update tools

agent_smolagent.py

@@ -14,7 +14,15 @@ from smolagents.default_tools import (DuckDuckGoSearchTool,
                                       SpeechToTextTool,
                                       PythonInterpreterTool)
 #from final_answer import FinalAnswerTool, check_reasoning, ensure_formatting
-#from tools import (use_vision_model, youtube_frames_to_images,
+from tools_smolagent import (read_file, 
+                             extract_text_from_image, 
+                             analyze_csv_file, 
+                             analyze_excel_file, 
+                             youtube_transcribe, 
+                             transcribe_audio, 
+                             wikipedia_search)
+
+#(use_vision_model, youtube_frames_to_images,
 #                         read_file, 
 #                         extract_text_from_image, analyze_csv_file, 
 #                         analyze_excel_file, youtube_transcribe,
@@ -50,17 +58,20 @@ def build_agent():
     return CodeAgent(
         model=model_desp,
         tools=[#FinalAnswerTool(), 
-            #DuckDuckGoSearchTool(), 
+            DuckDuckGoSearchTool(), 
             VisitWebpageTool(max_output_length=500000), 
-            WikipediaSearchTool(extract_format='HTML'),
+            #WikipediaSearchTool(extract_format='HTML'),
             SpeechToTextTool(),
+            read_file, 
+            extract_text_from_image, 
+            analyze_csv_file, 
+            analyze_excel_file, 
+            youtube_transcribe,
+            transcribe_audio,
+            wikipedia_search
             #youtube_transcribe,
             #    use_vision_model,
             #    youtube_frames_to_images,
-            #    review_youtube_video,
-            #read_file, 
-            #extract_text_from_image, 
-            #analyze_csv_file, analyze_excel_file,
             #transcribe_audio,
             ],
         managed_agents=[],
@@ -195,7 +206,7 @@ if __name__ == "__main__":
     '''
     task = {
         "task_id": "8e867cd7-cff9-4e6c-867a-ff5ddc2550be",
-        "question": "Examine the video at https://www.youtube.com/watch?v=1htKBjuUWec. What does Teal'c say in response to the question ""Isn't that hot?",
+        "question": question1,
         "Level": "1",
         "file_name": "",
     }

requirements.txt

@@ -13,10 +13,12 @@ pydub
 #tavily-python
 #wikipedia
 pytesseract
-#openai-whisper
+openai-whisper
 
 wikipedia-api
 smolagents 
 smolagents[transformers]
 smolagents[openai]
 duckduckgo_search
+yt-dlp
+wikipedia

tools_smolagent.py

@@ -2,16 +2,20 @@ import tempfile
 import requests
 import os
 
-from time import sleep
-from urllib.parse import urlparse
+#from time import sleep
+from dotenv import load_dotenv
+#from urllib.parse import urlparse
 from typing import Optional, List
 import yt_dlp
-from google.genai import types
+import wikipedia
+
+from smolagents import tool
+
+#from google.genai import types
 
 from PIL import Image
-from smolagents import CodeAgent, tool, OpenAIServerModel, LiteLLMModel
-from google import genai
-from dotenv import load_dotenv
+#from google import genai
+#from dotenv import load_dotenv
 #from model_provider import create_react_model, create_vision_model
 #import imageio
 
@@ -112,7 +116,7 @@ def youtube_frames_to_images(url: str, sample_interval_frames: int = 24) -> List
 
         reader.close()
         return images
-'''
+
 @tool
 def review_youtube_video(url: str, question: str) -> str:
     """
@@ -138,7 +142,7 @@ def review_youtube_video(url: str, question: str) -> str:
         return response.text
     except Exception as e:
         return f"Error asking {model} about video: {str(e)}"
-
+'''
 
 @tool
 def read_file(filepath: str ) -> str:
@@ -312,4 +316,34 @@ def transcribe_audio(audio_file_path: str) -> str:
     # Load model
     model = whisper.load_model(model_size)
     result = model.transcribe(audio_file_path)
-    return result['text']
+    return result['text']
+
+
+@tool
+def wikipedia_search(query: str) -> dict:
+    """
+    Search Wikipedia for a given query and return the first 10 results with summaries.
+    
+    Args:
+        query: The search term or topic.
+    Returns:
+        A dictionary with a 'wiki_results' key containing formatted Wikipedia summaries.
+    """
+    wikipedia.set_lang("en")
+    try:
+        results = wikipedia.search(query, results=10)
+        summaries = []
+        for title in results:
+            try:
+                summary = wikipedia.summary(title, sentences=2)
+                summaries.append(f"## {title}\n{summary}")
+            except wikipedia.exceptions.DisambiguationError as e:
+                summaries.append(f"## {title}\nDisambiguation required. Example options: {e.options[:3]}")
+            except wikipedia.exceptions.PageError:
+                summaries.append(f"## {title}\nPage not found.")
+        
+        formatted = "\n\n---\n\n".join(summaries)
+        return {"wiki_results": formatted}
+    
+    except Exception as e:
+        return {"wiki_results": f"Error during Wikipedia search: {str(e)}"}