Add Whisper transcription support and config

.env.example

@@ -14,14 +14,15 @@ RATE_LIMIT_WINDOW=60
 # Logging configuration
 LOG_LEVEL=INFO
 
-# yt-dlp configuration
-YT_DLP_TIMEOUT_LIST=30
-YT_DLP_TIMEOUT_DOWNLOAD=60
+# yt-dlp configuration (audio download timeout in seconds)
+YT_DLP_TIMEOUT_DOWNLOAD=120
+
+# Whisper configuration
+# Models: tiny, base, small, medium, large-v2, large-v3
+WHISPER_MODEL=base
 
 # Embedding model configuration
 EMBEDDING_MODEL=mixedbread-ai/mxbai-embed-large-v1
-# Models cache directory (set by Docker, optional for local dev)
-# SENTENCE_TRANSFORMERS_HOME=models
 
 # Server configuration
 HOST=0.0.0.0
@@ -29,4 +30,6 @@ PORT=8000
 RELOAD=true
 
 # Container-specific (set automatically in Dockerfile)
-# DISABLE_FILE_LOGGING=true
+# DISABLE_FILE_LOGGING=true
+# SENTENCE_TRANSFORMERS_HOME=/app/models
+# WHISPER_MODELS_DIR=/app/models

Dockerfile

@@ -45,6 +45,7 @@ ENV PYTHONPATH=/app
 ENV HF_HOME=/app/models
 ENV TRANSFORMERS_CACHE=/app/models
 ENV SENTENCE_TRANSFORMERS_HOME=/app/models
+ENV WHISPER_MODELS_DIR=/app/models
 
 # Disable file logging in container
 ENV DISABLE_FILE_LOGGING=true

app/apis/subtitles/service.py

@@ -1,10 +1,12 @@
-"""Subtitle extraction service using yt-dlp with caching."""
+"""Subtitle extraction service using yt-dlp audio download and Whisper transcription."""
 
 import asyncio
+import os
 import sys
 import tempfile
+import threading
 from pathlib import Path
-from typing import List, Optional, Tuple
+from typing import List, Tuple
 
 from cachetools import TTLCache
 
@@ -15,109 +17,89 @@ from app.core.exceptions import (
     SubtitleExtractionError,
     InvalidVideoURLError
 )
-from app.apis.subtitles.utils import extract_video_id, convert_vtt_to_text
+from app.apis.subtitles.utils import extract_video_id
+from app.core.logging import get_logger
 
+logger = get_logger(__name__)
 
 SUBTITLE_CACHE: TTLCache = TTLCache(maxsize=100, ttl=3600)
-
-ALTERNATIVE_LANGUAGES = {
-    "en": ["en-US", "en-GB", "en-orig"],
-    "es": ["es-ES", "es-MX", "es-419"],
-    "fr": ["fr-FR", "fr-CA"],
-    "de": ["de-DE"],
-    "it": ["it-IT"],
-    "pt": ["pt-BR", "pt-PT"],
-    "ja": ["ja-JP"],
-    "ko": ["ko-KR"],
-    "zh": ["zh-CN", "zh-TW", "zh-Hans", "zh-Hant"]
-}
+MODELS_DIR = os.environ.get("WHISPER_MODELS_DIR", os.environ.get("SENTENCE_TRANSFORMERS_HOME", "models"))
 
 
 class SubtitleService:
-    """Service for extracting subtitles from YouTube videos."""
+    """Service for extracting subtitles from videos via Whisper transcription."""
 
     def __init__(self) -> None:
-        self.timeout_list = settings.yt_dlp_timeout_list
+        self._whisper_model = None
+        self._model_name = settings.whisper_model
+        self._lock = threading.Lock()
         self.timeout_download = settings.yt_dlp_timeout_download
 
+    def _load_whisper_model(self):
+        """Lazy load the Whisper model on first use."""
+        if self._whisper_model is None:
+            with self._lock:
+                if self._whisper_model is None:
+                    logger.info(f"Loading Whisper model: {self._model_name}")
+                    logger.info(f"Models directory: {MODELS_DIR}")
+                    from faster_whisper import WhisperModel
+                    self._whisper_model = WhisperModel(
+                        self._model_name,
+                        device="cpu",
+                        compute_type="int8",
+                        download_root=MODELS_DIR
+                    )
+                    logger.info("Whisper model loaded successfully")
+
     async def extract_subtitles(self, url: str, lang: str = "en") -> Tuple[str, List[str]]:
         """
-        Extract subtitles from a YouTube video.
+        Extract subtitles from a video by downloading audio and transcribing with Whisper.
 
         Args:
             url: YouTube video URL
-            lang: Language code for subtitles
+            lang: Language code for transcription
 
         Returns:
             Tuple of (video_id, subtitle_lines)
-
-        Raises:
-            SubtitlesNotFoundError: If no subtitles are found
-            DownloadTimeoutError: If the operation times out
-            SubtitleExtractionError: If extraction fails
         """
         video_id = extract_video_id(url)
         cache_key = f"{video_id}:{lang}"
 
         if cache_key in SUBTITLE_CACHE:
+            logger.info(f"Cache hit for {cache_key}")
             return SUBTITLE_CACHE[cache_key]
 
         with tempfile.TemporaryDirectory() as temp_dir:
-            subtitle_content = await self._download_subtitles(url, lang, temp_dir, video_id)
+            audio_path = await self._download_audio(url, temp_dir, video_id)
 
-            if not subtitle_content:
-                subtitle_content = await self._try_alternative_languages(url, lang, temp_dir, video_id)
+            if not audio_path or not audio_path.exists():
+                raise SubtitleExtractionError("Failed to download audio from video")
 
-            if not subtitle_content:
-                raise SubtitlesNotFoundError(f"No subtitles available in language '{lang}' or alternatives")
+            subtitle_lines = await self._transcribe_audio(audio_path, lang)
 
-            clean_lines = convert_vtt_to_text(subtitle_content)
+            if not subtitle_lines:
+                raise SubtitlesNotFoundError("Transcription produced no text")
 
-            if not clean_lines:
-                raise SubtitlesNotFoundError("Subtitles found but appear to be empty after cleaning")
-
-            result = (video_id, clean_lines)
+            result = (video_id, subtitle_lines)
             SUBTITLE_CACHE[cache_key] = result
             return result
 
