fixing it for huggingface

manimator/services/voiceover.py

@@ -9,6 +9,9 @@ from typing import Optional, Dict, Any
 
 logger = logging.getLogger(__name__)
 
+# Get the base directory (2 levels up from this file: manimator/services -> manimator -> texttoanimation)
+BASE_DIR = Path(__file__).parent.parent.parent
+
 
 # Edge-TTS Voice mapping - high quality neural voices
 EDGE_TTS_VOICES = {
@@ -47,11 +50,11 @@ class SimpleElevenLabsService:
         if not self.api_key:
             logger.warning("ELEVENLABS_API_KEY not set. Will use Edge TTS (free).")
             
-        # Use provided cache_dir or default
+        # Use provided cache_dir or default (use absolute path for containerized environments)
         if cache_dir:
             self.cache_dir = Path(cache_dir) if not isinstance(cache_dir, Path) else cache_dir
         else:
-            self.cache_dir = Path("media/voiceover/elevenlabs")
+            self.cache_dir = BASE_DIR / "media" / "voiceover" / "elevenlabs"
         
         # Ensure cache directory exists
         self.cache_dir.mkdir(parents=True, exist_ok=True)
@@ -125,8 +128,8 @@ class SimpleElevenLabsService:
         try:
             import edge_tts
             
-            # Use a separate cache for edge-tts
-            edge_cache_dir = Path("media/voiceover/edge_tts")
+            # Use absolute path for edge-tts cache (important for containerized environments)
+            edge_cache_dir = BASE_DIR / "media" / "voiceover" / "edge_tts"
             edge_cache_dir.mkdir(parents=True, exist_ok=True)
             
             # Map the voice name to edge-tts voice
@@ -174,7 +177,8 @@ class SimpleElevenLabsService:
         try:
             from gtts import gTTS
             
-            gtts_cache_dir = Path("media/voiceover/gtts")
+            # Use absolute path for gTTS cache (important for containerized environments)
+            gtts_cache_dir = BASE_DIR / "media" / "voiceover" / "gtts"
             gtts_cache_dir.mkdir(parents=True, exist_ok=True)
             
             content_hash = hashlib.md5(text.encode("utf-8")).hexdigest()

start.sh

@@ -18,6 +18,7 @@ export DIRECT_MODE=true
 # Ensure all required directories exist
 mkdir -p media/voiceover/elevenlabs
 mkdir -p media/voiceover/gtts
+mkdir -p media/voiceover/edge_tts
 mkdir -p media/videos
 
 # Start Streamlit with Hugging Face-compatible settings

streamlit_app.py

@@ -512,6 +512,7 @@ def generate_video_direct(input_type: str, input_data: str, quality: str, catego
         media_dir = base_dir / "media"
         (media_dir / "voiceover" / "elevenlabs").mkdir(parents=True, exist_ok=True)
         (media_dir / "voiceover" / "gtts").mkdir(parents=True, exist_ok=True)
+        (media_dir / "voiceover" / "edge_tts").mkdir(parents=True, exist_ok=True)
         (media_dir / "videos").mkdir(parents=True, exist_ok=True)
         
         for attempt in range(max_retries):