Ma/auto-caption
1
0
Fork 0
mirror of https://github.com/HiMeditator/auto-caption.git synced 2026-02-04 04:14:42 +08:00
Code Issues Packages Projects Releases Wiki Activity

feat(control): 重构项目，增加字幕引擎配置

Browse Source
This commit is contained in:
himeditator
2025-06-19 18:59:13 +08:00
parent 5da9c3c038
commit 54c618aa3f
17 changed files with 152 additions and 180 deletions
Download Patch File Download Diff File Expand all files Collapse all files

107
python-prototype/gummy.ipynb
View File

@@ -156,112 +156,9 @@
},
{
"cell_type": "code",
"execution_count": 5,
"execution_count": null,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"流式翻译开始...\n",
"\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n",
"(4410, 2)\n"
]
},
{
"ename": "KeyboardInterrupt",
"evalue": "",
"output_type": "error",
"traceback": [
"\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
"\u001b[1;31mKeyboardInterrupt\u001b[0m Traceback (most recent call last)",
"\u001b[1;32m~\\AppData\\Local\\Temp\\ipykernel_29036\\3259296939.py\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m 21\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m 22\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mRATE\u001b[0m \u001b[1;33m/\u001b[0m \u001b[0mCHUNK\u001b[0m \u001b[1;33m*\u001b[0m \u001b[0mRECORD_SECONDS\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 23\u001b[1;33m \u001b[0mdata\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mstream\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mread\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mCHUNK\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m 24\u001b[0m \u001b[0mdata_np\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfrombuffer\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mint16\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m 25\u001b[0m \u001b[0mdata_np_r\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mdata_np\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mreshape\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m-\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mCHANNELS\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
"\u001b[1;32md:\\ML\\anaconda3\\envs\\mystd\\lib\\site-packages\\pyaudiowpatch\\__init__.py\u001b[0m in \u001b[0;36mread\u001b[1;34m(self, num_frames, exception_on_overflow)\u001b[0m\n\u001b[0;32m 638\u001b[0m paCanNotReadFromAnOutputOnlyStream)\n\u001b[0;32m 639\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 640\u001b[1;33m \u001b[1;32mreturn\u001b[0m \u001b[0mpa\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mread_stream\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_stream\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mnum_frames\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mexception_on_overflow\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m 641\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m 642\u001b[0m \u001b[1;32mdef\u001b[0m \u001b[0mget_read_available\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
"\u001b[1;31mKeyboardInterrupt\u001b[0m: "
]
}
],
"outputs": [],
"source": [
"RECORD_SECONDS = 20 # 监听时长(s)\n",
"\n",

16
python-subprocess/main.py
View File

@@ -1,23 +1,27 @@
from sysaudio.win import LoopbackStream, mergeStreamChannels
from audio2text.gummy import GummyTranslator
import sys
import argparse
def main():
def convert_audio_to_text(s_lang, t_lang, audio_source):
sys.stdout.reconfigure(line_buffering=True)
loopback = LoopbackStream()
loopback.openStream()
gummy = GummyTranslator(loopback.RATE, "zh", "en")
gummy = GummyTranslator(loopback.RATE, s_lang, t_lang)
gummy.translator.start()
for _ in range(0, 400):
while True:
if not loopback.stream: continue
data = loopback.stream.read(loopback.CHUNK)
data = mergeStreamChannels(data, loopback.CHANNELS)
gummy.translator.send_audio_frame(data)
gummy.translator.stop()
loopback.closeStream()
if __name__ == "__main__":
main()
parser = argparse.ArgumentParser(description='Convert system audio stream to text')
parser.add_argument('-s', '--s_lang', default='en', help='Source language code')
parser.add_argument('-t', '--t_lang', default='zh', help='Target language code')
parser.add_argument('-a', '--audio', default=0, help='Audio stream source: 0 for output audio stream, 1 for input audio stream')
args = parser.parse_args()
convert_audio_to_text(args.s_lang, args.t_lang, args.audio)

