import gradio as gr
import textstat
from langchain_huggingface import HuggingFaceEndpoint
import os

# Set up Hugging Face API token and model endpoint
HF_TOKEN = os.getenv("HF_TOKEN")  # Ensure you have your token set in your environment
print(HF_TOKEN)
llm = HuggingFaceEndpoint(
    repo_id="mistralai/Mistral-7B-Instruct-v0.3",
    huggingfacehub_api_token=HF_TOKEN.strip(),
    temperature=0.7,
    max_new_tokens=200
)

def check_and_improve_seo(content):
    # Define basic SEO criteria
    keywords = ["SEO", "content", "optimization", "keywords", "readability"]
    keyword_found = any(keyword.lower() in content.lower() for keyword in keywords)
    
    # Check readability score
    readability_score = textstat.flesch_reading_ease(content)
    
    # Prepare a prompt for the LLM to improve content
    prompt = (
        "Optimize the following content for SEO. Ensure it includes relevant keywords, "
        "is easy to read, and meets SEO best practices.\n\n"
        "Content:\n" + content
    )
    
    # Generate SEO-optimized content using the Hugging Face model
    response = llm(prompt)
    optimized_content = response

    # Define SEO checks
    seo_checks = {
        "Keywords Present": keyword_found,
        "Readability Score (Flesch)": readability_score,
        "Optimized Content": optimized_content
    }
    
    return seo_checks

# Define Gradio interface
interface = gr.Interface(
    fn=check_and_improve_seo,
    inputs=gr.Textbox(lines=10, placeholder="Enter your content here..."),
    outputs="json",
    title="SEO Compatibility Checker and Optimizer",
    description="Check if the given content is SEO compatible and get an improved version based on SEO best practices."
)

# Launch the app
if __name__ == "__main__":
    interface.launch()