import gradio as gr
import fitz  # PyMuPDF
from models import evaluate_with_gpt, evaluate_with_gemma, evaluate_with_bloom, evaluate_with_jabir, evaluate_with_llama

def extract_text_from_pdf(pdf_file):
    """Extracts and returns the text from a PDF file."""
    document = fitz.open(pdf_file)
    return "".join([page.get_text() for page in document])

def evaluate_resume(resume_text, job_description, model):
    """Evaluates the resume text using the specified model."""
    if model == "GPT-4o":
        return evaluate_with_gpt(resume_text, job_description)
    elif model == "Gemma":
        return evaluate_with_gemma(resume_text, job_description)
    elif model == "Bloom":
        return evaluate_with_bloom(resume_text, job_description)
    elif model == "jabir":
        return evaluate_with_jabir(resume_text, job_description)
    elif model == "llama":
        return evaluate_with_llama(resume_text, job_description)
    else:
        # If "All" is selected, evaluate with all models and return combined results.
        return evaluate_all_models(resume_text, job_description)

def evaluate_multiple_resumes(resume_files, job_description, model):
    """Evaluates multiple resumes and returns the results."""
    results = []
    for resume_file in resume_files:
        title = resume_file.name
        resume_text = extract_text_from_pdf(resume_file)
        result = evaluate_resume(resume_text, job_description, model)
        results.append(f"Result for {title}:\n{result}\n\n")
    return "\n".join(results)

iface = gr.Interface(
    fn=evaluate_multiple_resumes,
    inputs=[
        gr.File(type="filepath", label="Upload Resumes PDF", file_count="multiple"),
        gr.Textbox(lines=10, label="Job Description"),
        gr.Radio(choices=["GPT-4o", "Gemma", "Bloom", "jabir", "llama", "All"], label="Choose Model")
    ],
    outputs="text",
    title="Multiple Resume Evaluator"
)

iface.launch()