8
src/main/caption.ts
View File

@@ -3,8 +3,7 @@ import path from 'path'
import { is } from '@electron-toolkit/utils'
import icon from '../../resources/icon.png?asset'
import { controlWindow } from './control'
import { sendStyles, sendCaptionLog } from './data'
import { send } from 'vite'
import { sendStyles, sendCaptionLog } from './utils/config'
class CaptionWindow {
window: BrowserWindow | undefined;
@@ -40,7 +39,6 @@ class CaptionWindow {
})
this.window.on('closed', () => {
console.log('INFO caption window closed')
this.window = undefined
})
@@ -63,7 +61,6 @@ class CaptionWindow {
ipcMain.on('caption.controlWindow.activate', () => {
if(!controlWindow.window){
controlWindow.createWindow()
console.log('GET caption.controlWindow.activate')
}
else {
controlWindow.window.show()
@@ -71,21 +68,18 @@ class CaptionWindow {
})
// 字幕窗口高度发生变化
ipcMain.on('caption.windowHeight.change', (_, height) => {
console.log('GET caption.window.height.change', height)
if(this.window){
this.window.setSize(this.window.getSize()[0], height)
}
})
// 关闭字幕窗口
ipcMain.on('caption.window.close', () => {
console.log('GET caption.window.close')
if(this.window){
this.window.close()
}
})
// 是否固定在最前面
ipcMain.on('caption.pin.set', (_, pinned) => {
console.log('GET caption.pin.set', pinned)
if(this.window){
this.window.setAlwaysOnTop(pinned)
}

14
src/main/control.ts
View File

@@ -2,8 +2,13 @@ import { shell, BrowserWindow, ipcMain } from 'electron'
import path from 'path'
import { is } from '@electron-toolkit/utils'
import icon from '../../resources/icon.png?asset'
import { setStyles, sendStyles, sendCaptionLog } from './data'
import { captionWindow } from './caption'
import {
setStyles,
sendStyles,
sendCaptionLog,
setControls
} from './utils/config'
class ControlWindow {
window: BrowserWindow | undefined;
@@ -38,7 +43,6 @@ class ControlWindow {
})
this.window.on('closed', () => {
console.log('INFO control window closed')
this.window = undefined
})
@@ -57,7 +61,6 @@ class ControlWindow {
public handleMessage() {
// 控制窗口样式更新
ipcMain.on('control.style.change', (_, args) => {
console.log('GET control.style.change', args)
setStyles(args)
if(captionWindow.window){
sendStyles(captionWindow.window)
@@ -67,12 +70,15 @@ class ControlWindow {
ipcMain.on('control.captionWindow.activate', () => {
if(!captionWindow.window){
captionWindow.createWindow()
console.log('GET control.captionWindow.activate')
}
else {
captionWindow.window.show()
}
})
// 字幕引擎控制配置更新
ipcMain.on('control.control.change', (_, args) => {
setControls(args)
})
}
}

4
src/main/index.ts
View File

@@ -3,7 +3,7 @@ import { electronApp, optimizer } from '@electron-toolkit/utils'
import { controlWindow } from './control'
import { captionWindow } from './caption'
import { PythonProcess } from './pythonProcess'
import { PythonProcess } from './utils/pythonProcess'
const pySubProcess = new PythonProcess()
app.whenReady().then(() => {
@@ -18,7 +18,7 @@ app.whenReady().then(() => {
controlWindow.createWindow()
pySubProcess.start()
// pySubProcess.start()
app.on('activate', function () {
if (BrowserWindow.getAllWindows().length === 0){

29
src/main/types/index.ts Normal file
View File

@@ -0,0 +1,29 @@
export interface Styles {
fontFamily: string,
fontSize: number,
fontColor: string,
background: string,
opacity: number,
transDisplay: boolean,
transFontFamily: string,
transFontSize: number,
transFontColor: string
}
export interface CaptionItem {
index: number,
time_s: string,
time_t: string,
text: string,
translation: string
}
export interface Controls {
sourceLang: string,
targetLang: string,
engine: string,
translation: boolean,
customized: boolean,
customizedApp: string,
customizedCommand: string
}

