Update app.py

app.py

@@ -19,6 +19,7 @@ from scripts.text_web_browser import (
     VisitTool,
 )
 from scripts.visual_qa import visualizer
+from scripts.text_cleaner_tool import TextCleanerTool
 
 from smolagents import (
     CodeAgent,
@@ -32,7 +33,7 @@ from smolagents import (
 from smolagents.agent_types import AgentText, AgentImage, AgentAudio
 from smolagents.gradio_ui import pull_messages_from_step, handle_agent_output_types
 
-# ------------------------ Configuration and Setup ------------------------
+
 # Constants and configurations
 AUTHORIZED_IMPORTS = [
     "requests",
@@ -60,24 +61,26 @@ AUTHORIZED_IMPORTS = [
     "clean-text",
 ]
 
-user_agent = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/119.0.0.0 Safari/537.36 Edg/119.0.0.0"
+USER_AGENT = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/119.0.0.0 Safari/537.36 Edg/119.0.0.0"
 BROWSER_CONFIG = {
     "viewport_size": 1024 * 5,
     "downloads_folder": "downloads_folder",
     "request_kwargs": {
-        "headers": {"User-Agent": user_agent},
+        "headers": {"User-Agent": USER_AGENT},
         "timeout": 300,
     },
     "serpapi_key": os.getenv("SERPAPI_API_KEY"),
 }
 
-custom_role_conversions = {"tool-call": "assistant", "tool-response": "user"}
+CUSTOM_ROLE_CONVERSIONS = {"tool-call": "assistant", "tool-response": "user"}
+
 
-# Multimedia file types supported:
 ALLOWED_FILE_TYPES = [
     "application/pdf",
     "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
     "text/plain",
+    "text/markdown",  # Added Markdown support
+    "application/json",  # Added JSON support
     "image/png",
     "image/webp",
     "image/jpeg",  # Added JPEG support
@@ -92,14 +95,14 @@ ALLOWED_FILE_TYPES = [
 def setup_environment():
     """Initialize environment variables and authentication."""
     load_dotenv(override=True)
-    if os.getenv("HF_TOKEN"):  # Check if token is actually set
-        login(os.getenv("HF_TOKEN"))
-        print("HF_TOKEN (last 10 characters):", os.getenv("HF_TOKEN")[-10:])
+    hf_token = os.getenv("HF_TOKEN")
+    if hf_token:  # Check if token is actually set
+        login(hf_token)
+        print("HF_TOKEN (last 10 characters):", hf_token[-10:])
     else:
         print("HF_TOKEN not found in environment variables.")
 
 
-# ------------------------ Model and Tool Management ------------------------
 class ModelManager:
     """Manages model loading and initialization."""
 
@@ -110,13 +113,13 @@ class ModelManager:
             if chosen_inference == "hf_api":
                 return HfApiModel(model_id=model_id)
 
-            elif chosen_inference == "hf_api_provider":
+            if chosen_inference == "hf_api_provider":
                 return HfApiModel(provider="together")
 
-            elif chosen_inference == "litellm":
+            if chosen_inference == "litellm":
                 return LiteLLMModel(model_id=model_id)
 
-            elif chosen_inference == "openai":
+            if chosen_inference == "openai":
                 if not key_manager:
                     raise ValueError("Key manager required for OpenAI model")
 
@@ -124,15 +127,14 @@ class ModelManager:
                     model_id=model_id, api_key=key_manager.get_key("openai_api_key")
                 )
 
-            elif chosen_inference == "transformers":
+            if chosen_inference == "transformers":
                 return TransformersModel(
                     model_id="HuggingFaceTB/SmolLM2-1.7B-Instruct",
                     device_map="auto",
                     max_new_tokens=1000,
                 )
 
-            else:
-                raise ValueError(f"Invalid inference type: {chosen_inference}")
+            raise ValueError(f"Invalid inference type: {chosen_inference}")
 
         except Exception as e:
             print(f"✗ Couldn't load model: {e}")
@@ -169,14 +171,22 @@ class ToolRegistry:
             print(f"✗ Couldn't initialize image generation tool: {e}")
             raise
 
+    @staticmethod
+    def load_clean_text_tool():
+        """Initialize and return image generation tools."""
+        try:
+            return TextCleanerTool
+        except Exception as e:
+            print(f"✗ Couldn't initialize clean text tool: {e}")
+            raise
+
 
-# ------------------------ Agent Creation and Execution ------------------------
 def create_agent():
     """Creates a fresh agent instance with properly configured tools."""
     # Initialize model
     model = LiteLLMModel(
-        custom_role_conversions=custom_role_conversions,
-        model_id="openrouter/google/gemini-2.0-flash-001",  # currently serving:
+        custom_role_conversions=CUSTOM_ROLE_CONVERSIONS,
+        model_id="openrouter/deepseek/deepseek-chat-v3-0324:free",  # currently serving:
     )  # DEEPSEEK = openrouter/perplexity/r1-1776 <--- boss model
 
     # Initialize tools
@@ -186,9 +196,10 @@ def create_agent():
     # Collect all tools in a single list
     web_tools = ToolRegistry.load_web_tools(model, browser, text_limit)
     image_generator = ToolRegistry.load_image_generation_tools()
+    clean_text = ToolRegistry.load_clean_text_tool()
 
     # Combine all tools into a single list (not a tuple)
-    all_tools = [visualizer] + web_tools + [image_generator]
+    all_tools = [visualizer, web_tools, image_generator, clean_text]
 
     # Validate tools before creating agent
     for tool in all_tools:
@@ -200,8 +211,8 @@ def create_agent():
     return CodeAgent(
         model=model,
         tools=all_tools,  # Pass a single list containing all tools
-        max_steps=10,
-        verbosity_level=1,
+        max_steps=12,
+        verbosity_level=2,
         additional_authorized_imports=AUTHORIZED_IMPORTS,
         planning_interval=4,
     )
@@ -245,7 +256,6 @@ def stream_to_gradio(
         )
 
 
-# ------------------------ Gradio UI Components ------------------------
 class GradioUI:
     """A one-line interface to launch your agent in Gradio."""
 
@@ -254,8 +264,7 @@ class GradioUI:
         self.file_upload_folder = file_upload_folder
 
         if self.file_upload_folder is not None:
-            if not os.path.exists(file_upload_folder):
-                os.mkdir(file_upload_folder)
+            os.makedirs(file_upload_folder, exist_ok=True)
 
     def interact_with_agent(self, prompt, messages, session_state):
         """Main interaction handler with the agent."""
@@ -344,7 +353,7 @@ class GradioUI:
         """Process user message and handle file references."""
         message = text_input
 
-        if len(file_uploads_log) > 0:
+        if file_uploads_log:
             message += f"\nYou have been provided with these files, which might be helpful or not: {file_uploads_log}"  # Added file list
 
         return (
@@ -409,7 +418,7 @@ class GradioUI:
             with gr.Sidebar():
                 gr.Markdown(
                     """#OpenDeepResearch - 3theSmolagents!
-                Model_id: google/gemini-2.0-flash-001"""
+                Model_id: deepseek/deepseek-chat-v3-0324:free"""
                 )
                 with gr.Group():
                     gr.Markdown("**What's on your mind mate?**", container=True)
@@ -580,7 +589,6 @@ class GradioUI:
         )
 
 
-# ------------------------ Execution ------------------------
 def main():
     """Main entry point for the application."""
     # Initialize environment