Update gradio_app.py

gradio_app.py

@@ -15,7 +15,6 @@ from mllm_tools.litellm import LiteLLMWrapper
 from src.config.config import Config
 from generate_video import EnhancedVideoGenerator, VideoGenerationConfig, allowed_models
 from provider import provider_manager
-from llm_config import LLMConfigurationFacade, LLMConfiguration
 
 # Configure logging
 logging.basicConfig(
@@ -34,17 +33,8 @@ os.makedirs("thumbnails", exist_ok=True)
 # Global dictionary to track job status
 job_status = {}
 
-# Initialize the new LLM configuration system
-llm_config_facade = LLMConfigurationFacade()
-
-# Model descriptions for better user understanding
-MODEL_DESCRIPTIONS = {
-    "openai/gpt-4o": "🚀 Latest OpenAI model with enhanced capabilities"
-}
-
-# Set default model and provider
-DEFAULT_MODEL = "openai/gpt-4o"
-DEFAULT_PROVIDER = "openai"
+# Default model setting - simplified to use GPT-4o-mini for all operations
+DEFAULT_MODEL = "openai/gpt-4o-mini"
 
 def cancel_job(job_id):
     """Cancel a running job."""
@@ -96,8 +86,8 @@ def get_job_statistics():
 
 def init_video_generator(params):
     """Initialize the EnhancedVideoGenerator with the given parameters."""
-    model_name = params.get('model', 'gemini/gemini-2.5-flash-preview-04-17')
-    helper_model_name = params.get('helper_model', model_name)
+    model_name = params.get('model', DEFAULT_MODEL)
+    helper_model_name = params.get('helper_model', DEFAULT_MODEL)
     verbose = params.get('verbose', True)  # Set verbose to True by default for better debugging
     max_scene_concurrency = params.get('max_scene_concurrency', 1)
     
@@ -303,7 +293,7 @@ def start_async_job(job_id, params):
     thread.start()
     return thread
 
-def submit_job(topic, description, model, helper_model, max_retries, use_rag, use_visual_fix_code, temperature, use_context_learning, verbose, max_scene_concurrency):
+def submit_job(topic, description, max_retries, use_rag, use_visual_fix_code, temperature, use_context_learning, verbose, max_scene_concurrency, api_key):
     """Submit a new video generation job."""
     # Input validation
     if not topic.strip():
@@ -318,13 +308,13 @@ def submit_job(topic, description, model, helper_model, max_retries, use_rag, us
     if len(description.strip()) < 10:
         return "❌ Error: Description must be at least 10 characters long", None, gr.update(visible=False)
     
-    # Validate LLM configuration
-    current_config = llm_config_facade.load_configuration()
-    if not current_config or not current_config.api_key:
-        return "❌ Error: Please configure and validate your API key first", None, gr.update(visible=False)
+    # Set default model
+    model = DEFAULT_MODEL
+    helper_model = DEFAULT_MODEL
     
-    if not model:
-        return "❌ Error: Please select a model", None, gr.update(visible=False)
+    # Validate API key
+    if not api_key or not api_key.strip():
+        return "❌ Error: Please enter your OpenAI API key", None, gr.update(visible=False)
     
     try:
         # Generate job ID
@@ -342,7 +332,7 @@ def submit_job(topic, description, model, helper_model, max_retries, use_rag, us
             'message': 'Job submitted, waiting to start...'
         }
         
-        # Prepare parameters with validated configuration
+        # Prepare parameters with default configuration
         params = {
             'topic': topic,
             'description': description,
@@ -356,9 +346,9 @@ def submit_job(topic, description, model, helper_model, max_retries, use_rag, us
             'verbose': verbose,
             'max_scene_concurrency': max_scene_concurrency,
             'output_dir': Config.OUTPUT_DIR,
-            # Add validated configuration
-            'provider': current_config.provider,
-            'api_key': current_config.api_key
+            # Use OpenAI as default provider
+            'provider': 'openai',
+            'api_key': api_key.strip()
         }
         
