Update app.py

app.py

@@ -6,26 +6,53 @@ from datetime import datetime
 from typing import Optional
 from smolagents import CodeAgent, HfApiModel, tool
 from duckduckgo_search import DDGS
+from transformers import BlipProcessor, BlipForConditionalGeneration
 from PIL import Image
-import markdownify
 
 # === Tool 1: Current Time Tool ===
 @tool
 def get_current_time_in_timezone(timezone: Optional[str] = None) -> str:
+    """
+    Returns the current time. If a timezone is provided, it is noted in the message.
+
+    Args:
+        timezone: An optional timezone name or location.
+
+    Returns:
+        A string with the current UTC time or specified timezone info.
+    """
     now = datetime.utcnow()
     if timezone:
         return f"The current time in {timezone} is approximately {now.strftime('%Y-%m-%d %H:%M:%S')} UTC."
-    return f"The current UTC time is {now.strftime('%Y-%m-%d %H:%M:%S')}"
+    return f"The current UTC time is {now.strftime('%Y-%m-%d %H:%M:%S')}."
 
 # === Tool 2: CSV Parsing Tool ===
 @tool
 def parse_csv(file_path: str) -> str:
+    """
+    Parses a CSV file and returns the first few rows.
+
+    Args:
+        file_path: Path to the CSV file to be parsed.
+
+    Returns:
+        A preview of the first rows as a string.
+    """
     df = pd.read_csv(file_path)
     return df.head(5).to_string(index=False)
 
 # === Tool 3: Audio Transcription Tool ===
 @tool
 def transcribe_audio(audio_path: str) -> str:
+    """
+    Transcribes spoken words in an audio file.
+
+    Args:
+        audio_path: Path to the .mp3 or .wav file.
+
+    Returns:
+        A transcription of the audio content as a string.
+    """
     import whisper
     model = whisper.load_model("base")
     result = model.transcribe(audio_path)
@@ -34,38 +61,57 @@ def transcribe_audio(audio_path: str) -> str:
 # === Tool 4: Web Search Tool ===
 @tool
 def duckduckgo_search(query: str) -> str:
+    """
+    Performs a DuckDuckGo web search and returns the top 3 results.
+
+    Args:
+        query: The search query string.
+
+    Returns:
+        A formatted string with the top 3 search result titles and URLs.
+    """
     results = []
     with DDGS() as ddgs:
-        for r in ddgs.text(query):
-            results.append(f"{r['title']} - {r['href']}")
-            if len(results) >= 3:
-                break
-    return "\n".join(results) if results else "No results found."
+        for r in ddgs.text(query, region='wt-wt', safesearch='Off', max_results=3):
+            results.append(f"- {r['title']}\n{r['href']}")
+    return "\n\n".join(results) if results else "No results found."
 
 # === Tool 5: Image Captioning Tool ===
 @tool
 def describe_image_with_blip(image_path: str) -> str:
-    from transformers import BlipProcessor, BlipForConditionalGeneration
-    import torch
+    """
+    Generates a caption describing the contents of an image.
 
+    Args:
+        image_path: Path to the image file to be described.
+
+    Returns:
+        A string caption generated by the BLIP vision-language model.
+    """
     processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
     model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")
-
-    raw_image = Image.open(image_path).convert('RGB')
-    inputs = processor(raw_image, return_tensors="pt")
+    image = Image.open(image_path).convert("RGB")
+    inputs = processor(image, return_tensors="pt")
     out = model.generate(**inputs)
     caption = processor.decode(out[0], skip_special_tokens=True)
     return caption
 
+# === Constants ===
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
+# === Agent Definition ===
 class BasicAgent:
     def __init__(self):
         model = HfApiModel(model_id="Qwen/Qwen2.5-Coder-32B-Instruct")
         self.agent = CodeAgent(
             model=model,
-            tools=[get_current_time_in_timezone, parse_csv, transcribe_audio, duckduckgo_search, describe_image_with_blip],
-            add_base_tools=False,
+            tools=[
+                get_current_time_in_timezone,
+                parse_csv,
+                transcribe_audio,
+                duckduckgo_search,
+                describe_image_with_blip
+            ],
             max_steps=6,
             name="EvalAgent",
             description="Certification agent using vision, search, time, audio, and csv tools"
@@ -79,6 +125,7 @@ class BasicAgent:
             response += f"\nFinal Answer: {last_line}"
         return response
 
+# === Evaluation Logic ===
 def run_and_submit_all(profile: gr.OAuthProfile | None):
     space_id = os.getenv("SPACE_ID")
 
@@ -155,6 +202,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     except Exception as e:
         return f"Submission Failed: {e}", pd.DataFrame(results_log)
 
+# === Gradio UI ===
 with gr.Blocks() as demo:
     gr.Markdown("# Basic Agent Evaluation Runner")
     gr.Markdown("""