Update app.py

app.py

@@ -1,7 +1,8 @@
 from flask import Flask, request, jsonify, send_file
 from flask_cors import CORS
-from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer
+from transformers import AutoModelForCausalLM, AutoTokenizer
 from knowledgebase import KiswahiliKnowledgeBase, enhance_with_kiswahili
+from video_generation import FreeVideoGenerator
 import torch
 import time
 import re
@@ -14,6 +15,7 @@ import requests
 from PIL import Image
 import os
 import random
+import tempfile
 
 # Configure logging
 logging.basicConfig(level=logging.INFO)
@@ -25,16 +27,20 @@ CORS(app)
 # Initialize Kiswahili Knowledge Base
 kb = KiswahiliKnowledgeBase()
 
+# Initialize Video Generator
+HF_TOKEN = os.getenv('HF_TOKEN', 'your_hugging_face_token_here')
+video_gen = FreeVideoGenerator(HF_TOKEN)
+
 model = None
 tokenizer = None
 model_loaded = False
 
 # Hugging Face Configuration
-HF_TOKEN = os.getenv('HF_TOKEN', 'your_hugging_face_token_here')
 HF_API_URLS = {
     "text": "https://api-inference.huggingface.co/models/Qwen/Qwen2.5-7B-Instruct",
     "image": "https://api-inference.huggingface.co/models/runwayml/stable-diffusion-v1-5",
-    "fast_image": "https://api-inference.huggingface.co/models/stabilityai/stable-diffusion-2-1"
+    "fast_image": "https://api-inference.huggingface.co/models/stabilityai/stable-diffusion-2-1",
+    "video": "https://api-inference.huggingface.co/models/cerspense/zeroscope_v2_576w"
 }
 
 # Performance optimizations
@@ -44,25 +50,26 @@ request_timeout = 30
 
 # Advanced System Prompt (Optimized for Hugging Face)
 STANLEY_AI_SYSTEM = """You are STANLEY AI - an advanced AI assistant created by Stanley Samwel Owino, a Machine Learning Engineer from Kenya.
-
 CORE CAPABILITIES:
 - Provide detailed, comprehensive responses
 - Integrate Kiswahili phrases naturally when relevant
 - Share cultural insights and proverbs
 - Reference Lion King lore accurately
-- Generate and describe images
+- Generate and describe images and videos
 - Be helpful, knowledgeable, and engaging
-
 KISWAHILI INTEGRATION:
 Use phrases like "Habari", "Asante", "Karibu", "Pole sana" appropriately
 Explain cultural concepts with authenticity
 Share Swahili proverbs when relevant
-
-IMAGE GENERATION:
-You can generate images based on user descriptions
+IMAGE & VIDEO GENERATION:
+You can generate images and videos based on user descriptions
 Enhance prompts with cultural context when relevant
-Describe generated images in detail
-
+Describe generated content in detail
+VIDEO CAPABILITIES:
+- Create 4-second videos from text
+- Generate cultural theme videos
+- Create animations from text
+- Make slideshows from images
 RESPONSE STYLE: Be concise yet comprehensive, culturally aware, and genuinely helpful."""
 
 def load_model():
@@ -336,7 +343,6 @@ def generate_image_fallback(prompt):
     """Create simple placeholder images"""
     try:
         from PIL import Image, ImageDraw
-        import random
         
         # Create colorful placeholder
         width, height = 512, 512
@@ -384,21 +390,30 @@ def enhance_prompt_with_kiswahili(prompt):
 def home():
     return jsonify({
         "message": "🚀 STANLEY AI - Created by Stanley Samwel Owino (Machine Learning Engineer)", 
-        "version": "2.2",
+        "version": "3.0",
         "creator": "Stanley Samwel Owino",
         "role": "Machine Learning Engineer",
         "features": [
             "Hugging Face Optimized",
             "Fast Text Generation", 
             "Free Image Generation",
+            "Free Video Generation",
             "Kiswahili Integration",
             "Cultural Knowledge",
             "Response Caching",
             "API Fallbacks"
         ],
+        "video_capabilities": [
+            "Text-to-Video",
+            "Image-to-Video", 
+            "Slideshow Creation",
+            "Text Animation",
+            "Cultural Themes"
+        ],
         "status": "active",
         "model": "Qwen2.5 + HF Inference",
-        "image_generation": "Hugging Face API"
+        "image_generation": "Hugging Face API",
+        "video_generation": "Hugging Face Free Models"
     })
 
