disable local models for now

app.py

@@ -13,12 +13,12 @@ from core.code_loader_demo import CodeLoader
 from core.llm_demo import LLM
 from core.model_config import (
     PROVIDER_PRESETS,
-    create_local_model_config,
+    # create_local_model_config,  # TODO: Re-enable when local models are fixed
     create_provider_model_config,
     get_api_key_env_name,
     get_provider_from_model,
 )
-from core.ollama_models import fetch_ollama_models
+# from core.ollama_models import fetch_ollama_models  # TODO: Re-enable when local models are fixed
 from core.openrouter_models import fetch_free_models, get_model_config
 from core.prompt_demo import Prompt
 from core.token_counter_demo import TokenCounter
@@ -392,11 +392,10 @@ def main():
             
             **LLM Provider Recommendations:**
             - **Free Models (OpenRouter)**: Best for quick checks of already public paper+code combinations
-            - **Local Models (Ollama/vLLM)**: Best for privacy-sensitive content, e.g. for unpublished papers or code
             - **Provider Models (OpenAI, Anthropic, etc.)**: Best for high precision and best recall
             
             **Features:**
-            - Support for multiple LLM providers (free, local, or premium models)
+            - Support for multiple LLM providers (free or premium models)
             - Automatic content fetching from arXiv and GitHub
             - File upload support for custom papers and repositories
             - Secure API key handling (keys never stored or logged)
@@ -440,8 +439,10 @@ def main():
         # Model type selection
         model_type = st.radio(
             "Model Type",
-            options=["Free Models (OpenRouter)", "Local Model (Ollama/vLLM)", "Provider (OpenAI, Anthropic, Gemini, etc.)"],
-            help="Select free models (no API key), local models (Ollama/vLLM), or provider models (requires API key)",
+            options=["Free Models (OpenRouter)", "Provider (OpenAI, Anthropic, Gemini, etc.)"],
+            # options=["Free Models (OpenRouter)", "Local Model (Ollama/vLLM)", "Provider (OpenAI, Anthropic, Gemini, etc.)"],  # TODO: Re-enable Local Model option when fixed
+            help="Select free models (no API key) or provider models (requires API key)",
+            # help="Select free models (no API key), local models (Ollama/vLLM), or provider models (requires API key)",  # TODO: Re-enable when local models are fixed
             key="model_type_radio",
             index=0,  # Default to Free Models
         )
@@ -467,7 +468,7 @@ def main():
                 # Show privacy warning
                 st.warning(
                     "⚠️ **Privacy Notice**: Free models are provided via [OpenRouter](https://openrouter.ai). "
-                    "The model provider may log your prompts and outputs. For enhanced privacy, consider using Local or Provider models with your own API keys."
+                    "The model provider may log your prompts and outputs. For enhanced privacy, consider using Provider models with your own API keys."
                 )
                 # Create model options from fetched models
                 model_options = {get_model_config(m)["name"]: get_model_config(m) for m in free_models_raw}
@@ -494,112 +495,113 @@ def main():
                     st.error("⚠️ No free models available. Please try again later or use a different model type.")
                     model_config = None
         
