app.py

# seo_blog_writer_app.py

import gradio as gr
import google.generativeai as genai
import json
from datetime import datetime
import re
import os

# Example system prompt for Gemini
SYSTEM_PROMPT = '''
You are an expert SEO blog writer for a fermentation-focused website.

Write a long-form SEO blog post targeting the keyword: "{KEYWORD}".

Use the following writing guidelines:
- Tone: {TONE} (e.g., Informative, Conversational, Friendly)
- Point of View: {POV} (e.g., Third Person, Second Person)
- Audience: Beginners interested in fermentation

Content structure:
- Start with a compelling introduction (hook)
- Use <h2> and <h3> tags for subheadings
- Include a bullet-point section with beginner-friendly tips
- Add a step-by-step guide (if applicable)
- Include an FAQ section with at least 3 relevant questions and answers
- Add a strong conclusion with a CTA (Call to Action)
- Add meta title and meta description at the beginning
- End with <script type="application/ld+json"> containing valid FAQ schema

Formatting:
- Output only valid WordPress-ready HTML (no markdown)
- Use readable paragraphs and short sentences
- Emphasize clarity and structure

### 📌 INTERNAL LINK MAP
The following keywords should be hyperlinked when they appear naturally in the content:

{LINKS_JSON}

Use this format for links:
<a href="[URL]">[Keyword]</a>
'''

# Helper to parse keyword:url text into JSON
def parse_links_to_json(raw_text):
    try:
        link_map = {}
        if raw_text.strip():
            lines = raw_text.strip().splitlines()
            for line in lines:
                if ":" in line:
                    parts = line.split(":", 1)
                    keyword = parts[0].strip()
                    url = parts[1].strip()
                    if keyword and url:
                        link_map[keyword] = url
        return json.dumps(link_map, indent=2)
    except Exception as e:
        return "{}"

# Function to insert internal links from user-defined JSON
def insert_internal_links(text, internal_links_json):
    try:
        if not internal_links_json or internal_links_json == "{}":
            return text
            
        links = json.loads(internal_links_json)
        for keyword, url in links.items():
            # Only replace the first occurrence to avoid duplicate links
            if keyword in text and f'href="{url}"' not in text:
                text = text.replace(keyword, f'<a href="{url}">{keyword}</a>', 1)
    except Exception as e:
        return text + f"\n<!-- Error in internal linking: {str(e)} -->"
    return text


def generate_blog(api_key, keyword, tone, pov, internal_links_raw):
    try:
        # Validate inputs
        if not api_key.strip():
            return "Error: Please provide a Gemini API key.", "<p>Please provide a Gemini API key.</p>", gr.File(visible=False)
        
        if not keyword.strip():
            return "Error: Please provide a keyword.", "<p>Please provide a keyword.</p>", gr.File(visible=False)

        # Configure Gemini
        genai.configure(api_key=api_key)
        model = genai.GenerativeModel("gemini-1.5-pro")

        # Parse internal links
        internal_links_json = parse_links_to_json(internal_links_raw)
        
        # Format links for the prompt
        formatted_links = "No internal links provided."
        if internal_links_json != "{}":
            try:
                links_dict = json.loads(internal_links_json)
                formatted_links = "\n".join([f"- {k}: {v}" for k, v in links_dict.items()])
            except:
                formatted_links = "Error parsing internal links."

        # Create the prompt
        prompt = SYSTEM_PROMPT.format(
            KEYWORD=keyword,
            TONE=tone,
            POV=pov,
            LINKS_JSON=formatted_links
        )

        # Generate content
        response = model.generate_content(prompt)
        raw_html = response.text

        # Insert internal links
        linked_html = insert_internal_links(raw_html, internal_links_json)

        # Create clean preview HTML without custom styling
        preview = f"""
<div style="max-height: 500px; overflow-y: auto; padding: 10px;">
    <p><strong>Keyword:</strong> {keyword} | <strong>Tone:</strong> {tone} | <strong>POV:</strong> {pov}</p>
    <hr>
    {linked_html}
</div>
"""

        # Create full HTML file for download (clean version)
        full_html = f"""
<!DOCTYPE html>
<html>
<head>
    <meta charset='utf-8'>
    <title>{keyword.title()}</title>
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
</head>
<body>
    <h1>SEO Blog Post: {keyword.title()}</h1>
    <p><em>Tone: {tone} | Point of View: {pov}</em></p>
    <hr>
    {linked_html}
</body>
</html>
"""