53
src/main/data.ts → src/main/utils/config.ts
View File

@@ -1,26 +1,7 @@
import { Styles, CaptionItem, Controls } from '../types'
import { BrowserWindow } from 'electron'
export interface Styles {
fontFamily: string,
fontSize: number,
fontColor: string,
background: string,
opacity: number,
transDisplay: boolean,
transFontFamily: string,
transFontSize: number,
transFontColor: string
}
export interface CaptionItem {
index: number,
time_s: string,
time_t: string,
text: string,
translation: string
}
export let styles: Styles = {
export const styles: Styles = {
fontFamily: 'sans-serif',
fontSize: 24,
fontColor: '#000000',
@@ -32,6 +13,18 @@ export let styles: Styles = {
transFontColor: '#000000'
}
export const captionLog: CaptionItem[] = []
export const controls: Controls = {
sourceLang: 'en',
targetLang: 'zh',
engine: 'gummy',
translation: true,
customized: false,
customizedApp: '',
customizedCommand: ''
}
export function setStyles(args: any) {
styles.fontFamily = args.fontFamily
styles.fontSize = args.fontSize
@@ -42,18 +35,16 @@ export function setStyles(args: any) {
styles.transFontFamily = args.transFontFamily
styles.transFontSize = args.transFontSize
styles.transFontColor = args.transFontColor
console.log('[INFO] Set Styles:', styles)
}
export function sendStyles(window: BrowserWindow) {
window.webContents.send('caption.style.set', styles)
console.log('SNED caption.style.set')
console.log('[INFO] Send Styles:', styles)
}
export let captionLog: CaptionItem[] = []
export function sendCaptionLog(window: BrowserWindow) {
window.webContents.send('both.log.set', captionLog)
console.log('SEND both.log.set')
}
export function addCaptionLog(log: CaptionItem) {
@@ -64,8 +55,18 @@ export function addCaptionLog(log: CaptionItem) {
else {
captionLog.push(log)
}
console.log('ADD caption')
for(const window of BrowserWindow.getAllWindows()){
sendCaptionLog(window)
}
}
export function setControls(args: any) {
controls.sourceLang = args.sourceLang
controls.targetLang = args.targetLang
controls.engine = args.engine
controls.translation = args.translation
controls.customized = args.customized
controls.customizedApp = args.customizedApp
controls.customizedCommand = args.customizedCommand
console.log('[INFO] Set Controls:', controls)
}

2
src/main/pythonProcess.ts → src/main/utils/pythonProcess.ts
View File

@@ -1,7 +1,7 @@
import { spawn } from 'child_process'
import { app } from 'electron'
import path from 'path'
import { addCaptionLog } from './data'
import { addCaptionLog } from './config'
export class PythonProcess {
public start() {

2
src/renderer/index.html
View File

@@ -2,7 +2,7 @@
<html>
<head>
<meta charset="UTF-8" />
<title>Auto Caption Player</title>
<title>Auto Caption</title>
<!-- https://developer.mozilla.org/en-US/docs/Web/HTTP/CSP -->
<meta
http-equiv="Content-Security-Policy"

0
src/renderer/src/assets/styles/reset.css → src/renderer/src/assets/reset.css
View File

BIN
src/renderer/src/assets/video/example-ja.mp4
View File

Binary file not shown.

56
src/renderer/src/components/CaptionControl.vue
View File

@@ -9,7 +9,6 @@
<span class="control-label">源语言</span>
<a-select
class="control-input"
ref="select"
v-model:value="currentSourceLang"
:options="langList"
></a-select>
@@ -18,7 +17,6 @@
<span class="control-label">翻译语言</span>
<a-select
class="control-input"
ref="select"
v-model:value="currentTargetLang"
:options="langList.filter((item) => item.value !== 'auto')"
></a-select>
@@ -27,23 +25,34 @@
<span class="control-label">字幕引擎</span>
<a-select
class="control-input"
ref="select"
v-model:value="currentEngine"
:options="captionEngine"
></a-select>
</div>
<div class="control-item">
<span class="control-label">端口号</span>
<a-input
class="control-input"
ref="select"
type="number"
v-model:value="currentPort"
></a-input>
</div>
<div class="control-item">
<span class="control-label">启用翻译</span>
<a-switch v-model:checked="currentTranslation" />
<span class="control-label">自定义引擎</span>
<a-switch v-model:checked="currentCustomized" />
</div>
<div v-show="currentCustomized">
<a-card size="small" title="自定义字幕引擎">
<p class="customize-note">说明：允许用户使用自定义字幕引擎提供字幕。提供的引擎要能通过 <code>child_process.spawn()</code> 进行启动，且需要通过 IPC 与项目 node.js 后端进行通信。具体通信接口见后端实现。</p>
<div class="control-item">
<span class="control-label">引擎路径</span>
<a-input
class="control-input"
v-model:value="currentCustomizedApp"
></a-input>
</div>
<div class="control-item">
<span class="control-label">引擎指令</span>
<a-input
class="control-input"
v-model:value="currentCustomizedCommand"
></a-input>
</div>
</a-card>
</div>
</a-card>
<div style="height: 20px;"></div>
@@ -60,9 +69,12 @@ const { captionEngine } = storeToRefs(captionControl)
const currentSourceLang = ref('auto')
const currentTargetLang = ref('zh')
const currentEngine = ref('gummy')
const currentPort = ref(8765)
const currentTranslation = ref<boolean>(false)
const currentCustomized = ref<boolean>(false)
const currentCustomizedApp = ref('')
const currentCustomizedCommand = ref('')
const langList = computed(() => {
for(let item of captionEngine.value){
if(item.value === currentEngine.value) {
@@ -76,16 +88,24 @@ function applyChange(){
captionControl.sourceLang = currentSourceLang.value
captionControl.targetLang = currentTargetLang.value
captionControl.engine = currentEngine.value
captionControl.port = currentPort.value
captionControl.translation = currentTranslation.value
captionControl.customized = currentCustomized.value
captionControl.customizedApp = currentCustomizedApp.value
captionControl.customizedCommand = currentCustomizedCommand.value
captionControl.sendControlChange()
}
function cancelChange(){
currentSourceLang.value = captionControl.sourceLang
currentTargetLang.value = captionControl.targetLang
currentEngine.value = captionControl.engine
currentPort.value = captionControl.port
currentTranslation.value = captionControl.translation
currentCustomized.value = captionControl.customized
currentCustomizedApp.value = captionControl.customizedApp
currentCustomizedCommand.value = captionControl.customizedCommand
}
</script>
@@ -101,6 +121,12 @@ function cancelChange(){
margin-right: 10px;
}
.customize-note {
padding: 0 20px;
color: red;
font-size: 12px;
}
.control-input {
width: calc(100% - 100px);
min-width: 100px;