-        elif model_type == "Local Model (Ollama/vLLM)":
-            st.info("🖥️ **Local Model**: Use models running locally via Ollama or vLLM (OpenAI-compatible server).")
-            
-            local_model_type = st.radio(
-                "Local Server Type",
-                options=["Ollama", "vLLM (OpenAI-compatible)"],
-                help="Select the type of local server",
-                key="local_server_type",
-            )
-            
-            if local_model_type == "Ollama":
-                # API Base URL comes first
-                api_base = st.text_input(
-                    "API Base URL",
-                    value="http://localhost:11434",
-                    help="Ollama API base URL",
-                    key="ollama_api_base",
-                )
-                
-                # Query Ollama for available models if API base is provided
-                model_input = None
-                if api_base and api_base.strip():
-                    try:
-                        with st.spinner("Fetching available models from Ollama..."):
-                            available_models = fetch_ollama_models(api_base.strip())
-                        
-                        if available_models:
-                            model_input = st.selectbox(
-                                "Select Model",
-                                options=available_models,
-                                help="Select a model from your Ollama server",
-                                key="ollama_model_select",
-                            )
-                        else:
-                            st.warning("⚠️ No models found or unable to connect to Ollama. You can still enter a model name manually.")
-                            model_input = st.text_input(
-                                "Model Name (manual entry)",
-                                placeholder="e.g., llama2, mistral, codellama",
-                                help="Enter the Ollama model name manually (without 'ollama/' prefix)",
-                                key="ollama_model_input_manual",
-                            )
-                    except Exception as e:
-                        logger.error(f"Error fetching Ollama models: {e}")
-                        st.warning(f"⚠️ Could not fetch models from Ollama: {str(e)}. You can still enter a model name manually.")
-                        model_input = st.text_input(
-                            "Model Name (manual entry)",
-                            placeholder="e.g., llama2, mistral, codellama",
-                            help="Enter the Ollama model name manually (without 'ollama/' prefix)",
-                            key="ollama_model_input_manual",
-                        )
-                else:
-                    st.info("💡 Enter the API Base URL above to see available models, or enter a model name manually below.")
-                    model_input = st.text_input(
-                        "Model Name",
-                        placeholder="e.g., llama2, mistral, codellama",
-                        help="Enter the Ollama model name (without 'ollama/' prefix)",
-                        key="ollama_model_input",
-                    )
-                
-                max_context = st.number_input(
-                    "Max Context (tokens)",
-                    min_value=1000,
-                    max_value=1000000,
-                    value=131072,
-                    step=1000,
-                    help="Maximum context window size in tokens",
-                    key="ollama_max_context",
-                )
-                
-                if model_input and api_base:
-                    model_name = f"ollama/{model_input}"
-                    model_config = create_local_model_config(
-                        model=model_name,
-                        api_base=api_base.strip(),
-                        max_context=max_context,
-                    )
-            else:  # vLLM
-                model_input = st.text_input(
-                    "Model Name",
-                    placeholder="e.g., gpt-3.5-turbo, mistralai/Mistral-7B-Instruct-v0.1",
-                    help="Enter the model name for vLLM",
-                    key="vllm_model_input",
-                )
-                api_base = st.text_input(
-                    "API Base URL",
-                    value="http://localhost:8000/v1",
-                    help="vLLM API base URL (OpenAI-compatible endpoint)",
-                    key="vllm_api_base",
-                )
-                max_context = st.number_input(
-                    "Max Context (tokens)",
-                    min_value=1000,
-                    max_value=1000000,
-                    value=131072,
-                    step=1000,
-                    help="Maximum context window size in tokens",
-                    key="vllm_max_context",
-                )
-                
-                if model_input:
-                    model_name = model_input
-                    model_config = create_local_model_config(
-                        model=model_name,
-                        api_base=api_base,
-                        max_context=max_context,
-                    )
+        # TODO: Re-enable when local models are fixed
+        # elif model_type == "Local Model (Ollama/vLLM)":
+        #     st.info("🖥️ **Local Model**: Use models running locally via Ollama or vLLM (OpenAI-compatible server).")
+        #     
+        #     local_model_type = st.radio(
+        #         "Local Server Type",
+        #         options=["Ollama", "vLLM (OpenAI-compatible)"],