 @app.route('/api/chat', methods=['POST'])
@@ -426,6 +441,19 @@ def chat():
                 "response_time": round(time.time() - start_time, 2)
             })
         
+        # Check if user wants to generate a video
+        if video_gen.detect_video_request(user_message):
+            video_prompt = video_gen.extract_video_prompt(user_message)
+            enhanced_video_prompt = video_gen.enhance_prompt_with_context(video_prompt)
+            
+            return jsonify({
+                "response": f"🎬 I can create a video for: '{enhanced_video_prompt}'. Use the video generation feature below!",
+                "video_suggestion": enhanced_video_prompt,
+                "status": "success",
+                "suggest_video": True,
+                "response_time": round(time.time() - start_time, 2)
+            })
+        
         response = generate_comprehensive_response(user_message)
         response_time = round(time.time() - start_time, 2)
         
@@ -550,6 +578,282 @@ def generate_cultural_image():
             "status": "error"
         }), 500
 
+@app.route('/api/generate-video', methods=['POST'])
+def generate_video_endpoint():
+    """Generate videos from text prompts"""
+    try:
+        start_time = time.time()
+        data = request.get_json()
+        prompt = data.get('prompt', '')
+        
+        if not prompt:
+            return jsonify({"error": "Tafadhali provide a video prompt"}), 400
+        
+        # Enhance prompt with cinematic and cultural context
+        enhanced_prompt = video_gen.enhance_prompt_with_context(prompt)
+        
+        logger.info(f"🎬 Generating video: {enhanced_prompt[:50]}...")
+        
+        # Generate video
+        video_data = video_gen.generate_text_to_video(enhanced_prompt)
+        
+        if video_data:
+            generation_time = round(time.time() - start_time, 2)
+            
+            return jsonify({
+                "video": video_data,
+                "prompt": prompt,
+                "enhanced_prompt": enhanced_prompt,
+                "status": "success",
+                "generation_time": generation_time,
+                "format": "mp4",
+                "duration": "3-4 seconds",
+                "provider": "hugging_face_free",
+                "message": "Video generated successfully!"
+            })
+        else:
+            return jsonify({
+                "error": "Pole! Video generation service is busy. Try again in a moment.",
+                "status": "error",
+                "fallback_suggestion": "You can try generating individual images instead."
+            }), 500
+            
+    except Exception as e:
+        logger.error(f"Video endpoint error: {e}")
+        return jsonify({
+            "error": f"Pole! Video generation failed: {str(e)}",
+            "status": "error"
+        }), 500
+
+@app.route('/api/generate-cultural-video', methods=['POST'])
+def generate_cultural_video_endpoint():
+    """Generate videos with specific Kiswahili cultural themes"""
+    try:
+        start_time = time.time()
+        data = request.get_json()
+        theme = data.get('theme', 'safari')
+        style = data.get('style', 'animated')
+        
+        logger.info(f"🌍 Generating cultural video: {theme} ({style} style)")
+        
+        # Generate cultural video
+        video_data = video_gen.create_cultural_video(theme, style)
+        
+        if video_data:
+            generation_time = round(time.time() - start_time, 2)
+            
+            # Get cultural description
+            theme_descriptions = {
+                "safari": "African wildlife and landscapes",
+                "dance": "Traditional dance and celebration",
+                "market": "Vibrant African market scene",
+                "coastal": "Swahili coast and ocean",
+                "wildlife": "African wildlife documentary",
+                "village": "Traditional village life"
+            }
+            
+            return jsonify({
+                "video": video_data,
+                "theme": theme,
+                "style": style,
+                "description": theme_descriptions.get(theme, "Cultural scene"),