-    async def _try_alternative_languages(
-        self, url: str, lang: str, temp_dir: str, video_id: str
-    ) -> Optional[str]:
-        """Try downloading subtitles in alternative language codes concurrently."""
-        alt_langs = ALTERNATIVE_LANGUAGES.get(lang, [f"{lang}-{lang.upper()}"])
-
-        tasks = [
-            self._download_subtitles(url, alt_lang, temp_dir, video_id)
-            for alt_lang in alt_langs
-        ]
-
-        results = await asyncio.gather(*tasks, return_exceptions=True)
-
-        for result in results:
-            if isinstance(result, str) and result:
-                return result
-
-        return None
-
-    async def _download_subtitles(
-        self, url: str, lang: str, temp_dir: str, video_id: str
-    ) -> Optional[str]:
-        """Download subtitles for a specific language."""
-        output_template = str(Path(temp_dir) / f"{video_id}.%(ext)s")
+    async def _download_audio(self, url: str, temp_dir: str, video_id: str) -> Path:
+        """Download audio from video URL using yt-dlp."""
+        output_path = Path(temp_dir) / f"{video_id}.mp3"
 
         cmd = [
             sys.executable, "-m", "yt_dlp",
-            "--write-subs",
-            "--write-auto-subs",
-            "--sub-lang", lang,
-            "--skip-download",
+            "--extract-audio",
+            "--audio-format", "mp3",
+            "--audio-quality", "5",
             "--no-warnings",
-            "--output", output_template,
-            "--user-agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36",
+            "--output", str(Path(temp_dir) / f"{video_id}.%(ext)s"),
             url
         ]
 
         try:
+            logger.info(f"Downloading audio for video: {video_id}")
             process = await asyncio.create_subprocess_exec(
                 *cmd,
                 stdout=asyncio.subprocess.PIPE,
@@ -131,24 +113,56 @@ class SubtitleService:
 
             if process.returncode != 0:
                 error_msg = stderr.decode('utf-8', errors='ignore')
+                logger.error(f"yt-dlp error: {error_msg}")
                 if "Video unavailable" in error_msg or "Private video" in error_msg:
                     raise InvalidVideoURLError("Video is unavailable, private, or does not exist")
-                return None
-
-            temp_path = Path(temp_dir)
-            subtitle_files = list(temp_path.glob(f"{video_id}*.vtt"))
+                raise SubtitleExtractionError(f"Failed to download audio: {error_msg[:200]}")
 
-            if not subtitle_files:
-                return None
+            # Find the downloaded audio file
+            audio_files = list(Path(temp_dir).glob(f"{video_id}.*"))
+            if audio_files:
+                logger.info(f"Audio downloaded: {audio_files[0]}")
+                return audio_files[0]
 
-            return subtitle_files[0].read_text(encoding='utf-8', errors='ignore')
+            raise SubtitleExtractionError("Audio file not found after download")
 
         except asyncio.TimeoutError:
-            raise DownloadTimeoutError(f"Timeout while downloading subtitles for language '{lang}'")
-        except (InvalidVideoURLError, DownloadTimeoutError):
-            raise
-        except Exception as e:
-            raise SubtitleExtractionError(f"Error downloading subtitles: {str(e)}")
+            raise DownloadTimeoutError("Timeout while downloading audio")
+
+    async def _transcribe_audio(self, audio_path: Path, lang: str) -> List[str]:
+        """Transcribe audio file using Whisper."""
+        self._load_whisper_model()
+
+        logger.info(f"Transcribing audio: {audio_path}")
+
+        # Run transcription in thread pool to not block event loop
+        loop = asyncio.get_event_loop()
+        segments = await loop.run_in_executor(
+            None,
+            self._run_transcription,
+            audio_path,
+            lang
+        )
+
+        return segments
+
+    def _run_transcription(self, audio_path: Path, lang: str) -> List[str]:
+        """Run the actual transcription (blocking)."""
+        segments, info = self._whisper_model.transcribe(
+            str(audio_path),
+            language=lang if lang != "auto" else None,
+            beam_size=5,
+            vad_filter=True
+        )
+
+        lines = []
+        for segment in segments:
+            text = segment.text.strip()
+            if text:
+                lines.append(text)
+
+        logger.info(f"Transcription complete: {len(lines)} segments")
+        return lines
 
 
 subtitle_service = SubtitleService()

app/apis/subtitles/utils.py

@@ -4,96 +4,30 @@ import re
 from typing import List
 
 
-def clean_subtitle_text(lines: List[str]) -> List[str]:
-    """
-    Clean subtitle text by removing timestamps, duplicates, and empty lines.
-    
-    Args:
-        lines: Raw subtitle lines
-        
-    Returns:
-        Cleaned subtitle lines
-    """
-    clean_lines = []
-    seen_recently = set()  # Track recent lines to avoid duplicates
-    
-    for line in lines:
-        line = line.strip()
-        
-        # Skip VTT header lines
-        if line.startswith(("WEBVTT", "NOTE")):
-            continue
-            
-        # Skip timestamp lines (format: 00:00:00.000 --> 00:00:00.000)
-        if "-->" in line and re.match(r'\d+:\d+:\d+\.\d+', line):
-            continue
-            
-        # Skip empty lines and sequence numbers
-        if not line or line.isdigit():
-            continue
-        
-        # Remove HTML tags that might be in subtitles
-        line = re.sub(r'<[^>]+>', '', line)
-        
-        # Remove common subtitle formatting
-        line = re.sub(r'^\d+$', '', line)  # Remove standalone numbers
-        line = re.sub(r'^-\s*', '', line)  # Remove leading dashes
-        
-        # Skip if line is too short or just punctuation
-        if len(line.strip()) < 3:
-            continue
-            
-        # Avoid recent duplicates (check last 5 lines)
-        if line not in seen_recently:
-            clean_lines.append(line)
-            seen_recently.add(line)
-            
-            # Keep only last 5 lines in memory for duplicate checking
-            if len(seen_recently) > 5:
-                seen_recently.clear()
-    
-    return clean_lines
-
-
 def extract_video_id(url: str) -> str:
     """
     Extract video ID from YouTube URL.
-    
+
     Args:
         url: YouTube URL
-        
+
     Returns:
-        Video ID
+        Video ID string
     """
-    # Handle different YouTube URL formats
     patterns = [
         r'(?:youtube\.com/watch\?v=|youtu\.be/)([a-zA-Z0-9_-]+)',
         r'youtube\.com/embed/([a-zA-Z0-9_-]+)',
         r'youtube\.com/v/([a-zA-Z0-9_-]+)'
     ]
-    
+
     for pattern in patterns:
         match = re.search(pattern, url)
         if match:
             return match.group(1)
-    
+
     # Fallback: try to extract anything that looks like a video ID
     match = re.search(r'[a-zA-Z0-9_-]{11}', url)
     if match:
         return match.group(0)
-    
-    return "unknown"
 
-
-def convert_vtt_to_text(vtt_content: str) -> List[str]:
-    """
-    Convert VTT subtitle content to clean text lines.
-    
-    Args:
-        vtt_content: Raw VTT file content
-        
-    Returns:
-        Cleaned text lines
-    """
-    lines = vtt_content.split('\n')
-    return clean_subtitle_text(lines)
+    return "unknown"

app/core/config.py

@@ -22,8 +22,10 @@ class Settings(BaseSettings):
     log_level: str = "INFO"
 
     # yt-dlp configuration
-    yt_dlp_timeout_list: int = 30
-    yt_dlp_timeout_download: int = 60
+    yt_dlp_timeout_download: int = 120
+
+    # Whisper configuration
+    whisper_model: str = "base"
 
