Update app.py

app.py

@@ -1,7 +1,7 @@
 import streamlit as st
 from pathlib import Path
 import torch
-from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer
+from transformers import pipeline
 from PIL import Image, ImageDraw, ImageFont
 import tempfile
 import os
@@ -13,8 +13,6 @@ from concurrent.futures import ThreadPoolExecutor
 import io
 import unicodedata
 import re
-import requests
-import random
 import logging
 from typing import Optional, List, Dict, Tuple
 
@@ -24,13 +22,13 @@ class EnhancedVideoGenerator:
         try:
             self.setup_logging()
             self.setup_device()
-            self.initialize_models()
             self.setup_workspace()
             self.load_assets()
             self.setup_themes()
+            self.initialize_models()
         except Exception as e:
-            logging.error(f"Initialization failed: {str(e)}")
-            raise RuntimeError("Failed to initialize video generator")
+            self.logger.error(f"Initialization failed: {str(e)}")
+            raise RuntimeError(f"Failed to initialize video generator: {str(e)}")
 
     def setup_logging(self):
         """Configure logging for the application"""
@@ -52,22 +50,15 @@ class EnhancedVideoGenerator:
     def initialize_models(self):
         """Initialize all AI models"""
         try:
-            # Text generation model
+            # Text generation model - using a smaller model for better compatibility
             self.text_generator = pipeline(
                 'text-generation',
-                model='gpt2',
+                model='distilgpt2',
                 device=0 if self.device == "cuda" else -1
             )
-
-            # Initialize free image generation model
-            self.image_model = AutoModelForCausalLM.from_pretrained(
-                "CompVis/stable-diffusion-v1-4",
-                torch_dtype=torch.float16 if self.device == "cuda" else torch.float32
-            ).to(self.device)
-
         except Exception as e:
             self.logger.error(f"Model initialization failed: {str(e)}")
-            raise
+            self.text_generator = None
 
     def setup_workspace(self):
         """Set up working directory and resources"""
@@ -117,24 +108,29 @@ class EnhancedVideoGenerator:
 
         except Exception as e:
             self.logger.error(f"Asset loading failed: {str(e)}")
+            self.font = ImageFont.load_default()
 
     def generate_visual_assets(self, script: str, style: str) -> List[Dict]:
-        """Generate relevant visual assets based on script content"""
+        """Generate or fetch relevant visual assets based on script content"""
         try:
-            # Extract key topics from script
-            topics = self.extract_key_topics(script)
+            # Generate simple colored backgrounds instead of AI images
+            colors = [
+                (240, 248, 255),  # AliceBlue
+                (240, 255, 255),  # Azure
+                (245, 245, 245),  # WhiteSmoke
+                (255, 250, 250),  # Snow
+                (248, 248, 255)   # GhostWhite
+            ]
             
             assets = []
-            for topic in topics:
-                # Generate AI image
-                image = self.generate_ai_image(topic, style)
-                if image:
-                    assets.append({
-                        'type': 'image',
-                        'data': image,
-                        'topic': topic
-                    })
-
+            for i in range(5):  # Generate 5 different backgrounds
+                img = Image.new('RGB', (1920, 1080), colors[i])
+                assets.append({
+                    'type': 'image',
+                    'data': img,
+                    'topic': f'background_{i}'
+                })
+            
             return assets
 
         except Exception as e:
@@ -148,13 +144,12 @@ class EnhancedVideoGenerator:
         frame_number: int,
         total_frames: int,
         background_image: Optional[Image.Image] = None,
-        size: Tuple[int, int] = (1920, 1080)  # Upgraded to 1080p
+        size: Tuple[int, int] = (1920, 1080)
     ) -> np.ndarray:
         """Create a visually enhanced frame with background, text, and effects"""
         try:
             # Create base frame
             if background_image:
-                # Resize and crop background to fit
                 bg = background_image.resize(size, Image.LANCZOS)
                 frame = np.array(bg)
             else:
@@ -164,15 +159,6 @@ class EnhancedVideoGenerator:
             img = Image.fromarray(frame)
             draw = ImageDraw.Draw(img, 'RGBA')
 
-            # Add subtle gradient overlay
-            overlay = Image.new('RGBA', size, (0, 0, 0, 0))
-            overlay_draw = ImageDraw.Draw(overlay)
-            overlay_draw.rectangle(
-                [0, 0, size[0], size[1]],
-                fill=(255, 255, 255, 100)  # Semi-transparent white
-            )
-            img = Image.alpha_composite(img.convert('RGBA'), overlay)
-
             # Add text with improved styling
             text = self.clean_text(text)
             wrapped_text = textwrap.fill(text, width=50)
@@ -182,7 +168,7 @@ class EnhancedVideoGenerator:
             text_width = text_bbox[2] - text_bbox[0]
             text_height = text_bbox[3] - text_bbox[1]
             text_x = (size[0] - text_width) // 2
-            text_y = size[1] - text_height - 100  # Position at bottom
+            text_y = size[1] - text_height - 100
 