+                "status": "success",
+                "generation_time": generation_time,
+                "cultural_context": "kiswahili_theme",
+                "message": f"Cultural video generated: {theme.replace('_', ' ').title()}"
+            })
+        else:
+            return jsonify({
+                "error": "Pole! Cultural video generation failed",
+                "status": "error",
+                "suggestion": "Try a different theme or style"
+            }), 500
+            
+    except Exception as e:
+        logger.error(f"Cultural video error: {e}")
+        return jsonify({
+            "error": f"Pole! Cultural video failed: {str(e)}",
+            "status": "error"
+        }), 500
+
+@app.route('/api/animate-text', methods=['POST'])
+def animate_text_endpoint():
+    """Create animated text videos"""
+    try:
+        start_time = time.time()
+        data = request.get_json()
+        text = data.get('text', 'Hello World!')
+        animation_type = data.get('animation_type', 'pulse')
+        
+        if not text:
+            return jsonify({"error": "Tafadhali provide text to animate"}), 400
+        
+        # Generate text animation
+        video_data = video_gen.generate_animation_from_text(text)
+        
+        if video_data:
+            generation_time = round(time.time() - start_time, 2)
+            
+            return jsonify({
+                "video": video_data,
+                "text": text,
+                "animation_type": animation_type,
+                "status": "success",
+                "generation_time": generation_time,
+                "format": "mp4",
+                "duration": "3 seconds",
+                "message": "Text animation created!"
+            })
+        else:
+            return jsonify({
+                "error": "Pole! Text animation failed",
+                "status": "error"
+            }), 500
+            
+    except Exception as e:
+        logger.error(f"Text animation error: {e}")
+        return jsonify({
+            "error": f"Pole! Text animation failed: {str(e)}",
+            "status": "error"
+        }), 500
+
+@app.route('/api/create-slideshow', methods=['POST'])
+def create_slideshow_endpoint():
+    """Create video slideshow from images"""
+    try:
+        start_time = time.time()
+        data = request.get_json()
+        images = data.get('images', [])  # List of base64 images
+        duration = data.get('duration', 2.0)
+        
+        if not images or len(images) < 2:
+            return jsonify({
+                "error": "Tafadhali provide at least 2 images",
+                "status": "error"
+            }), 400
+        
+        # Create slideshow
+        video_data = video_gen.create_slideshow_video(images, duration)
+        
+        if video_data:
+            generation_time = round(time.time() - start_time, 2)
+            
+            return jsonify({
+                "video": video_data,
+                "image_count": len(images),
+                "duration_per_image": duration,
+                "total_duration": len(images) * duration,
+                "status": "success",
+                "generation_time": generation_time,
+                "format": "mp4",
+                "message": f"Slideshow created with {len(images)} images"
+            })
+        else:
+            return jsonify({
+                "error": "Pole! Slideshow creation failed",
+                "status": "error"
+            }), 500
+            
+    except Exception as e:
+        logger.error(f"Slideshow error: {e}")
+        return jsonify({
+            "error": f"Pole! Slideshow failed: {str(e)}",
+            "status": "error"
+        }), 500
+
+@app.route('/api/image-to-video', methods=['POST'])
+def image_to_video_endpoint():
+    """Generate video from an image"""
+    try:
+        start_time = time.time()
+        data = request.get_json()
+        image_data = data.get('image', '')
+        prompt = data.get('prompt', '')
+        
+        if not image_data:
+            return jsonify({"error": "Tafadhali provide an image"}), 400
+        
+        # Generate video from image
