问题：如何把豆包的双语实时字幕抓取下来？这是豆包的双语实时字幕，目前没有下载字幕的功能，我如何下载？我会写程序，请用python语言帮我解决这个问题。
如何导出或者下载字幕
实现的效果：
capture-doubao-subtitles.ps1
bash
param(
[int]$X = 20,
[int]$Y = 145,
[int]$Width = 1180,
[int]$Height = 815,
[double]$IntervalSeconds = 1.0,
[string]$OutputPath = "C:\Users\dannel\Documents\Codex\2026-06-08\new-chat\outputs\doubao-subtitles.txt",
[string]$DebugImagePath = "C:\Users\dannel\Documents\Codex\2026-06-08\new-chat\work\last-subtitle-capture.png",
[switch]$KeepDuplicateScreens,
[switch]$Once
)
Add-Type -AssemblyName System.Drawing
Add-Type -AssemblyName System.Windows.Forms
Add-Type -AssemblyName System.Runtime.WindowsRuntime
$null = [Windows.Globalization.Language, Windows.Foundation, ContentType = WindowsRuntime]
$null = [Windows.Media.Ocr.OcrEngine, Windows.Foundation, ContentType = WindowsRuntime]
$null = [Windows.Storage.StorageFile, Windows.Storage, ContentType = WindowsRuntime]
$null = [Windows.Graphics.Imaging.BitmapDecoder, Windows.Graphics, ContentType = WindowsRuntime]
$asTaskMethods = [System.WindowsRuntimeSystemExtensions].GetMethods() | Where-Object { $_.Name -eq "AsTask" }
function Await-WinRtOperation {
param(
[Parameter(Mandatory = $true)] $Operation,
[Parameter(Mandatory = $true)] [Type] $ResultType
)
$method = $asTaskMethods |
Where-Object {
$_.IsGenericMethodDefinition -and
$_.GetParameters().Count -eq 1 -and
$_.GetParameters()[0].ParameterType.Name -eq 'IAsyncOperation`1'
} |
Select-Object -First 1
if ($null -eq $method) {
throw "Cannot find WindowsRuntimeSystemExtensions.AsTask for IAsyncOperation."
}
$task = $method.MakeGenericMethod($ResultType).Invoke($null, @($Operation))
$task.Wait()
return $task.Result
}
function Capture-Region {
param(
[int]$Left,
[int]$Top,
[int]$CaptureWidth,
[int]$CaptureHeight,
[string]$Path
)
$bitmap = New-Object System.Drawing.Bitmap $CaptureWidth, $CaptureHeight
$graphics = [System.Drawing.Graphics]::FromImage($bitmap)
try {
$graphics.CopyFromScreen($Left, $Top, 0, 0, $bitmap.Size)
$bitmap.Save($Path, [System.Drawing.Imaging.ImageFormat]::Png)
}
finally {
$graphics.Dispose()
$bitmap.Dispose()
}
}
function Read-OcrText {
param([string]$Path)
$file = Await-WinRtOperation ([Windows.Storage.StorageFile]::GetFileFromPathAsync($Path)) ([Windows.Storage.StorageFile])
$stream = Await-WinRtOperation ($file.OpenReadAsync()) ([Windows.Storage.Streams.IRandomAccessStreamWithContentType])
try {
$decoder = Await-WinRtOperation ([Windows.Graphics.Imaging.BitmapDecoder]::CreateAsync($stream)) ([Windows.Graphics.Imaging.BitmapDecoder])
$softwareBitmap = Await-WinRtOperation ($decoder.GetSoftwareBitmapAsync()) ([Windows.Graphics.Imaging.SoftwareBitmap])
$language = New-Object Windows.Globalization.Language "zh-Hans-CN"
$engine = [Windows.Media.Ocr.OcrEngine]::TryCreateFromLanguage($language)
if ($null -eq $engine) {
$engine = [Windows.Media.Ocr.OcrEngine]::TryCreateFromUserProfileLanguages()
}
if ($null -eq $engine) {
throw "Windows OCR engine is not available."
