import gradio as gr
from huggingface_hub import InferenceClient
from PyPDF2 import PdfFileReader
import requests
from bs4 import BeautifulSoup

# Initialize the Inference Client
client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")

def extract_text_from_pdf(file):
    reader = PdfFileReader(file)
    text = ""
    for page in range(reader.getNumPages()):
        text += reader.getPage(page).extract_text()
    return text

def ats_friendly_checker(file):
    resume_text = extract_text_from_pdf(file)
    # Implement ATS-friendly checker logic using LLM
    system_message = "Evaluate the following resume for ATS-friendliness and provide a score and feedback."
    message = resume_text
    response = client.chat_completion(
        [{"role": "system", "content": system_message}, {"role": "user", "content": message}],
        max_tokens=512,
        temperature=0.7,
        top_p=0.95
    ).choices[0].message["content"]
    
    score = response.split("\n")[0].split(":")[-1].strip()
    feedback = "\n".join(response.split("\n")[1:])
    return score, feedback

def scrape_job_description(url):
    response = requests.get(url)
    soup = BeautifulSoup(response.text, 'html.parser')
    job_description = soup.get_text(separator=" ", strip=True)
    return job_description

def resume_match_checker(file, job_url):
    resume_text = extract_text_from_pdf(file)
    job_description = scrape_job_description(job_url)
    # Implement resume match checker logic using LLM
    system_message = "Compare the following resume with the job description and provide a match score."
    message = f"Resume: {resume_text}\n\nJob Description: {job_description}"
    response = client.chat_completion(
        [{"role": "system", "content": system_message}, {"role": "user", "content": message}],
        max_tokens=512,
        temperature=0.7,
        top_p=0.95
    ).choices[0].message["content"]

    match_score = response.split(":")[-1].strip()
    return match_score

def resume_quality_score(file):
    resume_text = extract_text_from_pdf(file)
    # Implement resume quality scoring logic using LLM
    system_message = "Evaluate the following resume for overall quality and provide a score."
    message = resume_text
    response = client.chat_completion(
        [{"role": "system", "content": system_message}, {"role": "user", "content": message}],
        max_tokens=512,
        temperature=0.7,
        top_p=0.95
    ).choices[0].message["content"]

    quality_score = response.split(":")[-1].strip()
    return quality_score

def text_to_overleaf(resume_text):
    # Implement the conversion to Overleaf code using LLM
    system_message = "Convert the following resume text to Overleaf code."
    message = resume_text
    response = client.chat_completion(
        [{"role": "system", "content": system_message}, {"role": "user", "content": message}],
        max_tokens=512,
        temperature=0.7,
        top_p=0.95
    ).choices[0].message["content"]
    
    overleaf_code = response
    return overleaf_code

# Define the Gradio interface
with gr.Blocks() as demo:
    gr.Markdown("# Resume Enhancement Tool\nEnhance your resume with the following features.")
    
    with gr.Tab("ATS-Friendly Checker"):
        with gr.Row():
            resume = gr.File(label="Upload your Resume (PDF)")
            score = gr.Number(label="ATS Score", interactive=False)
            feedback = gr.Textbox(label="Feedback", interactive=False)
        resume.upload(ats_friendly_checker, resume, [score, feedback])
    
    with gr.Tab("Resume Match Checker"):
        with gr.Row():
            resume = gr.File(label="Upload your Resume (PDF)")
            job_url = gr.Textbox(label="Job Description URL")
            match_score = gr.Number(label="Match Score", interactive=False)
        gr.Button("Check Match").click(resume_match_checker, [resume, job_url], match_score)
    
    with gr.Tab("Resume Quality Score"):
        with gr.Row():
            resume = gr.File(label="Upload your Resume (PDF)")
            quality_score = gr.Number(label="Quality Score", interactive=False)
        resume.upload(resume_quality_score, resume, quality_score)
    
    with gr.Tab("Text to Overleaf Code"):
        with gr.Row():
            resume_text = gr.Textbox(label="Resume Text")
            overleaf_code = gr.Textbox(label="Overleaf Code", interactive=False)
        resume_text.submit(text_to_overleaf, resume_text, overleaf_code)

    gr.Markdown("---\nBuilt with love by [Bahae Eddine HALIM](https://www.linkedin.com/in/halimbahae/)")

if __name__ == "__main__":
    demo.launch()