2
src/renderer/src/main.ts
View File

@@ -1,4 +1,4 @@
import './assets/styles/reset.css'
import './assets/reset.css'
import { createPinia } from 'pinia'
import { createApp } from 'vue'

35
src/renderer/src/stores/captionControl.ts
View File

@@ -26,17 +26,36 @@ export const useCaptionControlStore = defineStore('captionControl', () => {
])
const sourceLang = ref<string>('auto')
const targetLang = ref<string>('')
const targetLang = ref<string>('zh')
const engine = ref<string>('gummy')
const port = ref<number>(8765)
const translation = ref<boolean>(false)
const customized = ref<boolean>(false)
const customizedApp = ref<string>('')
const customizedCommand = ref<string>('')
function sendControlChange() {
const controls = {
sourceLang: sourceLang.value,
targetLang: targetLang.value,
engine: engine.value,
translation: translation.value,
customized: customized.value,
customizedApp: customizedApp.value,
customizedCommand: customizedCommand.value
}
window.electron.ipcRenderer.send('control.control.change', controls)
}
return {
captionEngine, // 字幕引擎
sourceLang, // 源语言
targetLang, // 目标语言
engine, // 字幕引擎
port, // 端口
translation // 是否启用翻译
captionEngine, // 字幕引擎
sourceLang, // 源语言
targetLang, // 目标语言
engine, // 字幕引擎
translation, // 是否启用翻译
customized, // 是否使用自定义字幕引擎
customizedApp, // 自定义字幕引擎的应用程序
customizedCommand, // 自定义字幕引擎的命令
sendControlChange // 发送最新控制消息到后端
}
})

1
src/renderer/src/stores/captionLog.ts
View File

@@ -14,7 +14,6 @@ export const useCaptionLogStore = defineStore('captionLog', () => {
window.electron.ipcRenderer.on('both.log.set', (_, logs) => {
captionData.value = logs
console.log('GET both.log.set', logs)
})
return {

2
src/renderer/src/stores/captionStyle.ts
View File

@@ -36,7 +36,6 @@ export const useCaptionStyleStore = defineStore('captionStyle', () => {
transFontColor: transFontColor.value
}
window.electron.ipcRenderer.send('control.style.change', styles)
console.log('SEND control.style.change', styles)
}
window.electron.ipcRenderer.on('caption.style.set', (_, args) => {
@@ -50,7 +49,6 @@ export const useCaptionStyleStore = defineStore('captionStyle', () => {
transFontSize.value = args.transFontSize
transFontColor.value = args.transFontColor
changeSignal.value = true
console.log('GET caption.style.set', args)
})
return {

1
src/renderer/src/views/CaptionPage.vue
View File

@@ -58,7 +58,6 @@ onMounted(() => {
for (const entry of entries) {
if(windowHeight.value !== Math.floor(entry.contentRect.height) + 2) {
windowHeight.value = Math.floor(entry.contentRect.height) + 2;
console.log('INFO window height change', windowHeight.value);
window.electron.ipcRenderer.send('caption.windowHeight.change', windowHeight.value)
}
}