+        video_data = video_gen.generate_image_to_video(image_data, prompt)
+        
+        if video_data:
+            generation_time = round(time.time() - start_time, 2)
+            
+            return jsonify({
+                "video": video_data,
+                "prompt": prompt if prompt else "Generated from image",
+                "status": "success",
+                "generation_time": generation_time,
+                "format": "mp4",
+                "duration": "4 seconds",
+                "provider": "hugging_face_free",
+                "message": "Video generated from image!"
+            })
+        else:
+            return jsonify({
+                "error": "Pole! Image to video conversion failed",
+                "status": "error"
+            }), 500
+            
+    except Exception as e:
+        logger.error(f"Image to video error: {e}")
+        return jsonify({
+            "error": f"Pole! Image to video failed: {str(e)}",
+            "status": "error"
+        }), 500
+
+@app.route('/api/video/info')
+def video_info_endpoint():
+    """Get information about video generation capabilities"""
+    try:
+        info = video_gen.get_video_info()
+        
+        return jsonify({
+            "status": "success",
+            "video_capabilities": info,
+            "creator": "Stanley Samwel Owino",
+            "note": "Free video generation powered by Hugging Face open-source models",
+            "supported_formats": ["MP4", "WebM"],
+            "max_resolution": "576x320",
+            "typical_generation_time": "30-90 seconds",
+            "free_models_available": True
+        })
+        
+    except Exception as e:
+        return jsonify({
+            "error": f"Pole! Could not get video info: {str(e)}",
+            "status": "error"
+        }), 500
+
+@app.route('/api/detect-video-request', methods=['POST'])
+def detect_video_request_endpoint():
+    """Detect if user wants to generate a video"""
+    try:
+        data = request.get_json()
+        text = data.get('text', '')
+        
+        if not text:
+            return jsonify({"error": "Tafadhali provide text"}), 400
+        
+        is_video_request = video_gen.detect_video_request(text)
+        prompt = None
+        
+        if is_video_request:
+            prompt = video_gen.extract_video_prompt(text)
+        
+        return jsonify({
+            "is_video_request": is_video_request,
+            "extracted_prompt": prompt,
+            "suggested_endpoint": "/api/generate-video" if is_video_request else None,
+            "status": "success"
+        })
+        
+    except Exception as e:
+        return jsonify({
+            "error": f"Pole! Detection failed: {str(e)}",
+            "status": "error"
+        }), 500
+
 @app.route('/api/quick-chat', methods=['POST'])
 def quick_chat():
     """Faster chat endpoint for simple queries"""
@@ -566,10 +870,13 @@ def quick_chat():
             'hi': 'Habari! Karibu kwa Stanley AI. How can I help you today?',
             'thanks': 'Asante sana! Karibu tena.',
             'thank you': 'Asante! Happy to help.',
-            'help': 'Ninaweza kukupa: Maelezo, Picha, Maarifa ya Kiswahili, na zaidi!',
+            'help': 'Ninaweza kukupa: Maelezo, Picha, Video, Maarifa ya Kiswahili, na zaidi!',
             'who created you': 'I was created by Stanley Samwel Owino, a Machine Learning Engineer from Kenya.',
             'who made you': 'Stanley Samwel Owino - Machine Learning Engineer and AI researcher from Kenya.',
-            'creator': 'Stanley Samwel Owino - Machine Learning Engineer passionate about AI and cultural integration.'
+            'creator': 'Stanley Samwel Owino - Machine Learning Engineer passionate about AI and cultural integration.',
+            'can you make videos': 'Ndio! I can create videos from text, images, and cultural themes. Try the video generation feature!',
+            'video capabilities': 'I can generate: 1) Text-to-video, 2) Image-to-video, 3) Slideshows, 4) Text animations, 5) Cultural theme videos',
+            'make a video': 'Sure! Just use the video generation endpoint or tell me what you want to create a video of.'
         }
         