         # Start job asynchronously
@@ -551,27 +541,22 @@ with gr.Blocks(
                     )
             with gr.Column(scale=1):
                 with gr.Group():
-                    gr.Markdown("### 🌐 LLM Configuration")
-                    # Remove provider selection, always use GPT-4o
-                    model_input = gr.Dropdown(
-                        label="🤖 AI Model",
-                        choices=[DEFAULT_MODEL],
-                        value=DEFAULT_MODEL,
-                        interactive=False,
-                        info="Model is fixed to GPT-4o for all generations"
-                    )
-                    model_description = gr.Markdown(
-                        value=MODEL_DESCRIPTIONS[DEFAULT_MODEL],
-                        visible=True,
-                        elem_classes=["model-description"]
-                    )
-                    helper_model_input = gr.Dropdown(
-                        label="🔧 Helper Model", 
-                        choices=[DEFAULT_MODEL],
-                        value=DEFAULT_MODEL,
-                        interactive=False,
-                        info="Helper model is fixed to GPT-4o"
+                    gr.Markdown("### 🔑 API Configuration")
+                    
+                    # Simple API key input
+                    api_key_input = gr.Textbox(
+                        label="🔑 OpenAI API Key",
+                        placeholder="Enter your OpenAI API key",
+                        type="password",
+                        value=os.getenv('OPENAI_API_KEY', ''),
+                        interactive=True,
+                        info="Your OpenAI API key for GPT-4o-mini access"
                     )
+                    
+                    # Display current model setting
+                    gr.Markdown(f"**🤖 AI Model:** Using {DEFAULT_MODEL} for all operations")
+                    
+                    # Temperature control
                     temperature_input = gr.Slider(
                         label="🌡️ Creativity (Temperature)", 
                         minimum=0.0, 
@@ -768,11 +753,9 @@ with gr.Blocks(
           - Examples or applications to demonstrate
           - Preferred video length or depth
         
-        ### 🤖 Step 2: Model Selection
-        - **Gemini 1.5 Pro**: Best for complex mathematical reasoning
-        - **Gemini 2.0 Flash**: Fastest processing, good for simple topics
-        - **GPT-4**: Reliable and consistent output
-        - **Claude**: Excellent for detailed explanations
+        ### 🔑 Step 2: API Configuration
+        - **OpenAI API Key**: Enter your OpenAI API key for access
+        - **AI Model**: Automatically uses GPT-4o-mini for optimal cost and performance
         
         ### ⚙️ Step 3: Advanced Settings
         - **Temperature**: 0.3-0.5 for factual content, 0.7-0.9 for creative explanations
@@ -806,10 +789,7 @@ with gr.Blocks(
     
     # Event handlers with improved functionality
     def clear_form():
-        return ("", "", 0.7, False, True, False, True, 1, 1, "Form cleared! Ready for new input.")
-    
-    def update_model_description(model):
-        return MODEL_DESCRIPTIONS.get(model, "No description available")
+        return ("", "", 0.7, False, True, False, True, 1, 1, os.getenv('OPENAI_API_KEY', ''), "Form cleared! Ready for new input.")
     
     def update_stats():
         stats = get_job_statistics()
@@ -830,12 +810,21 @@ with gr.Blocks(
         job_status.clear()
         return f"Cleared all {count} jobs"
     
+    # Connect simplified event handlers - no model selection needed
+    
+    clear_form_btn.click(
+        fn=clear_form,
+        outputs=[topic_input, description_input, temperature_input, 
+                use_rag_input, use_visual_fix_code_input, use_context_learning_input, 
+                verbose_input, max_retries_input, max_scene_concurrency_input, api_key_input, result_text]
+    )
+    
     submit_btn.click(
         fn=submit_job,
         inputs=[
-            topic_input, description_input, model_input, helper_model_input, max_retries_input,
+            topic_input, description_input, max_retries_input,
             use_rag_input, use_visual_fix_code_input, temperature_input, use_context_learning_input,
-            verbose_input, max_scene_concurrency_input
+            verbose_input, max_scene_concurrency_input, api_key_input
         ],
         outputs=[result_text, job_id_output, status_container]
     ).then(