     # Embedding configuration
     embedding_model: str = "mixedbread-ai/mxbai-embed-large-v1"

poetry.lock

@@ -33,6 +33,66 @@ doc = ["Sphinx", "packaging", "sphinx-autodoc-typehints (>=1.2.0)", "sphinx-rtd-
 test = ["anyio[trio]", "coverage[toml] (>=4.5)", "hypothesis (>=4.0)", "mock (>=4) ; python_version < \"3.8\"", "psutil (>=5.9)", "pytest (>=7.0)", "pytest-mock (>=3.6.1)", "trustme", "uvloop (>=0.17) ; python_version < \"3.12\" and platform_python_implementation == \"CPython\" and platform_system != \"Windows\""]
 trio = ["trio (<0.22)"]
 
+[[package]]
+name = "av"
+version = "16.1.0"
+description = "Pythonic bindings for FFmpeg's libraries."
+optional = false
+python-versions = ">=3.10"
+groups = ["main"]
+files = [
+    {file = "av-16.1.0-cp310-cp310-macosx_11_0_x86_64.whl", hash = "sha256:2395748b0c34fe3a150a1721e4f3d4487b939520991b13e7b36f8926b3b12295"},
+    {file = "av-16.1.0-cp310-cp310-macosx_14_0_arm64.whl", hash = "sha256:72d7ac832710a158eeb7a93242370aa024a7646516291c562ee7f14a7ea881fd"},
+    {file = "av-16.1.0-cp310-cp310-manylinux_2_28_aarch64.whl", hash = "sha256:6cbac833092e66b6b0ac4d81ab077970b8ca874951e9c3974d41d922aaa653ed"},
+    {file = "av-16.1.0-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:eb990672d97c18f99c02f31c8d5750236f770ffe354b5a52c5f4d16c5e65f619"},
+    {file = "av-16.1.0-cp310-cp310-musllinux_1_2_aarch64.whl", hash = "sha256:05ad70933ac3b8ef896a820ea64b33b6cca91a5fac5259cb9ba7fa010435be15"},
+    {file = "av-16.1.0-cp310-cp310-musllinux_1_2_x86_64.whl", hash = "sha256:d831a1062a3c47520bf99de6ec682bd1d64a40dfa958e5457bb613c5270e7ce3"},
+    {file = "av-16.1.0-cp310-cp310-win_amd64.whl", hash = "sha256:358ab910fef3c5a806c55176f2b27e5663b33c4d0a692dafeb049c6ed71f8aff"},
+    {file = "av-16.1.0-cp311-cp311-macosx_11_0_x86_64.whl", hash = "sha256:e88ad64ee9d2b9c4c5d891f16c22ae78e725188b8926eb88187538d9dd0b232f"},
+    {file = "av-16.1.0-cp311-cp311-macosx_14_0_arm64.whl", hash = "sha256:cb296073fa6935724de72593800ba86ae49ed48af03960a4aee34f8a611f442b"},
+    {file = "av-16.1.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:720edd4d25aa73723c1532bb0597806d7b9af5ee34fc02358782c358cfe2f879"},
+    {file = "av-16.1.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:c7f2bc703d0df260a1fdf4de4253c7f5500ca9fc57772ea241b0cb241bcf972e"},
+    {file = "av-16.1.0-cp311-cp311-musllinux_1_2_aarch64.whl", hash = "sha256:d69c393809babada7d54964d56099e4b30a3e1f8b5736ca5e27bd7be0e0f3c83"},
+    {file = "av-16.1.0-cp311-cp311-musllinux_1_2_x86_64.whl", hash = "sha256:441892be28582356d53f282873c5a951592daaf71642c7f20165e3ddcb0b4c63"},
+    {file = "av-16.1.0-cp311-cp311-win_amd64.whl", hash = "sha256:273a3e32de64819e4a1cd96341824299fe06f70c46f2288b5dc4173944f0fd62"},
+    {file = "av-16.1.0-cp312-cp312-macosx_11_0_x86_64.whl", hash = "sha256:640f57b93f927fba8689f6966c956737ee95388a91bd0b8c8b5e0481f73513d6"},
+    {file = "av-16.1.0-cp312-cp312-macosx_14_0_arm64.whl", hash = "sha256:ae3fb658eec00852ebd7412fdc141f17f3ddce8afee2d2e1cf366263ad2a3b35"},
+    {file = "av-16.1.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:27ee558d9c02a142eebcbe55578a6d817fedfde42ff5676275504e16d07a7f86"},
+    {file = "av-16.1.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:7ae547f6d5fa31763f73900d43901e8c5fa6367bb9a9840978d57b5a7ae14ed2"},
+    {file = "av-16.1.0-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:8cf065f9d438e1921dc31fc7aa045790b58aee71736897866420d80b5450f62a"},
+    {file = "av-16.1.0-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:a345877a9d3cc0f08e2bc4ec163ee83176864b92587afb9d08dff50f37a9a829"},
+    {file = "av-16.1.0-cp312-cp312-win_amd64.whl", hash = "sha256:f49243b1d27c91cd8c66fdba90a674e344eb8eb917264f36117bf2b6879118fd"},
+    {file = "av-16.1.0-cp313-cp313-macosx_11_0_x86_64.whl", hash = "sha256:ce2a1b3d8bf619f6c47a9f28cfa7518ff75ddd516c234a4ee351037b05e6a587"},
+    {file = "av-16.1.0-cp313-cp313-macosx_14_0_arm64.whl", hash = "sha256:408dbe6a2573ca58a855eb8cd854112b33ea598651902c36709f5f84c991ed8e"},
