Update app.py

app.py

@@ -1,58 +1,39 @@
-# Step 1: Install/Update necessary libraries
-# Using -U ensures the latest version of google-generativeai is installed
-#!pip install -q -U gradio duckduckgo_search google-generativeai python-dotenv
-
-# --- IMPORTANT: After this cell runs, RESTART THE RUNTIME ---
-# --- (Runtime -> Restart runtime) for the library update to take effect ---
-
-# Step 2: Import libraries
+# Step 1: Import libraries (Ensure these are in requirements.txt)
 import gradio as gr
 import google.generativeai as genai
 from duckduckgo_search import DDGS
 import os
 import textwrap
-#from google.colab import userdata # For Colab secrets
 import traceback # For detailed error logging if needed
 
-# Step 3: Configure API Key (Using Colab Secrets)
-# --- Instructions ---
-# 1. Go to the "Secrets" tab (key icon 🔑) on the left pane in Colab.
-# 2. Click "+ Add a new secret".
-# 3. Set the NAME exactly as: GOOGLE_API_KEY
-# 4. Paste your actual Google AI API key into the VALUE field.
-# 5. Ensure the "Notebook access" toggle is ON for this secret.
-# 6. Do NOT paste your API key directly into this code.
-
-# Initialize flag and key variable
+# --- Step 2: Configure API Key (Using Hugging Face Secrets) ---
 is_api_configured = False
 GOOGLE_API_KEY = None
 
-print("⚙️ Attempting to configure Google API Key...")
+print("⚙️ Attempting to configure Google API Key from HF Space secret...")
 try:
-    GOOGLE_API_KEY = userdata.get('GOOGLE_API_KEY')
+    # Read the secret value set in the Hugging Face Space settings
+    GOOGLE_API_KEY = os.getenv('GOOGLE_API_KEY')
 
     if GOOGLE_API_KEY:
         genai.configure(api_key=GOOGLE_API_KEY)
-        print("✅ Google API Key configured successfully.")
+        print("✅ Google API Key configured successfully from HF secret.")
         is_api_configured = True # Set flag to True ONLY if configure() succeeds
     else:
-        # Secret found but empty - treat as not configured
-        print("⚠️ Error: GOOGLE_API_KEY secret found but is empty.")
+        # Secret variable not found or empty in HF Space settings
+        print("❌ Error: GOOGLE_API_KEY secret not found or is empty in Space settings.")
+        print("➡️ Please go to your Space Settings -> Secrets and ensure 'GOOGLE_API_KEY' is added with your API key value.")
         is_api_configured = False
 
-except userdata.SecretNotFoundError:
-    print("❌ Error: Secret 'GOOGLE_API_KEY' not found.")
-    print("➡️ Please go to 'Secrets' (key icon 🔑) and add GOOGLE_API_KEY with your API key value.")
-    is_api_configured = False # Not configured if secret not found
 except Exception as e:
     print(f"❌ An unexpected error occurred during API Key configuration: {e}")
     is_api_configured = False # Not configured if any other error occurs
-    # traceback.print_exc() # Uncomment for detailed traceback during configuration
+    traceback.print_exc() # Print detailed traceback in HF logs for debugging
 
 # --- End of API Key Configuration ---
 
 
-# Step 4: Define Helper Functions
+# Step 3: Define Helper Functions
 
 # Function to perform web search
 def search_web(query, num_results=7):
@@ -60,8 +41,7 @@ def search_web(query, num_results=7):
     print(f"🔍 Searching the web for: '{query}'...")
     try:
         with DDGS() as ddgs:
-            # Using region='wt-wt' for potentially broader results
-            results = list(ddgs.text(query, region='wt-wt', safesearch='off', max_results=num_results)) # Added safesearch='off'
+            results = list(ddgs.text(query, region='wt-wt', safesearch='off', max_results=num_results))
             if not results:
                 print("⚠️ No search results found.")
                 return "No relevant search results found for the query."
@@ -78,7 +58,7 @@ def search_web(query, num_results=7):
             return context
     except Exception as e:
         print(f"❌ Error during web search: {e}")
-        # traceback.print_exc() # Uncomment for detailed traceback
+        traceback.print_exc() # Log details in HF
         return f"Error occurred during web search. Details: {e}"
 
 # Function to generate the case study using Gemini
@@ -89,7 +69,8 @@ def generate_case_study(topic, search_context):
     # --- Check 1: API Configuration ---
     if not is_api_configured:
          print("❌ Cannot generate: Google API Key not configured successfully.")
-         return "Error: Google API Key not configured successfully. Please check Colab Secrets setup and restart runtime."
+         # Provide error message tailored for HF environment
+         return "Error: Google API Key not configured successfully. Please check the GOOGLE_API_KEY secret in your Hugging Face Space settings and ensure it's correct. The space might need a restart after setting the secret."
 