             # Draw text background
             padding = 20
@@ -193,7 +179,7 @@ class EnhancedVideoGenerator:
                     text_x + text_width + padding,
                     text_y + text_height + padding
                 ],
-                fill=(0, 0, 0, 160)  # Semi-transparent black
+                fill=(0, 0, 0, 160)
             )
 
             # Draw text
@@ -204,7 +190,7 @@ class EnhancedVideoGenerator:
                 font=self.font
             )
 
-            # Add progress bar with animation
+            # Add progress bar
             self.draw_animated_progress_bar(
                 draw,
                 frame_number,
@@ -217,7 +203,6 @@ class EnhancedVideoGenerator:
 
         except Exception as e:
             self.logger.error(f"Frame creation failed: {str(e)}")
-            # Return fallback frame
             return np.full((size[1], size[0], 3), theme['bg'], dtype=np.uint8)
 
     def draw_animated_progress_bar(
@@ -228,10 +213,10 @@ class EnhancedVideoGenerator:
         size: Tuple[int, int],
         theme: dict
     ):
-        """Draw an animated progress bar with effects"""
+        """Draw an animated progress bar"""
         try:
             progress = frame_number / total_frames
-            bar_width = int(size[0] * 0.8)  # 80% of screen width
+            bar_width = int(size[0] * 0.8)
             bar_height = 6
             x_offset = (size[0] - bar_width) // 2
             y_position = size[1] - 40
@@ -242,28 +227,18 @@ class EnhancedVideoGenerator:
                 fill=(200, 200, 200, 160)
             )
 
-            # Draw progress with gradient effect
+            # Draw progress
             progress_width = int(bar_width * progress)
-            for x in range(progress_width):
-                alpha = int(255 * (x / bar_width))  # Gradient effect
-                draw.line(
-                    [x_offset + x, y_position, x_offset + x, y_position + bar_height],
-                    fill=(theme['accent'][0], theme['accent'][1], theme['accent'][2], alpha)
-                )
-
-            # Add animated highlight
-            highlight_pos = x_offset + progress_width
-            if highlight_pos < x_offset + bar_width:
-                draw.rectangle(
-                    [highlight_pos-2, y_position-1, highlight_pos+2, y_position + bar_height+1],
-                    fill=(255, 255, 255, 200)
-                )
+            draw.rectangle(
+                [x_offset, y_position, x_offset + progress_width, y_position + bar_height],
+                fill=theme['accent']
+            )
 
         except Exception as e:
             self.logger.error(f"Progress bar drawing failed: {str(e)}")
 
     def generate_voice_over(self, script: str) -> AudioFileClip:
-        """Generate voice-over audio using gTTS"""
+        """Generate voice-over audio"""
         try:
             audio_path = self.temp_dir / "voice.mp3"
             tts = gTTS(
@@ -285,7 +260,7 @@ class EnhancedVideoGenerator:
         duration: int,
         output_path: str
     ) -> str:
-        """Create full video with all enhanced features"""
+        """Create full video with all features"""
         try:
             # Generate visual assets
             assets = self.generate_visual_assets(script, style)
@@ -293,7 +268,7 @@ class EnhancedVideoGenerator:
             # Generate voice-over
             audio = self.generate_voice_over(script)
             
-            # Create frames with visual assets
+            # Create frames
             frames = []
             fps = 30
             total_frames = int(duration * fps)
@@ -302,16 +277,13 @@ class EnhancedVideoGenerator:
                 frame_futures = []
                 
                 for i in range(total_frames):
-                    # Calculate current text segment
                     progress = i / total_frames
                     text_index = int(progress * len(script.split()))
                     current_text = " ".join(script.split()[:text_index + 1])
                     
-                    # Get appropriate background
                     asset_index = int(progress * len(assets))
                     current_asset = assets[asset_index] if assets else None
                     
-                    # Submit frame creation to thread pool
                     future = executor.submit(
                         self.create_enhanced_frame,
                         current_text,
@@ -322,7 +294,6 @@ class EnhancedVideoGenerator:
                     )
                     frame_futures.append(future)
                 
-                # Collect frames
                 frames = [future.result() for future in frame_futures]
 
             # Create video clip
@@ -331,14 +302,6 @@ class EnhancedVideoGenerator:
             # Add voice-over
             video = video.set_audio(audio)
             
-            # Add background music (if available)
-            try:
-                music = AudioFileClip("assets/music/background.mp3")
-                music = music.volumex(0.1).loop(duration=video.duration)
-                video = video.set_audio(CompositeAudioClip([video.audio, music]))
-            except Exception as e:
-                self.logger.warning(f"Background music addition failed: {str(e)}")
-
             # Write final video
             video.write_videofile(
                 output_path,
@@ -361,84 +324,25 @@ class EnhancedVideoGenerator:
         if not isinstance(text, str):
             text = str(text)
         
-        # Normalize unicode characters
         text = unicodedata.normalize('NFKD', text)
-        
-        # Remove non-ASCII characters
         text = text.encode('ascii', 'ignore').decode('ascii')
         