+        #         help="Select the type of local server",
+        #         key="local_server_type",
+        #     )
+        #     
+        #     if local_model_type == "Ollama":
+        #         # API Base URL comes first
+        #         api_base = st.text_input(
+        #             "API Base URL",
+        #             value="http://localhost:11434",
+        #             help="Ollama API base URL",
+        #             key="ollama_api_base",
+        #         )
+        #         
+        #         # Query Ollama for available models if API base is provided
+        #         model_input = None
+        #         if api_base and api_base.strip():
+        #             try:
+        #                 with st.spinner("Fetching available models from Ollama..."):
+        #                     available_models = fetch_ollama_models(api_base.strip())
+        #                 
+        #                 if available_models:
+        #                     model_input = st.selectbox(
+        #                         "Select Model",
+        #                         options=available_models,
+        #                         help="Select a model from your Ollama server",
+        #                         key="ollama_model_select",
+        #                     )
+        #                 else:
+        #                     st.warning("⚠️ No models found or unable to connect to Ollama. You can still enter a model name manually.")
+        #                     model_input = st.text_input(
+        #                         "Model Name (manual entry)",
+        #                         placeholder="e.g., llama2, mistral, codellama",
+        #                         help="Enter the Ollama model name manually (without 'ollama/' prefix)",
+        #                         key="ollama_model_input_manual",
+        #                     )
+        #             except Exception as e:
+        #                 logger.error(f"Error fetching Ollama models: {e}")
+        #                 st.warning(f"⚠️ Could not fetch models from Ollama: {str(e)}. You can still enter a model name manually.")
+        #                 model_input = st.text_input(
+        #                     "Model Name (manual entry)",
+        #                     placeholder="e.g., llama2, mistral, codellama",
+        #                     help="Enter the Ollama model name manually (without 'ollama/' prefix)",
+        #                     key="ollama_model_input_manual",
+        #                 )
+        #         else:
+        #             st.info("💡 Enter the API Base URL above to see available models, or enter a model name manually below.")
+        #             model_input = st.text_input(
+        #                 "Model Name",
+        #                 placeholder="e.g., llama2, mistral, codellama",
+        #                 help="Enter the Ollama model name (without 'ollama/' prefix)",
+        #                 key="ollama_model_input",
+        #             )
+        #         
+        #         max_context = st.number_input(
+        #             "Max Context (tokens)",
+        #             min_value=1000,
+        #             max_value=1000000,
+        #             value=131072,
+        #             step=1000,
+        #             help="Maximum context window size in tokens",
+        #             key="ollama_max_context",
+        #         )
+        #         
+        #         if model_input and api_base:
+        #             model_name = f"ollama/{model_input}"
+        #             model_config = create_local_model_config(
+        #                 model=model_name,
+        #                 api_base=api_base.strip(),
+        #                 max_context=max_context,
+        #             )
+        #     else:  # vLLM
+        #         model_input = st.text_input(
+        #             "Model Name",
+        #             placeholder="e.g., gpt-3.5-turbo, mistralai/Mistral-7B-Instruct-v0.1",
+        #             help="Enter the model name for vLLM",
+        #             key="vllm_model_input",
+        #         )
+        #         api_base = st.text_input(
+        #             "API Base URL",
+        #             value="http://localhost:8000/v1",
+        #             help="vLLM API base URL (OpenAI-compatible endpoint)",
+        #             key="vllm_api_base",
+        #         )
+        #         max_context = st.number_input(
+        #             "Max Context (tokens)",
+        #             min_value=1000,
+        #             max_value=1000000,
+        #             value=131072,
+        #             step=1000,
+        #             help="Maximum context window size in tokens",
+        #             key="vllm_max_context",
+        #         )
+        #         
+        #         if model_input:
+        #             model_name = model_input
+        #             model_config = create_local_model_config(
+        #                 model=model_name,
+        #                 api_base=api_base,
+        #                 max_context=max_context,
+        #             )
         
         else:  # Provider Model
             st.info("🔑 **Provider Model**: Use your own API keys to access premium models. Your keys are never stored, logged, or displayed.")

core/ollama_models.py

@@ -21,7 +21,7 @@ def fetch_ollama_models(api_base: str) -> list[str]:
     try:
         # Ollama API endpoint for listing models
         url = f"{api_base.rstrip('/')}/api/tags"
-        response = requests.get(url, timeout=5)
+        response = requests.get(url, timeout=10)
         response.raise_for_status()
         
         data = response.json()