        # Create filename and save file
        filename = f"seo_post_{datetime.now().strftime('%Y%m%d_%H%M%S')}.html"
        
        # Save to current directory
        with open(filename, "w", encoding="utf-8") as f:
            f.write(full_html)

        return linked_html, preview, gr.File(value=filename, visible=True)

    except Exception as e:
        error_msg = f"Error generating blog post: {str(e)}"
        return error_msg, f"<p>{error_msg}</p>", gr.File(visible=False)


def test_faq_schema(blog_html):
    if not blog_html or "Error" in blog_html[:100]:
        return "<p>No content to analyze.</p>"
    
    try:
        if "application/ld+json" in blog_html:
            # Extract schema
            start_tag = "<script type=\"application/ld+json\">"
            end_tag = "</script>"
            
            start_idx = blog_html.find(start_tag)
            if start_idx != -1:
                start_idx += len(start_tag)
                end_idx = blog_html.find(end_tag, start_idx)
                if end_idx != -1:
                    schema_content = blog_html[start_idx:end_idx].strip()
                    return f"""
                    <div>
                        <h4>✅ FAQ Schema Detected</h4>
                        <details>
                            <summary>View Schema Code</summary>
                            <pre>{schema_content}</pre>
                        </details>
                    </div>
                    """
        
        return "<div><h4>⚠️ No FAQ Schema Found</h4><p>The generated content doesn't include FAQ structured data.</p></div>"
    
    except Exception as e:
        return f"<p>Error analyzing schema: {str(e)}</p>"


# Create the Gradio interface
with gr.Blocks(theme=gr.themes.Soft(), title="SEO Blog Generator") as demo:
    gr.Markdown("""
    # 🧠 SEO Blog Article Generator
    
    Generate WordPress-ready blog content with metadata, FAQ, internal links, and schema using Gemini 1.5 Pro.
    
    **Quick Start:**
    1. 🔑 Enter your Gemini API key
    2. 🎯 Specify your target keyword  
    3. 📝 Choose tone and point of view
    4. 🔗 Add internal links (optional)
    5. 🚀 Click Generate!
    """)

    # API Key Section
    gr.Markdown("### 🔐 API Configuration")
    api_key_input = gr.Textbox(
        label="Gemini API Key", 
        type="password",
        placeholder="Enter your Google Gemini API key...",
        info="Get your API key from Google AI Studio"
    )

    # Content Settings
    gr.Markdown("### ⚙️ Content Settings")
    with gr.Row():
        keyword_input = gr.Textbox(
            label="Target SEO Keyword", 
            placeholder="e.g., how to ferment carrots at home",
            scale=2
        )
        tone_input = gr.Dropdown(
            ["Informative", "Conversational", "Friendly", "Professional"], 
            value="Friendly", 
            label="Writing Tone",
            scale=1
        )
        pov_input = gr.Dropdown(
            ["First Person", "Second Person", "Third Person"], 
            value="Second Person", 
            label="Point of View",
            scale=1
        )

    # Internal Links Section
    gr.Markdown("### 🔗 Internal Links (Optional)")
    internal_links_box = gr.Textbox(
        label="Internal Links", 
        lines=3, 
        placeholder="kefir:https://example.com/kefir-guide\nkombucha:https://example.com/kombucha-basics",
        info="Format: keyword:url (one per line)"
    )

    # Generate Button
    generate_btn = gr.Button("🚀 Generate Blog Post", variant="primary", size="lg")
    
    # Results Section
    gr.Markdown("### 📊 Generated Content")
    with gr.Row():
        with gr.Column():
            result_html = gr.Textbox(
                label="Generated HTML Code", 
                lines=12,
                info="Copy this HTML to use in WordPress",
                max_lines=20,
                show_copy_button=True
            )
        with gr.Column():
            preview_html = gr.HTML(label="Live Preview")

    # Analysis & Download
    gr.Markdown("### 🔍 Analysis & Download")
    with gr.Row():
        with gr.Column():
            faq_test = gr.HTML(label="Schema Analysis")
        with gr.Column():
            download_file = gr.File(
                label="Download HTML File", 
                interactive=False,
                visible=True
            )

    # Event handlers
    generate_btn.click(
        fn=generate_blog,
        inputs=[api_key_input, keyword_input, tone_input, pov_input, internal_links_box],
        outputs=[result_html, preview_html, download_file]
    )

    result_html.change(
        fn=test_faq_schema, 
        inputs=[result_html], 
        outputs=[faq_test]
    )

    # Add examples
    gr.Markdown("### 💡 Example Inputs")
    gr.Examples(
        examples=[
            ["how to ferment vegetables at home", "Conversational", "Second Person", "fermentation:https://example.com/fermentation-guide\nvegetables:https://example.com/vegetable-guide"],
            ["benefits of drinking kefir daily", "Informative", "Third Person", "kefir:https://example.com/kefir-guide\nprobiotics:https://example.com/probiotics-benefits"],
            ["making kombucha for beginners", "Friendly", "Second Person", "kombucha:https://example.com/kombucha-basics\nSCOBY:https://example.com/what-is-scoby"]
        ],
        inputs=[keyword_input, tone_input, pov_input, internal_links_box]
    )

if __name__ == "__main__":
    demo.launch(
        share=True, 
        server_name="0.0.0.0",
        inbrowser=False,
        show_error=True,
        quiet=False
    )