+    {file = "av-16.1.0-cp313-cp313-manylinux_2_28_aarch64.whl", hash = "sha256:57f657f86652a160a8a01887aaab82282f9e629abf94c780bbdbb01595d6f0f7"},
+    {file = "av-16.1.0-cp313-cp313-manylinux_2_28_x86_64.whl", hash = "sha256:adbad2b355c2ee4552cac59762809d791bda90586d134a33c6f13727fb86cb3a"},
+    {file = "av-16.1.0-cp313-cp313-musllinux_1_2_aarch64.whl", hash = "sha256:f42e1a68ec2aebd21f7eb6895be69efa6aa27eec1670536876399725bbda4b99"},
+    {file = "av-16.1.0-cp313-cp313-musllinux_1_2_x86_64.whl", hash = "sha256:58fe47aeaef0f100c40ec8a5de9abbd37f118d3ca03829a1009cf288e9aef67c"},
+    {file = "av-16.1.0-cp313-cp313-win_amd64.whl", hash = "sha256:565093ebc93b2f4b76782589564869dadfa83af5b852edebedd8fee746457d06"},
+    {file = "av-16.1.0-cp313-cp313t-macosx_11_0_x86_64.whl", hash = "sha256:574081a24edb98343fd9f473e21ae155bf61443d4ec9d7708987fa597d6b04b2"},
+    {file = "av-16.1.0-cp313-cp313t-macosx_14_0_arm64.whl", hash = "sha256:9ab00ea29c25ebf2ea1d1e928d7babb3532d562481c5d96c0829212b70756ad0"},
+    {file = "av-16.1.0-cp313-cp313t-manylinux_2_28_aarch64.whl", hash = "sha256:a84a91188c1071f238a9523fd42dbe567fb2e2607b22b779851b2ce0eac1b560"},
+    {file = "av-16.1.0-cp313-cp313t-manylinux_2_28_x86_64.whl", hash = "sha256:c2cd0de4dd022a7225ff224fde8e7971496d700be41c50adaaa26c07bb50bf97"},
+    {file = "av-16.1.0-cp313-cp313t-musllinux_1_2_aarch64.whl", hash = "sha256:0816143530624a5a93bc5494f8c6eeaf77549b9366709c2ac8566c1e9bff6df5"},
+    {file = "av-16.1.0-cp313-cp313t-musllinux_1_2_x86_64.whl", hash = "sha256:e3a28053af29644696d0c007e897d19b1197585834660a54773e12a40b16974c"},
+    {file = "av-16.1.0-cp313-cp313t-win_amd64.whl", hash = "sha256:2e3e67144a202b95ed299d165232533989390a9ea3119d37eccec697dc6dbb0c"},
+    {file = "av-16.1.0-cp314-cp314-macosx_11_0_x86_64.whl", hash = "sha256:39a634d8e5a87e78ea80772774bfd20c0721f0d633837ff185f36c9d14ffede4"},
+    {file = "av-16.1.0-cp314-cp314-macosx_14_0_arm64.whl", hash = "sha256:0ba32fb9e9300948a7fa9f8a3fc686e6f7f77599a665c71eb2118fdfd2c743f9"},
+    {file = "av-16.1.0-cp314-cp314-manylinux_2_28_aarch64.whl", hash = "sha256:ca04d17815182d34ce3edc53cbda78a4f36e956c0fd73e3bab249872a831c4d7"},
+    {file = "av-16.1.0-cp314-cp314-manylinux_2_28_x86_64.whl", hash = "sha256:ee0e8de2e124a9ef53c955fe2add6ee7c56cc8fd83318265549e44057db77142"},
+    {file = "av-16.1.0-cp314-cp314-musllinux_1_2_aarch64.whl", hash = "sha256:22bf77a2f658827043a1e184b479c3bf25c4c43ab32353677df2d119f080e28f"},
+    {file = "av-16.1.0-cp314-cp314-musllinux_1_2_x86_64.whl", hash = "sha256:2dd419d262e6a71cab206d80bbf28e0a10d0f227b671cdf5e854c028faa2d043"},
+    {file = "av-16.1.0-cp314-cp314-win_amd64.whl", hash = "sha256:53585986fd431cd436f290fba662cfb44d9494fbc2949a183de00acc5b33fa88"},
+    {file = "av-16.1.0-cp314-cp314t-macosx_11_0_x86_64.whl", hash = "sha256:76f5ed8495cf41e1209a5775d3699dc63fdc1740b94a095e2485f13586593205"},
+    {file = "av-16.1.0-cp314-cp314t-macosx_14_0_arm64.whl", hash = "sha256:8d55397190f12a1a3ae7538be58c356cceb2bf50df1b33523817587748ce89e5"},
+    {file = "av-16.1.0-cp314-cp314t-manylinux_2_28_aarch64.whl", hash = "sha256:9d51d9037437218261b4bbf9df78a95e216f83d7774fbfe8d289230b5b2e28e2"},
+    {file = "av-16.1.0-cp314-cp314t-manylinux_2_28_x86_64.whl", hash = "sha256:0ce07a89c15644407f49d942111ca046e323bbab0a9078ff43ee57c9b4a50dad"},
+    {file = "av-16.1.0-cp314-cp314t-musllinux_1_2_aarch64.whl", hash = "sha256:cac0c074892ea97113b53556ff41c99562db7b9f09f098adac1f08318c2acad5"},
+    {file = "av-16.1.0-cp314-cp314t-musllinux_1_2_x86_64.whl", hash = "sha256:7dec3dcbc35a187ce450f65a2e0dda820d5a9e6553eea8344a1459af11c98649"},
+    {file = "av-16.1.0-cp314-cp314t-win_amd64.whl", hash = "sha256:6f90dc082ff2068ddbe77618400b44d698d25d9c4edac57459e250c16b33d700"},
+    {file = "av-16.1.0.tar.gz", hash = "sha256:a094b4fd87a3721dacf02794d3d2c82b8d712c85b9534437e82a8a978c175ffd"},
+]
+
 [[package]]
 name = "black"
 version = "23.12.1"
@@ -253,6 +313,24 @@ files = [
     {file = "colorama-0.4.6.tar.gz", hash = "sha256:08695f5cb7ed6e0531a20572697297273c47b8cae5a63ffc6d6ed5c201be6e44"},
 ]
 