         msg_lower = user_message.lower().strip()
@@ -598,10 +905,12 @@ def system_status():
         "role": "Machine Learning Engineer",
         "model_loaded": model_loaded,
         "hugging_face_available": True,
+        "video_generation_available": True,
         "cache_size": len(response_cache),
         "features": [
             "Text Generation",
             "Image Generation", 
+            "Video Generation",
             "Kiswahili Knowledge",
             "Cultural Integration",
             "Fast Responses"
@@ -610,7 +919,8 @@ def system_status():
             "Response Caching",
             "API Fallbacks",
             "Quick Responses",
-            "Cultural Prompts"
+            "Cultural Prompts",
+            "Video Compression"
         ]
     })
 
@@ -632,7 +942,6 @@ def clear_cache():
             "status": "error"
         }), 500
 
-# Keep all your existing Kiswahili knowledge endpoints
 @app.route('/api/kiswahili/proverbs')
 def get_proverbs():
     """Get random Swahili proverbs"""
@@ -662,16 +971,83 @@ def get_phrases():
         "I'm fine": "Nzuri",
         "Please": "Tafadhali",
         "Yes": "Ndio",
-        "No": "Hapana"
+        "No": "Hapana",
+        "I love you": "Nakupenda",
+        "See you later": "Tutaonana",
+        "Congratulations": "Hongera",
+        "Good luck": "Bahati njema",
+        "Have a good day": "Siku njema"
     }
     return jsonify(phrases)
 
+@app.route('/api/kiswahili/learn')
+def learn_kiswahili():
+    """Interactive Kiswahili learning"""
+    lessons = {
+        "basics": {
+            "greetings": {
+                "Habari": "Hello / How are you?",
+                "Nzuri": "Good / Fine",
+                "Asante": "Thank you",
+                "Karibu": "Welcome / You're welcome",
+                "Tafadhali": "Please"
+            },
+            "questions": {
+                "Unaitwa nani?": "What is your name?",
+                "Unatoka wapi?": "Where are you from?",
+                "Unaishi wapi?": "Where do you live?",
+                "Unafanya nini?": "What do you do?"
+            },
+            "responses": {
+                "Ninaitwa...": "My name is...",
+                "Ninatoka Kenya": "I am from Kenya",
+                "Naishi Nairobi": "I live in Nairobi",
+                "Mimi ni mwanafunzi": "I am a student"
+            }
+        },
+        "numbers": {
+            "1": "Moja",
+            "2": "Mbili",
+            "3": "Tatu",
+            "4": "Nne",
+            "5": "Tano",
+            "10": "Kumi",
+            "100": "Mia moja"
+        },
+        "culture": {
+            "Hakuna Matata": "No worries / No problems",
+            "Pole pole": "Slowly slowly (proverb)",
+            "Haraka haraka haina baraka": "Hurry hurry has no blessing",
+            "Jambo": "Hello / Matter / Issue"
+        }
+    }
+    return jsonify({
+        "lessons": lessons,
+        "language": "Kiswahili",
+        "region": "East Africa",
+        "note": "Interactive learning by Stanley AI"
+    })
+
 if __name__ == '__main__':
-    print("🚀 STANLEY AI - Hugging Face Optimized")
+    print("🚀 STANLEY AI - Complete Edition with Video Generation")
     print("👨‍💻 Created by: Stanley Samwel Owino - Machine Learning Engineer")
     print("🌍 Kiswahili Knowledge: Loaded")
     print("🖼️ Image Generation: Hugging Face API")
+    print("🎬 Video Generation: Free Hugging Face Models")
     print("⚡ Performance: Optimized")
     print("🔧 Fallbacks: Enabled")
+    print("📊 API Endpoints: Active")
+    print("=" * 50)
+    print("Available Endpoints:")
+    print("1. /api/chat - Text chat with AI")
+    print("2. /api/generate-image - Image generation")
+    print("3. /api/generate-video - Video generation from text")
+    print("4. /api/generate-cultural-video - Cultural theme videos")
+    print("5. /api/animate-text - Text animation")
+    print("6. /api/create-slideshow - Slideshow from images")
+    print("7. /api/image-to-video - Video from images")
+    print("8. /api/quick-chat - Fast responses")
+    print("9. /api/kiswahili/* - Kiswahili learning")
+    print("=" * 50)
     
     app.run(debug=True, host='0.0.0.0', port=7860, threaded=True)