-        # Replace problematic characters
         replacements = {
-            '–': '-',    # en dash
-            '—': '-',    # em dash
-            '"': '"',    # smart quotes
-            '"': '"',    # smart quotes
-            ''': "'",    # smart apostrophe
-            ''': "'",    # smart apostrophe
-            '…': '...',  # ellipsis
+            '–': '-',
+            '—': '-',
+            '"': '"',
+            '"': '"',
+            ''': "'",
+            ''': "'",
+            '…': '...',
         }
         for old, new in replacements.items():
             text = text.replace(old, new)
         
-        # Remove any remaining non-standard characters
         text = re.sub(r'[^\x00-\x7F]+', '', text)
         
         return text.strip()
 
-    def extract_key_topics(self, script: str) -> List[str]:
-        """Extract main topics from the script for visual asset generation"""
-        try:
-            # Simple keyword extraction based on noun phrases
-            # In a production environment, you might want to use a proper NLP library
-            sentences = script.split('.')
-            topics = []
-            
-            for sentence in sentences:
-                words = sentence.strip().split()
-                if len(words) >= 2:
-                    # Extract potential noun phrases (pairs of words)
-                    topics.append(' '.join(words[:2]))
-            
-            # Remove duplicates and limit to top 5 topics
-            return list(dict.fromkeys(topics))[:5]
-            
-        except Exception as e:
-            self.logger.error(f"Topic extraction failed: {str(e)}")
-            return ["default topic"]
-
-    def generate_ai_image(self, prompt: str, style: str) -> Optional[Image.Image]:
-        """Generate an AI image using Stability AI"""
-        try:
-            if not self.stability_api:
-                return None
-
-            # Enhance prompt based on style
-            style_prompts = {
-                'Professional': "professional, corporate, clean, modern",
-                'Creative': "artistic, vibrant, innovative, dynamic",
-                'Educational': "clear, informative, academic, detailed"
-            }
-            
-            enhanced_prompt = f"{prompt}, {style_prompts.get(style, '')}, high quality, 4k"
-            
-            # Generate image
-            response = self.stability_api.generate(
-                prompt=enhanced_prompt,
-                samples=1,
-                width=1920,
-                height=1080
-            )
-            
-            if response and len(response) > 0:
-                image_data = response[0].image
-                return Image.open(io.BytesIO(image_data))
-            
-            return None
-
-        except Exception as e:
-            self.logger.error(f"AI image generation failed: {str(e)}")
-            return None
-
     def cleanup(self):
         """Clean up temporary files and resources"""
         try:
@@ -463,7 +367,6 @@ class EnhancedVideoGenerator:
     def __exit__(self, exc_type, exc_val, exc_tb):
         self.cleanup()
 
-# Streamlit UI Class
 class VideoGeneratorUI:
     def __init__(self):
         self.generator = EnhancedVideoGenerator()
@@ -474,11 +377,10 @@ class VideoGeneratorUI:
         st.write("Create professional videos with AI-generated content")
 
         with st.form("video_generator_form"):
-            # Input fields
             prompt = st.text_area(
-                "Enter your video topic/prompt",
+                "Enter your video script",
                 height=100,
-                help="Describe what you want your video to be about"
+                help="Enter the text you want in your video"
             )
             
             col1, col2 = st.columns(2)
@@ -497,41 +399,17 @@ class VideoGeneratorUI:
                     step=10
                 )
 
-            advanced_options = st.expander("Advanced Options")
-            with advanced_options:
-                use_premium_voice = st.checkbox(
-                    "Use premium voice-over",
-                    value=False,
-                    help="Requires ElevenLabs API key"
-                )
-                
-                include_music = st.checkbox(
-                    "Include background music",
-                    value=True
-                )
-                
-                fps = st.slider(
-                    "Frames per second",
-                    min_value=24,
-                    max_value=60,
-                    value=30
-                )
-
             submit_button = st.form_submit_button("Generate Video")
 
             if submit_button:
                 if not prompt:
-                    st.error("Please enter a prompt for your video.")
+                    st.error("Please enter a script for your video.")
                     return
 
                 try:
                     with st.spinner("Generating your video..."):
                         output_path = f"generated_video_{int(time.time())}.mp4"
                         
-                        # Update generator settings based on advanced options
-                        self.generator.use_premium_voice = use_premium_voice
-                        
-                        # Generate video
                         video_path = self.generator.create_video(
                             prompt,
                             style,
@@ -539,7 +417,6 @@ class VideoGeneratorUI:
                             output_path
                         )
                         
-                        # Show success message and download button
                         st.success("Video generated successfully!")
                         
                         with open(video_path, 'rb') as f:
@@ -552,7 +429,7 @@ class VideoGeneratorUI:
                         
                 except Exception as e:
                     st.error(f"Failed to generate video: {str(e)}")
-                    st.error("Please try again with different settings or contact support.")
+                    st.error("Please try again with different settings.")
 
 if __name__ == "__main__":
     ui = VideoGeneratorUI()