+[[package]]
+name = "coloredlogs"
+version = "15.0.1"
+description = "Colored terminal output for Python's logging module"
+optional = false
+python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*"
+groups = ["main"]
+files = [
+    {file = "coloredlogs-15.0.1-py2.py3-none-any.whl", hash = "sha256:612ee75c546f53e92e70049c9dbfcc18c935a2b9a53b66085ce9ef6a6e5c0934"},
+    {file = "coloredlogs-15.0.1.tar.gz", hash = "sha256:7c991aa71a4577af2f82600d8f8f3a89f936baeaf9b50a9c197da014e5bf16b0"},
+]
+
+[package.dependencies]
+humanfriendly = ">=9.1"
+
+[package.extras]
+cron = ["capturer (>=2.4)"]
+
 [[package]]
 name = "coverage"
 version = "7.10.6"
@@ -354,6 +432,56 @@ files = [
 [package.extras]
 toml = ["tomli ; python_full_version <= \"3.11.0a6\""]
 
+[[package]]
+name = "ctranslate2"
+version = "4.6.3"
+description = "Fast inference engine for Transformer models"
+optional = false
+python-versions = ">=3.9"
+groups = ["main"]
+files = [
+    {file = "ctranslate2-4.6.3-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:d75d79e55a3a26964320445c03a56af60d7215d95561b744d93d04bad24c268a"},
+    {file = "ctranslate2-4.6.3-cp310-cp310-macosx_11_0_x86_64.whl", hash = "sha256:13ccb5011e67b831354c9a01bf4d824b4dc5535c54abcf492e0ae4e41894518e"},
+    {file = "ctranslate2-4.6.3-cp310-cp310-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:259ab216d4de93723f3db1805f2bac48b1a5732ce3de0e5a163b570821fcb063"},
+    {file = "ctranslate2-4.6.3-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:7a5e59a5a67c3f48133ffe6fe2a557922283c16eb4233e6dbb82e0b9a20782f2"},
+    {file = "ctranslate2-4.6.3-cp310-cp310-win_amd64.whl", hash = "sha256:6be735c7904ea98c22d7d02b338299c0a7f4cd4b1d0e9dd528e319e52bd78d66"},
+    {file = "ctranslate2-4.6.3-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:1ac0d2bec0961f0f9ee00cd5c55b4d5904ee309d9269778d9f9edd23c46c87ff"},
+    {file = "ctranslate2-4.6.3-cp311-cp311-macosx_11_0_x86_64.whl", hash = "sha256:db5f82661fa960a6a1bc0e738acf135a22da94a32cda198d8fb782d37ef4caa8"},
+    {file = "ctranslate2-4.6.3-cp311-cp311-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:5f1ec2cd9546f02ff9f1b2d21b115eadcce45c8ae5ac5811e7d382f9d9736aa4"},
+    {file = "ctranslate2-4.6.3-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:67f4b5802349a8cfa2e6105b161bf015e97aadab0f58a7034c97e78283cb29b8"},
+    {file = "ctranslate2-4.6.3-cp311-cp311-win_amd64.whl", hash = "sha256:fa2f3dcda893a3f4dedeb32b5059e4085738934d93ea8dccdce4bbef2be5d3dc"},
+    {file = "ctranslate2-4.6.3-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:32022dcf0ee2eace0b00345899b0e2be2f5a8b57d8467b1f5ecee40bb3e18746"},
+    {file = "ctranslate2-4.6.3-cp312-cp312-macosx_11_0_x86_64.whl", hash = "sha256:df88e7ac821b2def12ae6c71ba4180c13abc13713c1d1ae819e92f2db8556564"},
+    {file = "ctranslate2-4.6.3-cp312-cp312-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:487f57da179057e1a8498d3b61f2fcd826ddfe989ce43ff3b500ec805ca55d56"},
+    {file = "ctranslate2-4.6.3-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:a857a42b091f9e0b8b1f63cf1fb356822bb4905d555039f542ff95cf90fd592b"},
+    {file = "ctranslate2-4.6.3-cp312-cp312-win_amd64.whl", hash = "sha256:05ec48b44bb2f1e623e30acc57d34d22000d969e8998cae7762137231fae0d25"},
+    {file = "ctranslate2-4.6.3-cp313-cp313-macosx_11_0_arm64.whl", hash = "sha256:95ff7fdd70bd64d40834cb6ba82bcec15228a9f34dff587babd03a1c3064c302"},
+    {file = "ctranslate2-4.6.3-cp313-cp313-macosx_11_0_x86_64.whl", hash = "sha256:a562ef2fd48287423dd6158a0c7921b6c238a052f690bce510b998bba82fd3e2"},
+    {file = "ctranslate2-4.6.3-cp313-cp313-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:6cc539ed7c3531354971c78938da50f29ac08b8dc9140bc7ac377e8344bc63e2"},
+    {file = "ctranslate2-4.6.3-cp313-cp313-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:f08efa826707d095ade28410dca27f8d377520f3068843e00b349d5ca15cf174"},
+    {file = "ctranslate2-4.6.3-cp313-cp313-win_amd64.whl", hash = "sha256:9a6b6e80d79242761d0583bc0ad7e7ba4d09745d2b23e814bc35f6c842b0ca45"},
+    {file = "ctranslate2-4.6.3-cp314-cp314-macosx_11_0_arm64.whl", hash = "sha256:75f3e9d3ca7b3d91c87f67972f20998fc318a22d49c25b6d7144b947b5e3240e"},
+    {file = "ctranslate2-4.6.3-cp314-cp314-macosx_11_0_x86_64.whl", hash = "sha256:a0657885219e05a6575bb9d8ac4c055da25110d6c897dfed7a322f8c01267fb1"},
+    {file = "ctranslate2-4.6.3-cp314-cp314-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:53e975acf49bab2cd00290a2ece56925d087f8300d5bd7463b96c60002146034"},
+    {file = "ctranslate2-4.6.3-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:e411c7212f42899f12522b4d9a4b5a59542aa27d5b8e87e7e7bd2f52194fa984"},
+    {file = "ctranslate2-4.6.3-cp314-cp314-win_amd64.whl", hash = "sha256:40749b5ad208eb5224ea7ec9516ff290e77373974be0f41697eccf3cef2a44eb"},
+    {file = "ctranslate2-4.6.3-cp314-cp314t-macosx_11_0_arm64.whl", hash = "sha256:dd117643e9bae19d53e3fea4415862841c4e69fcff86dbc4dd397f6864390d84"},
+    {file = "ctranslate2-4.6.3-cp314-cp314t-macosx_11_0_x86_64.whl", hash = "sha256:e058b51372faee95780c0d0af513e7c5df268fffcd435a856476d998e65ebf67"},
+    {file = "ctranslate2-4.6.3-cp314-cp314t-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:4eca886e30e658bece2bd0fc331a37f4a5ad1e29a590d43d5082c7896eba59d7"},
+    {file = "ctranslate2-4.6.3-cp314-cp314t-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:5345d0d259383ddc106343744be5ada9646f0e2632a6676482fd9de6114c9ee2"},
+    {file = "ctranslate2-4.6.3-cp314-cp314t-win_amd64.whl", hash = "sha256:53ab04edc3f7280465cd54e6a359f26960eb63961eeae27cb9726f449b4b217e"},
+    {file = "ctranslate2-4.6.3-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:3f229dfcd14be23c4f76542873562ab9d8006e6e045fa585be83f82d224c30be"},
+    {file = "ctranslate2-4.6.3-cp39-cp39-macosx_11_0_x86_64.whl", hash = "sha256:11f88fe0d2d081f1fc4f7442477a7089a3fac9ad28c98fa2df1d9739a114524e"},
+    {file = "ctranslate2-4.6.3-cp39-cp39-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:8d0b028e161a8374467a7b77a4675a0aa88cd2dd24e0700c8277418cc31be4d9"},
+    {file = "ctranslate2-4.6.3-cp39-cp39-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:a7d0c858a4fecf288211b488ed281c00b93b77155e39e6d496646cc1ddbecda1"},
+    {file = "ctranslate2-4.6.3-cp39-cp39-win_amd64.whl", hash = "sha256:875fd4dcd7185589f07197b667ce547ec532ef42d68feed76615543e87f289ef"},
+]
+
+[package.dependencies]
+numpy = "*"
+pyyaml = ">=5.3,<7"
+setuptools = "*"
+
 [[package]]
 name = "deprecated"
 version = "1.3.1"
@@ -393,6 +521,29 @@ typing-extensions = ">=4.8.0"
 [package.extras]
 all = ["email-validator (>=2.0.0)", "httpx (>=0.23.0)", "itsdangerous (>=1.1.0)", "jinja2 (>=2.11.2)", "orjson (>=3.2.1)", "pydantic-extra-types (>=2.0.0)", "pydantic-settings (>=2.0.0)", "python-multipart (>=0.0.5)", "pyyaml (>=5.3.1)", "ujson (>=4.0.1,!=4.0.2,!=4.1.0,!=4.2.0,!=4.3.0,!=5.0.0,!=5.1.0)", "uvicorn[standard] (>=0.12.0)"]
 