}
$result = Await-WinRtOperation ($engine.RecognizeAsync($softwareBitmap)) ([Windows.Media.Ocr.OcrResult])
return ($result.Lines | ForEach-Object { $_.Text }) -join "`n"
}
finally {
if ($stream) {
$stream.Dispose()
}
}
}
function Remove-ToolbarText {
param([string]$Text)
$keptLines = foreach ($line in ($Text -split "`r?`n")) {
$trimmed = $line.Trim()
$compact = $trimmed -replace "\s+", ""
if ([string]::IsNullOrWhiteSpace($trimmed)) {
continue
}
if ($compact -match "^(翻译为[:：]?中文|关闭原文|Aa.*字号|收起字幕|展开字幕)$") {
continue
}
if ($compact.Length -le 1 -and $compact -match "^[^\p{L}\p{N}]$") {
continue
}
$trimmed
}
return ($keptLines -join "`n").Trim()
}
function Format-SubtitleText {
param([string]$Text)
$cjk = "[\u3400-\u9FFF\uF900-\uFAFF]"
$cjkOrAscii = "[\u3400-\u9FFF\uF900-\uFAFF0-9A-Za-z%]"
$cjkPunctuation = "[\u3001\u3002\uff0c\uff1a\uff1b\uff01\uff1f\uff09\uff1d]"
$openingPunctuation = "[\uff08]"
$formattedLines = foreach ($line in ($Text -split "`r?`n")) {
$value = $line.Trim()
if ([string]::IsNullOrWhiteSpace($value)) {
continue
}
$value = $value -replace "(?<=$cjkOrAscii)\s+(?=$cjk)", ""
$value = $value -replace "(?<=$cjk)\s+(?=$cjkOrAscii)", ""
$value = $value -replace "\s+($cjkPunctuation)", '$1'
$value = $value -replace "($cjkPunctuation)\s+(?=$cjkOrAscii)", '$1'
$value = $value -replace "($openingPunctuation)\s+", '$1'
$value
}
return ($formattedLines -join "`n").Trim()
}
function Normalize-SubtitleLine {
param([string]$Line)
return ($Line.ToLowerInvariant() -replace "[^\p{L}\p{N}]+", "").Trim()
}
function Get-NewSubtitleLines {
param(
[string]$Text,
[hashtable]$SeenLines
)
$newLines = New-Object System.Collections.Generic.List[string]
foreach ($line in ($Text -split "`r?`n")) {
$trimmed = $line.Trim()
if ([string]::IsNullOrWhiteSpace($trimmed)) {
continue
}
$key = Normalize-SubtitleLine $trimmed
if ($key.Length -lt 4) {
continue
}
if (-not $SeenLines.ContainsKey($key)) {
$SeenLines[$key] = $true
$newLines.Add($trimmed)
}
}
return ($newLines -join "`n").Trim()
}
New-Item -ItemType Directory -Force -Path (Split-Path -Parent $OutputPath) | Out-Null
New-Item -ItemType Directory -Force -Path (Split-Path -Parent $DebugImagePath) | Out-Null
$lastText = ""
$seenLines = @{}
Write-Output "Saving subtitles to: $OutputPath"
Write-Output "Press Ctrl+C to stop."
while ($true) {
Capture-Region -Left $X -Top $Y -CaptureWidth $Width -CaptureHeight $Height -Path $DebugImagePath
$text = Format-SubtitleText (Remove-ToolbarText (Read-OcrText -Path $DebugImagePath))
$textToSave = $text
if (-not $KeepDuplicateScreens) {
$textToSave = Get-NewSubtitleLines -Text $text -SeenLines $seenLines
}
if (-not [string]::IsNullOrWhiteSpace($textToSave) -and ($KeepDuplicateScreens -or $text -ne $lastText)) {
Add-Content -Path $OutputPath -Encoding UTF8 -Value $textToSave
Add-Content -Path $OutputPath -Encoding UTF8 -Value ""
Write-Output $textToSave
$lastText = $text
}
if ($Once) {
break
}
Start-Sleep -Milliseconds ([int]($IntervalSeconds * 1000))
}format-doubao-subtitles.ps1
bash
param(
[string]$InputPath = "C:\Users\dannel\Documents\Codex\2026-06-08\new-chat\outputs\doubao-subtitles.txt",
[string]$OutputPath = "C:\Users\dannel\Documents\Codex\2026-06-08\new-chat\outputs\doubao-subtitles-formatted.txt"
)
function Format-SubtitleText {
param([string]$Text)
$cjk = "[\u3400-\u9FFF\uF900-\uFAFF]"
$cjkOrAscii = "[\u3400-\u9FFF\uF900-\uFAFF0-9A-Za-z%]"
$cjkPunctuation = "[\u3001\u3002\uff0c\uff1a\uff1b\uff01\uff1f\uff09\uff1d]"
$openingPunctuation = "[\uff08]"
$formattedLines = foreach ($line in ($Text -split "`r?`n")) {
$value = $line.Trim()
if ([string]::IsNullOrWhiteSpace($value)) {
continue
}
if ($value -match "^\[\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}\]$") {
continue
}
$value = $value -replace "(?<=$cjkOrAscii)\s+(?=$cjk)", ""
$value = $value -replace "(?<=$cjk)\s+(?=$cjkOrAscii)", ""
$value = $value -replace "\s+($cjkPunctuation)", '$1'
$value = $value -replace "($cjkPunctuation)\s+(?=$cjkOrAscii)", '$1'
$value = $value -replace "($openingPunctuation)\s+", '$1'
$value
}
return ($formattedLines -join "`n").Trim()
}
$text = Get-Content -Path $InputPath -Raw
Format-SubtitleText $text | Set-Content -Path $OutputPath -Encoding UTF8
Write-Output $OutputPathstart-doubao-subtitle-capture.cmd
bash
@echo off
powershell -ExecutionPolicy Bypass -File "%~dp0capture-doubao-subtitles.ps1" -X 20 -Y 145 -Width 1180 -Height 815 -IntervalSeconds 1
pause实现的效果并不是太完美，感兴趣的coder可以一起完善。