app.py

import gradio as gr
import requests
from bs4 import BeautifulSoup
from nltk import download, sent_tokenize
import google.generativeai as genai
import os
import re
import tempfile
import asyncio
import time

# Download NLTK data
download('punkt')
download('punkt_tab')

# Configure Gemini API using environment variable
api_key = os.environ.get("GEMINI_API_KEY")
if not api_key:
    raise ValueError("GEMINI_API_KEY not found in environment variables. Please set it in your environment.")

genai.configure(api_key=api_key)

# Use gemini-1.5-flash for faster and more accessible text analysis
try:
    model = genai.GenerativeModel('gemini-1.5-flash')
except Exception as e:
    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.
2. Legal Policy Violations: Flag content that may violate common legal policies (e.g., copyright infringement, defamation, incitement to violence).
3. Crude/Abusive Language: Detect crude, offensive, or abusive language.
4. Sensitive Topics: Identify content related to sensitive topics such as racism, gender bias, or other forms of discrimination.

Return the results in the following markdown format:
# Blog Review Report

## Grammar Corrections
1. [Heading of issue]
   - 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):
    try:
        response = requests.get(url, timeout=10)
        response.raise_for_status()
        soup = BeautifulSoup(response.text, 'html.parser')
        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, progress=gr.Progress()):
    if text_input and not url_input:
        input_type = "Text"
        input_text = text_input
    elif url_input and not text_input:
        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)

    if not input_text:
        return "Review Blog", "Error: No input provided.", gr.update(visible=False)

    if input_type == "URL":
        progress(0, desc="Fetching content...")
        input_text = await fetch_url_content(input_text)
        if input_text.startswith("Error"):
            return "Review Blog", input_text, gr.update(visible=False)

    sentences = sent_tokenize(input_text)
    analysis_text = "\n".join(sentences)

    progress(0, desc="Generating report...")
    start_time = time.time()
    for i in range(1, 10):
        await asyncio.sleep(1)
        progress(i / 10, desc=f"Generating report... ({int(i * 10)}%)")
        if time.time() - start_time > 30:
            return "Review Blog", "Error: API request timed out after 30 seconds.", gr.update(visible=False)

    try:
        response = await asyncio.to_thread(model.generate_content, PROMPT + "\n\nText to analyze:\n" + analysis_text)
        report = response.text.strip()
        report = re.sub(r'^markdown\n|$', '', report, flags=re.MULTILINE)
    except Exception as e:
        report = f"Error analyzing content with Gemini: {str(e)}"
        print("Available models:")
        for m in genai.list_models():
            print(m.name)
        return "Review Blog", report, gr.update(visible=False)

    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
        progress(1.0, desc="Report generated!")
        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)

# Enhanced custom CSS for a more attractive and user-friendly UI
custom_css = """
@import url('https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&display=swap');
.gradio-container {
    font-family: 'Inter', sans-serif !important;
    background: linear-gradient(135deg, #f0f4f8 0%, #d9e2ec 100%);
    padding: 30px;
    border-radius: 15px;
    box-shadow: 0 6px 25px rgba(0, 0, 0, 0.15);
    color: #2d3748;
}
h1, h3 {
    color: #2b6cb0;
    text-align: center;
    font-weight: 700;
    margin-bottom: 15px;
    text-shadow: 1px 1px 3px rgba(0, 0, 0, 0.1);
}
.review-btn {
    background: linear-gradient(45deg, #4a90e2, #63b3ed);
    color: white;
    font-weight: 600;
    border: none;
    border-radius: 12px;
    padding: 14px 28px;
    transition: all 0.3s ease;
    box-shadow: 0 4px 12px rgba(74, 144, 226, 0.3);
    cursor: pointer;
}
.review-btn:hover {
    background: linear-gradient(45deg, #63b3ed, #4a90e2);
    transform: translateY(-3px);
    box-shadow: 0 6px 15px rgba(74, 144, 226, 0.4);
}
.review-btn:disabled {
    opacity: 0.7;
    cursor: not-allowed;
    transform: none;
    box-shadow: none;
}
.review-btn:disabled::after {
    content: ' ⏳';
}
.tab-nav button {
    font-family: 'Inter', sans-serif;
    font-weight: 600;
    background: #edf2f7;
    color: #2d3748;
    border-radius: 10px 10px 0 0;
    padding: 12px 24px;
    transition: all 0.2s ease;
    border: 1px solid #e2e8f0;
}
.tab-nav button:hover {
    background: #4a90e2;
    color: white;
    border-color: #4a90e2;
}
input, textarea {
    font-family: 'Inter', sans-serif;
    border: 2px solid #e2e8f0;
    border-radius: 10px;
    padding: 12px;
    background: #ffffff;
    box-shadow: 0 2px 6px rgba(0, 0, 0, 0.05);
    color: #2d3748;
    font-size: 16px;
}
.gr-textbox textarea {
    resize: vertical;
    min-height: 150px;
}
.gr-progress {
    background: #edf2f7;
    border-radius: 12px;
    overflow: hidden;
    height: 30px;
}
.gr-progress > div {
    background: linear-gradient(90deg, #48bb78, #38a169);
    height: 100%;
    transition: width 0.5s ease;
}
.markdown-container {
    background: #ffffff;
    border-radius: 12px;
    padding: 25px;
    box-shadow: 0 4px 15px rgba(0, 0, 0, 0.1);
    margin-top: 20px;
    color: #2d3748;
}
.download-btn {
    background: #48bb78;
    color: white;
    border-radius: 10px;
    padding: 12px 24px;
    font-weight: 600;
    transition: all 0.3s ease;
    cursor: pointer;
}
.download-btn:hover {
    background: #38a169;
    transform: translateY(-3px);
}
"""

# Gradio UI with enhanced layout
with gr.Blocks(theme=gr.themes.Monochrome(), css=custom_css) as demo:
    gr.Markdown(
        """
        # 📝 AI Blog Reviewer
        Easily analyze your blog for grammar, legal issues, crude language, and sensitive topics.
        """
    )
    with gr.Row():
        with gr.Column(scale=2):
            with gr.Tabs():
                with gr.TabItem("Text Input"):
                    text_input = gr.Textbox(
                        lines=10,
                        label="Paste Your Blog Content",
                        placeholder="Enter your blog text here...",
                        elem_classes=["input-text"]
                    )
                with gr.TabItem("URL Input"):
                    url_input = gr.Textbox(
                        lines=1,
                        label="Enter Blog URL",
                        placeholder="https://example.com/blog",
                        elem_classes=["input-url"]
                    )
        with gr.Column(scale=1):
            gr.Markdown(
                """
                ### How It Works
                1. **Input**: Paste text or provide a URL.
                2. **Analysis**: AI checks grammar, legal issues, and more.
                3. **Report**: Download a detailed markdown report.
                """
            )
            status_button = gr.Button("Review Blog", elem_classes=["review-btn"])
    
    gr.Markdown("### 📄 Review Report")
    with gr.Row():
        with gr.Column():
            report_output = gr.Markdown(elem_classes=["markdown-container"])
            download_btn = gr.File(label="Download Report", visible=False, elem_classes=["download-btn"])

    status_button.click(
        fn=review_blog,
        inputs=[text_input, url_input],
        outputs=[status_button, report_output, download_btn]
    )

demo.launch()