+[[package]]
+name = "faster-whisper"
+version = "1.2.1"
+description = "Faster Whisper transcription with CTranslate2"
+optional = false
+python-versions = ">=3.9"
+groups = ["main"]
+files = [
+    {file = "faster_whisper-1.2.1-py3-none-any.whl", hash = "sha256:79a66ad50688c0b794dd501dc340a736992a6342f7f95e5811be60b5224a26a7"},
+]
+
+[package.dependencies]
+av = ">=11"
+ctranslate2 = ">=4.0,<5"
+huggingface-hub = ">=0.21"
+onnxruntime = ">=1.14,<2"
+tokenizers = ">=0.13,<1"
+tqdm = "*"
+
+[package.extras]
+conversion = ["transformers[torch] (>=4.23)"]
+dev = ["black (==23.*)", "flake8 (==6.*)", "isort (==5.*)", "pytest (==7.*)"]
+
 [[package]]
 name = "filelock"
 version = "3.20.3"
@@ -422,6 +573,17 @@ mccabe = ">=0.7.0,<0.8.0"
 pycodestyle = ">=2.11.0,<2.12.0"
 pyflakes = ">=3.1.0,<3.2.0"
 
+[[package]]
+name = "flatbuffers"
+version = "25.12.19"
+description = "The FlatBuffers serialization format for Python"
+optional = false
+python-versions = "*"
+groups = ["main"]
+files = [
+    {file = "flatbuffers-25.12.19-py2.py3-none-any.whl", hash = "sha256:7634f50c427838bb021c2d66a3d1168e9d199b0607e6329399f04846d42e20b4"},
+]
+
 [[package]]
 name = "fsspec"
 version = "2026.1.0"
@@ -652,6 +814,21 @@ testing = ["InquirerPy (==0.3.4)", "Jinja2", "Pillow", "aiohttp", "authlib (>=1.
 torch = ["safetensors[torch]", "torch"]
 typing = ["types-PyYAML", "types-requests", "types-simplejson", "types-toml", "types-tqdm", "types-urllib3", "typing-extensions (>=4.8.0)"]
 
+[[package]]
+name = "humanfriendly"
+version = "10.0"
+description = "Human friendly output for text interfaces using Python"
+optional = false
+python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*"
+groups = ["main"]
+files = [
+    {file = "humanfriendly-10.0-py2.py3-none-any.whl", hash = "sha256:1697e1a8a8f550fd43c2865cd84542fc175a61dcb779b6fee18cf6b6ccba1477"},
+    {file = "humanfriendly-10.0.tar.gz", hash = "sha256:6b0b831ce8f15f7300721aa49829fc4e83921a9a301cc7f606be6686a2288ddc"},
+]
+
+[package.dependencies]
+pyreadline3 = {version = "*", markers = "sys_platform == \"win32\" and python_version >= \"3.8\""}
+
 [[package]]
 name = "idna"
 version = "3.10"
@@ -1299,6 +1476,46 @@ files = [
     {file = "nvidia_nvtx_cu12-12.6.77-py3-none-win_amd64.whl", hash = "sha256:2fb11a4af04a5e6c84073e6404d26588a34afd35379f0855a99797897efa75c0"},
 ]
 
+[[package]]
+name = "onnxruntime"
+version = "1.23.2"
+description = "ONNX Runtime is a runtime accelerator for Machine Learning models"
+optional = false
+python-versions = ">=3.10"
+groups = ["main"]
+files = [
+    {file = "onnxruntime-1.23.2-cp310-cp310-macosx_13_0_arm64.whl", hash = "sha256:a7730122afe186a784660f6ec5807138bf9d792fa1df76556b27307ea9ebcbe3"},
+    {file = "onnxruntime-1.23.2-cp310-cp310-macosx_13_0_x86_64.whl", hash = "sha256:b28740f4ecef1738ea8f807461dd541b8287d5650b5be33bca7b474e3cbd1f36"},
+    {file = "onnxruntime-1.23.2-cp310-cp310-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:8f7d1fe034090a1e371b7f3ca9d3ccae2fabae8c1d8844fb7371d1ea38e8e8d2"},
+    {file = "onnxruntime-1.23.2-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:4ca88747e708e5c67337b0f65eed4b7d0dd70d22ac332038c9fc4635760018f7"},
+    {file = "onnxruntime-1.23.2-cp310-cp310-win_amd64.whl", hash = "sha256:0be6a37a45e6719db5120e9986fcd30ea205ac8103fd1fb74b6c33348327a0cc"},
+    {file = "onnxruntime-1.23.2-cp311-cp311-macosx_13_0_arm64.whl", hash = "sha256:6f91d2c9b0965e86827a5ba01531d5b669770b01775b23199565d6c1f136616c"},
+    {file = "onnxruntime-1.23.2-cp311-cp311-macosx_13_0_x86_64.whl", hash = "sha256:87d8b6eaf0fbeb6835a60a4265fde7a3b60157cf1b2764773ac47237b4d48612"},
+    {file = "onnxruntime-1.23.2-cp311-cp311-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:bbfd2fca76c855317568c1b36a885ddea2272c13cb0e395002c402f2360429a6"},
+    {file = "onnxruntime-1.23.2-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:da44b99206e77734c5819aa2142c69e64f3b46edc3bd314f6a45a932defc0b3e"},
+    {file = "onnxruntime-1.23.2-cp311-cp311-win_amd64.whl", hash = "sha256:902c756d8b633ce0dedd889b7c08459433fbcf35e9c38d1c03ddc020f0648c6e"},
+    {file = "onnxruntime-1.23.2-cp312-cp312-macosx_13_0_arm64.whl", hash = "sha256:b8f029a6b98d3cf5be564d52802bb50a8489ab73409fa9db0bf583eabb7c2321"},
+    {file = "onnxruntime-1.23.2-cp312-cp312-macosx_13_0_x86_64.whl", hash = "sha256:218295a8acae83905f6f1aed8cacb8e3eb3bd7513a13fe4ba3b2664a19fc4a6b"},
+    {file = "onnxruntime-1.23.2-cp312-cp312-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:76ff670550dc23e58ea9bc53b5149b99a44e63b34b524f7b8547469aaa0dcb8c"},
+    {file = "onnxruntime-1.23.2-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:0f9b4ae77f8e3c9bee50c27bc1beede83f786fe1d52e99ac85aa8d65a01e9b77"},
+    {file = "onnxruntime-1.23.2-cp312-cp312-win_amd64.whl", hash = "sha256:25de5214923ce941a3523739d34a520aac30f21e631de53bba9174dc9c004435"},
+    {file = "onnxruntime-1.23.2-cp313-cp313-macosx_13_0_arm64.whl", hash = "sha256:2ff531ad8496281b4297f32b83b01cdd719617e2351ffe0dba5684fb283afa1f"},
+    {file = "onnxruntime-1.23.2-cp313-cp313-macosx_13_0_x86_64.whl", hash = "sha256:162f4ca894ec3de1a6fd53589e511e06ecdc3ff646849b62a9da7489dee9ce95"},
+    {file = "onnxruntime-1.23.2-cp313-cp313-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:45d127d6e1e9b99d1ebeae9bcd8f98617a812f53f46699eafeb976275744826b"},
+    {file = "onnxruntime-1.23.2-cp313-cp313-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:8bace4e0d46480fbeeb7bbe1ffe1f080e6663a42d1086ff95c1551f2d39e7872"},
+    {file = "onnxruntime-1.23.2-cp313-cp313-win_amd64.whl", hash = "sha256:1f9cc0a55349c584f083c1c076e611a7c35d5b867d5d6e6d6c823bf821978088"},
+    {file = "onnxruntime-1.23.2-cp313-cp313t-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:9d2385e774f46ac38f02b3a91a91e30263d41b2f1f4f26ae34805b2a9ddef466"},
+    {file = "onnxruntime-1.23.2-cp313-cp313t-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:e2b9233c4947907fd1818d0e581c049c41ccc39b2856cc942ff6d26317cee145"},
+]
+
+[package.dependencies]
+coloredlogs = "*"
+flatbuffers = "*"
+numpy = ">=1.21.6"
+packaging = "*"
+protobuf = "*"
+sympy = "*"
+
 [[package]]
 name = "packaging"
 version = "25.0"
@@ -1465,6 +1682,26 @@ files = [
 dev = ["pre-commit", "tox"]
 testing = ["coverage", "pytest", "pytest-benchmark"]
 