     # --- Check 2: Search Results ---
     if "Error occurred during web search" in search_context or "No relevant search results found" in search_context:
@@ -97,29 +78,27 @@ def generate_case_study(topic, search_context):
          return f"Cannot generate case study due to search issues:\n{search_context}"
 
     # --- Configure the Gemini model ---
-    model_name = 'gemini-1.5-flash-latest' # Using the recommended modern model
-    # model_name = 'gemini-1.0-pro' # Alternative if flash causes issues
-    # model_name = 'gemini-pro' # Less likely to work based on previous error
+    model_name = 'gemini-1.5-flash-latest'
     try:
         print(f"   Using model: {model_name}")
         model = genai.GenerativeModel(model_name)
     except Exception as e:
         print(f"❌ Error initializing GenerativeModel '{model_name}': {e}")
-        # traceback.print_exc()
+        traceback.print_exc() # Log details in HF
         # Try to list available models if initialization fails
+        error_message = f"Error setting up the AI model '{model_name}': {e}."
         try:
             available_models = [m.name for m in genai.list_models() if 'generateContent' in m.supported_generation_methods]
             print(f"   Available models supporting generateContent: {available_models}")
-            if not available_models:
-                 return f"Error setting up the AI model: {e}. Additionally, no compatible models were found via ListModels."
+            if available_models:
+                 suggested_model = next((m for m in available_models if 'flash' in m or 'pro' in m), available_models[0])
+                 error_message += f" You could try updating the model name in app.py to one of these like: '{suggested_model.split('/')[-1]}'"
             else:
-                 # Suggest trying one of the listed models
-                 suggested_model = next((m for m in available_models if 'flash' in m or 'pro' in m), available_models[0]) # Simple suggestion logic
-                 return f"Error setting up the AI model '{model_name}': {e}. You could try using one of the available models like: '{suggested_model.split('/')[-1]}'"
-
+                 error_message += " Additionally, no compatible models were found via ListModels."
         except Exception as list_e:
              print(f"   Additionally failed to list available models: {list_e}")
-             return f"Error setting up the AI model '{model_name}': {e}. Failed to list alternatives."
+             error_message += " Failed to list alternative models."
+        return error_message
 
 
     # --- Define the Prompt ---
@@ -132,21 +111,20 @@ def generate_case_study(topic, search_context):
     **Required Case Study Format:**
 
     **1. Title:** Create a concise and informative title based on the topic and findings.
-    **2. Introduction/Executive Summary:** Briefly introduce the subject and the core topic (challenge, initiative). State the key outcome *mentioned in the sources*.
+    **2. Introduction/Executive Summary:** Briefly introduce the subject and the core topic. State the key outcome *mentioned in the sources*.
     **3. The Company/Subject:** Provide background information *only from the search results*.
     **4. The Challenge/Problem:** Describe the specific business issue mentioned in the sources.
-    **5. The Solution:** Detail the implemented solution *based only on the sources*. Describe technology use (like GenAI) if mentioned.
+    **5. The Solution:** Detail the implemented solution *based only on the sources*.
     **6. Implementation/Process:** (Optional) Describe *only if information is available in the sources*.
-    **7. Results/Impact:** Quantify results and impact using data *from the sources*. Describe qualitative benefits mentioned. If no results are mentioned, state that.
-    **8. Conclusion:** Summarize key takeaways *based on the provided information*. Reiterate challenge, solution, outcomes *from sources*.
+    **7. Results/Impact:** Quantify results and impact using data *from the sources*. If no results are mentioned, state that.
+    **8. Conclusion:** Summarize key takeaways *based on the provided information*.
     **9. Sources:** List the URLs (`URL:` lines) from the search results that were most relevant.
 
     **Instructions:**
-    *   Adhere strictly to the format above. Use Markdown `##` for section headings (e.g., `## 1. Title`).
-    *   Base your writing ***exclusively*** on the information in the "Provided Search Context". Do not invent information or use external knowledge.
-    *   If details for a section are missing in the sources, explicitly state: "Information not available in the provided sources." Do not leave sections blank.
+    *   Adhere strictly to the format above. Use Markdown `##` for section headings.
+    *   Base your writing ***exclusively*** on the information in the "Provided Search Context". Do not invent information.
+    *   If details for a section are missing in the sources, explicitly state: "Information not available in the provided sources."
     *   Maintain an objective and professional tone.
-    *   Ensure coherence and logical flow.
     *   Format the output using Markdown.
 
     **Provided Search Context:**
@@ -159,68 +137,43 @@ def generate_case_study(topic, search_context):
 
     # --- Generate Content ---
     try:
-        # Optional: Add safety settings if needed
-        # safety_settings = [...]
-        # response = model.generate_content(prompt, safety_settings=safety_settings)
         response = model.generate_content(prompt)
 
         # --- Process Response Safely ---
-        # Check for content parts first
         if response.parts:
             generated_text = "".join(part.text for part in response.parts)
             print("✅ Case study generated successfully.")
             return generated_text
-        # Check for blocking reasons if no parts exist
         elif response.prompt_feedback and response.prompt_feedback.block_reason:
              block_reason = response.prompt_feedback.block_reason
              safety_info = f" Ratings: {response.prompt_feedback.safety_ratings}" if response.prompt_feedback.safety_ratings else ""
              print(f"⚠️ Generation blocked due to: {block_reason}")
              return f"Error: Generation failed. Blocked due to '{block_reason}'.{safety_info} Please try refining your topic or check content policies."
-        # Handle candidates being empty or other unexpected scenarios
         elif not response.candidates:
-             finish_reason = response.candidates[0].finish_reason if response.candidates else "UNKNOWN" # Attempt to get finish reason
+             finish_reason = response.candidates[0].finish_reason if response.candidates else "UNKNOWN"
              print(f"⚠️ Warning: Generation finished without valid content (Finish Reason: {finish_reason}).")
-             return f"Error: The AI model finished generation but produced no usable content (Reason: {finish_reason}). This might indicate an issue with the prompt, the model, or safety filters."
+             return f"Error: The AI model finished generation but produced no usable content (Reason: {finish_reason}). Check model compatibility or prompt complexity."
         else:
-             # Fallback for other unexpected empty response scenarios
              print("⚠️ Warning: Generation finished but produced no text content for unknown reasons.")
-             # You could inspect the raw response object here if needed: print(response)
-             return "Error: The AI model generated an empty response. This might be due to the input, content filters, or a temporary issue."
-
+             return "Error: The AI model generated an empty response. This might be due to input, filters, or a temporary issue."
 
     except Exception as e:
         print(f"❌ Error during case study generation: {e}")
-        # traceback.print_exc() # Uncomment for detailed traceback
+        traceback.print_exc() # Log details in HF
         error_message = f"An unexpected error occurred during AI generation: {e}"
-        # Check for common API errors
+        # Add specific error checks relevant to Gemini API
         if "API key not valid" in str(e) or "PermissionDenied" in str(e) or "AuthenticationError" in str(e):
-            error_message = "Error: Invalid, expired, or missing API Key. Please verify your GOOGLE_API_KEY secret in Colab and ensure the Gemini API is enabled in your Google Cloud project."
-        elif "Model not found" in str(e) or "models/" in str(e) and "is not found" in str(e):
-             # Make error more specific based on the original error message
-             error_message = f"Error: The specified AI model ('{model_name}') was not found or is not supported for generateContent with the current API version. Check the model name or try updating the library (`!pip install -U google-generativeai` + restart runtime)."
-             # Attempt to list models again here might be useful
-             try:
-                 available_models = [m.name for m in genai.list_models() if 'generateContent' in m.supported_generation_methods]
-                 error_message += f"\n   Available models supporting generateContent: {available_models}"
-             except Exception:
-                 error_message += "\n   Failed to retrieve list of available models."
-
+             error_message = "Error: Invalid, expired, or missing API Key. Please double-check the GOOGLE_API_KEY secret in Space settings and ensure the Gemini API is enabled in your Google Cloud project."
+        elif "Model not found" in str(e):
+             error_message = f"Error: The AI model ('{model_name}') was not found or is unsupported. Check the model name in app.py or try updating the google-generativeai library in requirements.txt."
         elif "Resource has been exhausted" in str(e) or "Quota" in str(e):
-             error_message = "Error: API quota exceeded. Please check your usage limits in Google Cloud Console or wait and try again."
-        elif "Invalid API key" in str(e):
-             error_message = "Error: The provided API key is invalid. Please double-check the GOOGLE_API_KEY secret value."
-        elif hasattr(e, 'message'): # General Google API error message structure
-             # Append original message if available and different
-             if str(e) != e.message:
-                 error_message = f"Error during AI generation: {e.message} (Details: {e})"
-             else:
-                 error_message = f"Error during AI generation: {e.message}"
-
-
+             error_message = "Error: API quota exceeded. Check your usage limits in Google Cloud Console."
+        elif hasattr(e, 'message') and str(e) != e.message:
+             error_message = f"Error during AI generation: {e.message} (Details: {e})"
         return error_message
 
 
-# Step 5: Define the main processing function for Gradio
+# Step 4: Define the main processing function for Gradio
 def create_case_study(company_or_topic):
     """Orchestrates the web search and case study generation process."""
     print("-" * 60) # Separator for new request
@@ -240,21 +193,17 @@ def create_case_study(company_or_topic):
     print("-" * 60) # Separator for end of request
     return case_study_markdown
 
-# Step 6: Create and Launch the Gradio Interface
+# Step 5: Create and Launch the Gradio Interface
 print("\n⚙️ Setting up Gradio interface...")
 
-# Add a final check before launching Gradio
+# Add a final check before defining Gradio Interface (optional but good practice)
 if not is_api_configured:
     print("\n" + "="*60)
-    print("‼️ WARNING: Google API Key not configured successfully. ‼️")
-    print("   The Gradio interface will launch, but case study generation WILL FAIL.")
-    print("   Please ensure the 'GOOGLE_API_KEY' secret is correctly set up in Colab")
-    print("   and that you have restarted the runtime after any library updates.")
+    print("‼️ WARNING: Google API Key not configured successfully at startup. ‼️")
+    print("   The Gradio interface will launch, but case study generation WILL FAIL until the API key is correctly set in secrets and the space potentially restarted.")
     print("="*60 + "\n")
-    # Optionally, prevent launch entirely:
-    # print("\n❌ ERROR: Google API Key not configured. Cannot launch Gradio interface.")
-    # exit() # Uncomment to stop execution
 
+# Define the Gradio interface
 iface = gr.Interface(
     fn=create_case_study,
     inputs=gr.Textbox(
@@ -263,11 +212,10 @@ iface = gr.Interface(
         label="Company Name or Topic"
     ),
     outputs=gr.Markdown( # Use Markdown output for better formatting
-        label="Generated Case Study",
-        # line_breaks=True # Uncomment if you prefer single newlines to create <br> tags
+        label="Generated Case Study"
     ),
     title="📄 AI Case Study Generator (Gemini + DuckDuckGo)",
-    description="Enter a company/topic. The app searches the web (DuckDuckGo) and uses Google's Gemini AI to write a case study *based only on the search results*. \n**Requires a correctly configured `GOOGLE_API_KEY` in Colab Secrets (🔑) and runtime restart after library updates.**",
+    description="Enter a company/topic. The app searches the web (DuckDuckGo) and uses Google's Gemini AI to write a case study *based only on the search results*. \n**Requires a correctly configured `GOOGLE_API_KEY` secret in Hugging Face Space Settings.**",
     allow_flagging="never",
     examples=[
         ["How Spotify uses AI for music recommendations"],
@@ -279,11 +227,11 @@ iface = gr.Interface(
 )
 
 print("🚀 Launching Gradio interface...")
-# Launch in debug mode for detailed logs, share=True creates a public link (useful in Colab)
-# Set share=False if you don't need a public link
+
+# Launch the interface (share=True is not needed on HF Spaces)
+# Use debug=True for more detailed logs initially, you can remove it later
 try:
-    iface.launch(debug=True, share=True)
+    iface.launch(debug=True)
 except Exception as e:
     print(f"❌ Failed to launch Gradio interface: {e}")
-    print("   This might be due to networking issues, Colab limitations, or conflicts. Check logs.")
-    # traceback.print_exc() # Uncomment for detailed traceback on launch failure
+    traceback.print_exc() # Log details in HF