block;
margin-left: 6px;
padding: 1px 6px;
font-size: 11px;
font-weight: 600;
color: #7c3aed;
background: #f3f0ff;
border: 1px solid #ddd6fe;
border-radius: 4px;
line-height: 16px;
vertical-align: middle;
flex-shrink: 0;
}
.xx-voice-star {