fix: harmonizing of model names and selectors

Author: jasperan

agentic_rag/gradio_app.py

@@ -298,46 +298,46 @@ def create_interface():
         # HF models first if token is available
         if hf_token:
             model_choices.extend([
-                "Local (Mistral)", 
-                "Local (Mistral) - 4-bit Quantized",
-                "Local (Mistral) - 8-bit Quantized",
+                "mistral", 
+                "mistral-4bit",
+                "mistral-8bit",
             ])
         # Then Ollama models (don't require HF token)
         model_choices.extend([
-            "Ollama - llama3",
-            "Ollama - phi-3",
-            "Ollama - qwen2",
+            "llama3",
+            "phi-3",
+            "qwen2",
             # New Ollama models
-            "Ollama - gemma3:1b",
-            "Ollama - gemma3",
-            "Ollama - gemma3:12b", 
-            "Ollama - gemma3:27b",
-            "Ollama - qwq",
-            "Ollama - deepseek-r1",
-            "Ollama - deepseek-r1:671b",
-            "Ollama - llama3.3",
-            "Ollama - llama3.2",
-            "Ollama - llama3.2:1b",
-            "Ollama - llama3.2-vision",
-            "Ollama - llama3.2-vision:90b",
-            "Ollama - llama3.1",
-            "Ollama - llama3.1:405b",
-            "Ollama - phi4",
-            "Ollama - phi4-mini",
-            "Ollama - mistral",
-            "Ollama - moondream",
-            "Ollama - neural-chat",
-            "Ollama - starling-lm",
-            "Ollama - codellama",
-            "Ollama - llama2-uncensored",
-            "Ollama - llava",
-            "Ollama - granite3.2"
+            "gemma3:1b",
+            "gemma3",
+            "gemma3:12b", 
+            "gemma3:27b",
+            "qwq",
+            "deepseek-r1",
+            "deepseek-r1:671b",
+            "llama3.3",
+            "llama3.2",
+            "llama3.2:1b",
+            "llama3.2-vision",
+            "llama3.2-vision:90b",
+            "llama3.1",
+            "llama3.1:405b",
+            "phi4",
+            "phi4-mini",
+            "mistral",
+            "moondream",
+            "neural-chat",
+            "starling-lm",
+            "codellama",
+            "llama2-uncensored",
+            "llava",
+            "granite3.2"
         ])
         if openai_key:
-            model_choices.append("OpenAI")
+            model_choices.append("openai")
 
-        # Set default model to Ollama - qwen2
-        default_model = "Ollama - qwen2"
+        # Set default model to qwen2
+        default_model = "qwen2"
 
         # Model Management Tab (First Tab)
         with gr.Tab("Model Management"):

agentic_rag/local_rag_agent.py

@@ -170,35 +170,25 @@ def __init__(self, vector_store: VectorStore = None, model_name: str = None,
         self.model_name = model_name
         # skip_analysis parameter kept for backward compatibility but no longer used
 
-        # Check if this is an Ollama model
-        self.is_ollama = model_name and (model_name.startswith("ollama:") or "Ollama - " in model_name)
+        # Check if this is an Ollama model (anything not Mistral is considered Ollama)
+        self.is_ollama = not (model_name and "mistral" in model_name.lower())
 
         if self.is_ollama:
-            # Extract the actual model name from the prefix
-            # If model_name contains 'ollama:' prefix, remove it
-            # If model_name is from gradio interface (e.g., "Ollama - llama3"), extract just the model name
-            if model_name.startswith("ollama:"):
-                ollama_model_name = model_name.replace("ollama:", "")
-            elif "Ollama - " in model_name:
-                ollama_model_name = model_name.replace("Ollama - ", "").strip()
-            else:
-                ollama_model_name = model_name
-            
             # Add :latest suffix if not present
-            if not ollama_model_name.endswith(":latest"):
-                ollama_model_name = f"{ollama_model_name}:latest"
+            if not model_name.endswith(":latest"):
+                model_name = f"{model_name}:latest"
 
             # Load Ollama model
             print("\nLoading Ollama model...")
-            print(f"Model: {ollama_model_name}")
+            print(f"Model: {model_name}")
             print("Note: Make sure Ollama is running on your system.")
 
             # Initialize Ollama model handler
-            self.ollama_handler = OllamaModelHandler(ollama_model_name)
+            self.ollama_handler = OllamaModelHandler(model_name)
 
             # Create pipeline-like interface
             self.pipeline = self.ollama_handler
-            print(f"Using Ollama model: {ollama_model_name}")
+            print(f"Using Ollama model: {model_name}")
         else:
             # Only initialize Mistral if no model is specified
             if not model_name: