Update app.py

app.py

@@ -10,6 +10,7 @@ import asyncio
 
 # Download NLTK data
 download('punkt')
+download('punkt_tab')
 
 # Configure Gemini API using environment variable
 api_key = os.environ.get("GEMINI_API_KEY")
@@ -18,8 +19,18 @@ if not api_key:
 
 genai.configure(api_key=api_key)
 
-model = genai.GenerativeModel('gemini-1.5-flash')
-
+# Use gemini-1.5-flash for faster and more accessible text analysis
+try:
+    model = genai.GenerativeModel('gemini-1.5-flash')
+except Exception as e:
+    # Fallback: List available models if the specified model is not found
+    print(f"Error initializing model: {str(e)}")
+    print("Available models:")
+    for m in genai.list_models():
+        print(m.name)
+    raise ValueError("Failed to initialize gemini-1.5-flash. Check available models above and update the model name.")
+
+# Prompt for Gemini to analyze text with specified output format
 PROMPT = """
 You are an AI content reviewer. Analyze the provided text for the following:
 1. *Grammar Issues*: Identify and suggest corrections for grammatical errors.
@@ -33,14 +44,21 @@ Return the results in the following markdown format:
    - CONTENT: [Exact line or part of text with the issue]
    - SUGGESTION: [Suggested correction]
    - ISSUE: [Description of the issue]
+2. [Heading of next issue]
+   - CONTENT: [Exact line or part of text with the issue]
+   - SUGGESTION: [Suggested correction]
+   - ISSUE: [Description of the issue]
+[Continue numbering for additional issues or state "None detected"]
 ## Legal Policy Violations
 - CONTENT: [Exact line or part of text with the issue]
   SUGGESTION: [Suggested action or correction]
   ISSUE: [Description of the legal violation]
+[Or state "None detected"]
 ## Crude/Abusive Language
 - [List instances of crude or abusive language or "None detected"]
 ## Sensitive Topics
 - [List instances of sensitive topics or "None detected"]
+For each issue, provide the exact text, a suggested correction or action, and a concise explanation. Be precise and ensure the output strictly follows the specified format.
 """
 
 async def fetch_url_content(url):
@@ -48,41 +66,73 @@ async def fetch_url_content(url):
         response = requests.get(url, timeout=10)
         response.raise_for_status()
         soup = BeautifulSoup(response.text, 'html.parser')
+        # Extract text from common content tags
         content = ' '.join([p.get_text(strip=True) for p in soup.find_all(['p', 'article', 'div'])])
         return content if content else "No readable content found on the page."
     except Exception as e:
         return f"Error fetching URL: {str(e)}"
 
 async def review_blog(text_input, url_input):
+    # Start loading effect immediately
+    button_text = "Processing..."
+
+    # Determine input type based on which field is populated
     if text_input and not url_input:
+        input_type = "Text"
         input_text = text_input
     elif url_input and not text_input:
-        input_text = await fetch_url_content(url_input)
-        if input_text.startswith("Error"):
-            return "Review Blog", input_text, gr.update(visible=False), ""
+        input_type = "URL"
+        input_text = url_input
     else:
-        return "Review Blog", "Error: Please provide input in either the Text or URL tab, but not both.", gr.update(visible=False), ""
+        return "Review Blog", "Error: Please provide input in either the Text or URL tab, but not both.", gr.update(visible=False)
+
+    # Handle empty input
+    if not input_text:
+        return "Review Blog", "Error: No input provided.", gr.update(visible=False)
 