+[[package]]
+name = "protobuf"
+version = "6.33.4"
+description = ""
+optional = false
+python-versions = ">=3.9"
+groups = ["main"]
+files = [
+    {file = "protobuf-6.33.4-cp310-abi3-win32.whl", hash = "sha256:918966612c8232fc6c24c78e1cd89784307f5814ad7506c308ee3cf86662850d"},
+    {file = "protobuf-6.33.4-cp310-abi3-win_amd64.whl", hash = "sha256:8f11ffae31ec67fc2554c2ef891dcb561dae9a2a3ed941f9e134c2db06657dbc"},
+    {file = "protobuf-6.33.4-cp39-abi3-macosx_10_9_universal2.whl", hash = "sha256:2fe67f6c014c84f655ee06f6f66213f9254b3a8b6bda6cda0ccd4232c73c06f0"},
+    {file = "protobuf-6.33.4-cp39-abi3-manylinux2014_aarch64.whl", hash = "sha256:757c978f82e74d75cba88eddec479df9b99a42b31193313b75e492c06a51764e"},
+    {file = "protobuf-6.33.4-cp39-abi3-manylinux2014_s390x.whl", hash = "sha256:c7c64f259c618f0bef7bee042075e390debbf9682334be2b67408ec7c1c09ee6"},
+    {file = "protobuf-6.33.4-cp39-abi3-manylinux2014_x86_64.whl", hash = "sha256:3df850c2f8db9934de4cf8f9152f8dc2558f49f298f37f90c517e8e5c84c30e9"},
+    {file = "protobuf-6.33.4-cp39-cp39-win32.whl", hash = "sha256:955478a89559fa4568f5a81dce77260eabc5c686f9e8366219ebd30debf06aa6"},
+    {file = "protobuf-6.33.4-cp39-cp39-win_amd64.whl", hash = "sha256:0f12ddbf96912690c3582f9dffb55530ef32015ad8e678cd494312bd78314c4f"},
+    {file = "protobuf-6.33.4-py3-none-any.whl", hash = "sha256:1fe3730068fcf2e595816a6c34fe66eeedd37d51d0400b72fabc848811fdc1bc"},
+    {file = "protobuf-6.33.4.tar.gz", hash = "sha256:dc2e61bca3b10470c1912d166fe0af67bfc20eb55971dcef8dfa48ce14f0ed91"},
+]
+
 [[package]]
 name = "pycodestyle"
 version = "2.11.1"
@@ -1647,6 +1884,22 @@ files = [
     {file = "pyflakes-3.1.0.tar.gz", hash = "sha256:a0aae034c444db0071aa077972ba4768d40c830d9539fd45bf4cd3f8f6992efc"},
 ]
 
+[[package]]
+name = "pyreadline3"
+version = "3.5.4"
+description = "A python implementation of GNU readline."
+optional = false
+python-versions = ">=3.8"
+groups = ["main"]
+markers = "sys_platform == \"win32\""
+files = [
+    {file = "pyreadline3-3.5.4-py3-none-any.whl", hash = "sha256:eaf8e6cc3c49bcccf145fc6067ba8643d1df34d604a1ec0eccbf7a18e6d3fae6"},
+    {file = "pyreadline3-3.5.4.tar.gz", hash = "sha256:8d57d53039a1c75adba8e50dd3d992b28143480816187ea5efbd5c78e6c885b7"},
+]
+
+[package.extras]
+dev = ["build", "flake8", "mypy", "pytest", "twine"]
+
 [[package]]
 name = "pytest"
 version = "7.4.4"
@@ -2152,7 +2405,6 @@ description = "Easily download, build, install, upgrade, and uninstall Python pa
 optional = false
 python-versions = ">=3.9"
 groups = ["main"]
-markers = "platform_system == \"Linux\" and platform_machine == \"x86_64\" or python_version >= \"3.12\""
 files = [
     {file = "setuptools-80.9.0-py3-none-any.whl", hash = "sha256:062d34222ad13e0cc312a4c02d73f059e86a4acbfbdea8f8f76b28c99f306922"},
     {file = "setuptools-80.9.0.tar.gz", hash = "sha256:f36b47402ecde768dbfafc46e8e4207b4360c654f1f3bb84475f0a28628fb19c"},
@@ -2954,4 +3206,4 @@ test = ["pytest (>=8.1,<9.0)", "pytest-rerunfailures (>=14.0,<15.0)"]
 [metadata]
 lock-version = "2.1"
 python-versions = "^3.11"
-content-hash = "c179a0fd91b7ec088e821057b465cd5ca717460f85d0d55e3f481456810881c2"
+content-hash = "ebb16163670f970e4d506a8230fbb256edc8d5a749255bbbc9dc9cd17fc870eb"

pyproject.toml

@@ -19,6 +19,7 @@ slowapi = "^0.1.9"
 cachetools = "^5.3.0"
 sentence-transformers = "^2.2.2"
 torch = "^2.0.0"
+faster-whisper = "^1.0.0"
 
 [tool.poetry.group.dev.dependencies]
 pytest = "^7.4.3"

tests/conftest.py

@@ -46,30 +46,6 @@ def temp_dir():
         yield tmp_dir
 
 
-@pytest.fixture
-def sample_vtt_content():
-    """Sample VTT subtitle content for testing."""
-    return """WEBVTT
-Kind: captions
-Language: en
-
-00:00:00.000 --> 00:00:03.000
-Never gonna give you up
-
-00:00:03.000 --> 00:00:06.000
-Never gonna let you down
-
-00:00:06.000 --> 00:00:09.000
-Never gonna run around and desert you
-
-00:00:09.000 --> 00:00:12.000
-Never gonna make you cry
-
-00:00:12.000 --> 00:00:15.000
-Never gonna say goodbye
-"""
-
-
 @pytest.fixture
 def sample_youtube_url():
     """Sample YouTube URL for testing."""

tests/test_subtitles.py

@@ -1,19 +1,11 @@
 """Tests for subtitle extraction functionality."""
 