     try:
-        sentences = sent_tokenize(input_text)
-        analysis_text = "\n".join(sentences)
-        response = await asyncio.to_thread(model.generate_content, PROMPT + "\n\nText to analyze:\n" + analysis_text)
-        report = response.text.strip()
-
-        suggestions = re.findall(r'- CONTENT: (.*?)\n\s*SUGGESTION: (.*?)\n\s*ISSUE: (.*?)\n', report, re.DOTALL)
-        highlighted = input_text
-        for content, suggestion, issue in suggestions:
-            highlighted = highlighted.replace(content, f'<span style="background-color: #ffcccc;" title="{issue} | Suggested: {suggestion}">{content}</span>')
-
-        with tempfile.NamedTemporaryFile(mode='w', suffix='.md', delete=False, encoding='utf-8') as temp_file:
-            temp_file.write(report)
-            temp_file_path = temp_file.name
-
-        report = f"**Report generated, please scroll down to view.**\n\n{report}"
-        return "Review Blog", report, gr.update(visible=True, value=temp_file_path), highlighted
+        async with asyncio.timeout(30):  # 30-second timeout for entire process
+            # Handle URL input
+            if input_type == "URL":
+                button_text = "Fetching content..."
+                input_text = await fetch_url_content(input_text)
+                if input_text.startswith("Error"):
+                    return "Review Blog", input_text, gr.update(visible=False)
+
+            # Tokenize input for analysis
+            sentences = sent_tokenize(input_text)
+            analysis_text = "\n".join(sentences)
+
+            # Update button for API call
+            button_text = "Generating report..."
+            try:
+                response = await asyncio.to_thread(model.generate_content, PROMPT + "\n\nText to analyze:\n" + analysis_text)
+                report = response.text.strip()
+                # Ensure the response is markdown by removing any code fences
+                report = re.sub(r'^markdown\n|$', '', report, flags=re.MULTILINE)
+            except Exception as e:
+                report = f"Error analyzing content with Gemini: {str(e)}. Please check your API key, network connection, or model availability."
+                # Fallback: List available models for debugging
+                print("Available models:")
+                for m in genai.list_models():
+                    print(m.name)
+                return "Review Blog", report, gr.update(visible=False)
+
+            # Create a temporary file to store the report
+            try:
+                with tempfile.NamedTemporaryFile(mode='w', suffix='.md', delete=False, encoding='utf-8') as temp_file:
+                    temp_file.write(report)
+                    temp_file_path = temp_file.name
+                # Add note to scroll to report
+                report = f"**Report generated, please scroll down to view.**\n\n{report}"
+                return "Review Blog", report, gr.update(visible=True, value=temp_file_path)
+            except Exception as e:
+                return "Review Blog", f"Error creating temporary file: {str(e)}", gr.update(visible=False)
+
     except asyncio.TimeoutError:
-        return "Timeout", "Error: Process timed out after 30 seconds.", gr.update(visible=False), ""
+        return "Timeout", "Error: Process timed out after 30 seconds.", gr.update(visible=False)
 
+# Custom CSS for hover effect, loading state, and Inter font
 custom_css = """
 @import url('https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&display=swap');
 .gradio-container {
@@ -106,6 +156,22 @@ custom_css = """
     opacity: 0.7;
     cursor: not-allowed;
 }
+.review-btn:disabled::before {
+    content: '';
+    display: inline-block;
+    width: 16px;
+    height: 16px;
+    border: 2px solid #fff;
+    border-radius: 50%;
+    border-top-color: transparent;
+    animation: spin 1s linear infinite;
+    margin-right: 8px;
+    vertical-align: middle;
+}
+@keyframes spin {
+    0% { transform: rotate(0deg); }
+    100% { transform: rotate(360deg); }
+}
 .tab-nav button {
     font-family: 'Inter', sans-serif;
     font-weight: 500;
@@ -115,9 +181,10 @@ input, textarea {
 }
 """
 
+# Gradio UI with Tabs
 with gr.Blocks(theme=gr.themes.Monochrome(), css=custom_css) as demo:
     gr.Markdown("# 📝 AI Blog Reviewer")
-    gr.Markdown("Enter blog text or a URL to review. Issues will be highlighted in the content and summarized in a downloadable report.")
+    gr.Markdown("Enter blog text or a URL to review for grammar, legal issues, crude language, and sensitive topics. The report is generated in markdown format.")
 
     with gr.Tabs():
         with gr.TabItem("Text"):
@@ -126,16 +193,16 @@ with gr.Blocks(theme=gr.themes.Monochrome(), css=custom_css) as demo:
             url_input = gr.Textbox(lines=1, label="Blog URL", placeholder="Enter the blog URL here...")
 
     status_button = gr.Button(value="Review Blog", elem_classes=["review-btn"])
+
     gr.Markdown("### 📄 Review Report")
     report_output = gr.Markdown()
     download_btn = gr.File(label="Download Report", visible=False)
-    gr.Markdown("### 🔍 Highlighted Blog Content")
-    highlighted_text = gr.HTML()
 
+    # Bind the review button to process inputs
     status_button.click(
         fn=review_blog,
         inputs=[text_input, url_input],
-        outputs=[status_button, report_output, download_btn, highlighted_text]
+        outputs=[status_button, report_output, download_btn]
     )
 
-demo.launch()
+demo.launch()