 import pytest
-from unittest.mock import AsyncMock, patch
-import asyncio
+from unittest.mock import patch, MagicMock
 
 from app.apis.subtitles.service import SubtitleService, SUBTITLE_CACHE
-from app.apis.subtitles.utils import (
-    clean_subtitle_text,
-    extract_video_id,
-    convert_vtt_to_text
-)
-from app.core.exceptions import (
-    SubtitlesNotFoundError,
-    DownloadTimeoutError
-)
+from app.apis.subtitles.utils import extract_video_id
+from app.core.exceptions import SubtitlesNotFoundError
 
 
 class TestSubtitleUtils:
@@ -34,38 +26,6 @@ class TestSubtitleUtils:
         url = "https://www.youtube.com/embed/dQw4w9WgXcQ"
         assert extract_video_id(url) == "dQw4w9WgXcQ"
 
-    def test_clean_subtitle_text(self):
-        """Test cleaning subtitle text."""
-        raw_lines = [
-            "WEBVTT",
-            "",
-            "1",
-            "00:00:00.000 --> 00:00:03.000",
-            "Never gonna give you up",
-            "",
-            "2",
-            "00:00:03.000 --> 00:00:06.000",
-            "Never gonna let you down",
-            "Never gonna give you up",
-            ""
-        ]
-
-        cleaned = clean_subtitle_text(raw_lines)
-        assert "Never gonna give you up" in cleaned
-        assert "Never gonna let you down" in cleaned
-        assert "WEBVTT" not in cleaned
-        assert "00:00:00.000 --> 00:00:03.000" not in cleaned
-        assert len([line for line in cleaned if line == "Never gonna give you up"]) == 1
-
-    def test_convert_vtt_to_text(self, sample_vtt_content):
-        """Test converting VTT content to clean text."""
-        result = convert_vtt_to_text(sample_vtt_content)
-
-        assert "Never gonna give you up" in result
-        assert "Never gonna let you down" in result
-        assert "WEBVTT" not in result
-        assert "00:00:00.000 --> 00:00:03.000" not in result
-
 
 class TestSubtitleService:
     """Test subtitle extraction service."""
@@ -81,33 +41,28 @@ class TestSubtitleService:
         return SubtitleService()
 
     @pytest.mark.asyncio
-    async def test_extract_subtitles_success(self, service, sample_youtube_url, sample_vtt_content):
+    async def test_extract_subtitles_success(self, service, sample_youtube_url):
         """Test successful subtitle extraction."""
-        with patch.object(service, '_download_subtitles') as mock_download:
-            mock_download.return_value = sample_vtt_content
+        with patch.object(service, '_download_audio') as mock_download, \
+             patch.object(service, '_transcribe_audio') as mock_transcribe:
+            mock_download.return_value = MagicMock()
+            mock_download.return_value.exists.return_value = True
+            mock_transcribe.return_value = ["Test subtitle line 1", "Test subtitle line 2"]
 
             video_id, subtitles = await service.extract_subtitles(sample_youtube_url, "en")
 
             assert video_id == "dQw4w9WgXcQ"
-            assert len(subtitles) > 0
-            assert "Never gonna give you up" in subtitles
-
-    @pytest.mark.asyncio
-    async def test_extract_subtitles_not_found(self, service, sample_youtube_url):
-        """Test subtitle extraction when no subtitles are found."""
-        with patch.object(service, '_download_subtitles') as mock_download, \
-             patch.object(service, '_try_alternative_languages') as mock_alt:
-            mock_download.return_value = None
-            mock_alt.return_value = None
-
-            with pytest.raises(SubtitlesNotFoundError):
-                await service.extract_subtitles(sample_youtube_url, "en")
+            assert len(subtitles) == 2
+            assert "Test subtitle line 1" in subtitles
 
     @pytest.mark.asyncio
-    async def test_extract_subtitles_uses_cache(self, service, sample_youtube_url, sample_vtt_content):
+    async def test_extract_subtitles_uses_cache(self, service, sample_youtube_url):
         """Test that cached results are returned."""
-        with patch.object(service, '_download_subtitles') as mock_download:
-            mock_download.return_value = sample_vtt_content
+        with patch.object(service, '_download_audio') as mock_download, \
+             patch.object(service, '_transcribe_audio') as mock_transcribe:
+            mock_download.return_value = MagicMock()
+            mock_download.return_value.exists.return_value = True
+            mock_transcribe.return_value = ["Cached subtitle"]
 
             result1 = await service.extract_subtitles(sample_youtube_url, "en")
             result2 = await service.extract_subtitles(sample_youtube_url, "en")
@@ -115,14 +70,26 @@ class TestSubtitleService:
             assert result1 == result2
             assert mock_download.call_count == 1
 
+    @pytest.mark.asyncio
+    async def test_extract_subtitles_empty_transcription(self, service, sample_youtube_url):
+        """Test error when transcription produces no text."""
+        with patch.object(service, '_download_audio') as mock_download, \
+             patch.object(service, '_transcribe_audio') as mock_transcribe:
+            mock_download.return_value = MagicMock()
+            mock_download.return_value.exists.return_value = True
+            mock_transcribe.return_value = []
+
+            with pytest.raises(SubtitlesNotFoundError):
+                await service.extract_subtitles(sample_youtube_url, "en")
+
 
 class TestSubtitleAPI:
     """Test subtitle API endpoints."""
 
-    def test_extract_subtitles_endpoint_success(self, client, api_key, sample_vtt_content):
+    def test_extract_subtitles_endpoint_success(self, client, api_key):
         """Test successful subtitle extraction via API."""
         with patch('app.apis.subtitles.service.subtitle_service.extract_subtitles') as mock_extract:
-            mock_extract.return_value = ("dQw4w9WgXcQ", ["Never gonna give you up", "Never gonna let you down"])
+            mock_extract.return_value = ("dQw4w9WgXcQ", ["Never gonna give you up"])
 
             response = client.post(
                 "/api/v1/subtitles/extract",
@@ -134,7 +101,6 @@ class TestSubtitleAPI:
             data = response.json()
             assert data["status"] == "success"
             assert data["video_id"] == "dQw4w9WgXcQ"
-            assert "Never gonna give you up" in data["subtitles"]
 
     def test_extract_subtitles_endpoint_invalid_api_key(self, client, invalid_api_key):
         """Test API endpoint with invalid API key."""
@@ -143,7 +109,6 @@ class TestSubtitleAPI:
             json={"url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ", "lang": "en"},
             headers={"x-api-key": invalid_api_key}
         )
-
         assert response.status_code == 401
 
     def test_extract_subtitles_endpoint_missing_api_key(self, client):
@@ -152,7 +117,6 @@ class TestSubtitleAPI:
             "/api/v1/subtitles/extract",
             json={"url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ", "lang": "en"}
         )
-
         assert response.status_code == 401
 
     def test_extract_subtitles_endpoint_invalid_url(self, client, api_key):
@@ -162,13 +126,11 @@ class TestSubtitleAPI:
             json={"url": "https://example.com/not-youtube", "lang": "en"},
             headers={"x-api-key": api_key}
         )
-
         assert response.status_code == 422
 
     def test_subtitles_health_endpoint(self, client):
         """Test subtitles health check endpoint."""
         response = client.get("/api/v1/subtitles/health")
-
         assert response.status_code == 200
         data = response.json()
         assert data["status"] == "healthy"