main.py

from flask import Flask, render_template, request, jsonify
from flask_cors import CORS
import fitz  # PyMuPDF for PDF text extraction
import spacy
from transformers import T5Tokenizer, T5ForConditionalGeneration
import torch
import os

app = Flask(__name__)
CORS(app)

# ===== Load Custom NER Model =====
try:
    nlp = spacy.load("custom_ner_model")  # Load your custom-trained NER model
    print("Custom NER model loaded successfully.")
except Exception as e:
    print(f"Error loading custom NER model: {e}")
    exit()

# ===== Load T5 Model for Job Title Prediction =====
tokenizer = T5Tokenizer.from_pretrained("t5-base")
model = T5ForConditionalGeneration.from_pretrained("t5-base")

# Load model weights
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
model.load_state_dict(torch.load("best.pth", map_location=device))
model.eval()
model.to(device)

print("T5 model for job title prediction loaded successfully.")

# ===== Helper Functions =====

# Extract text from PDF
def extract_text_from_pdf(pdf_file):
    doc = fitz.open(stream=pdf_file.read(), filetype="pdf")
    text = ""
    for page in doc:
        text += page.get_text()
    return text

# Extract entities using Custom NER
def extract_entities(text):
    text=text.replace("\\n","\n")
    doc = nlp(text)  # Process text with custom NER
    extracted_data = {}
    
    for ent in doc.ents:
        # Use only relevant labels
        if ent.label_ in ["SKILL", "ROLE", "LOCATION", "AREA", "INDUSTRY"]:
            if ent.label_ not in extracted_data:
                extracted_data[ent.label_] = []
            if ent.text not in extracted_data[ent.label_]:
                extracted_data[ent.label_].append(ent.text)
    
    # Format results as comma-separated strings
    for key in extracted_data:
        extracted_data[key] = ", ".join(extracted_data[key])
    return extracted_data

# Predict job title using T5 model
def predict_job_title(skills, area,roles,location,industry):
    input_text = f"Skills: {skills}; \nRole: {roles}; \nLocation: {location}; \nArea: {area}; \nIndustry: {industry}"
    inputs = tokenizer(input_text, return_tensors="pt", padding=True, truncation=True).to(device)
    
    with torch.no_grad():
        outputs = model.generate(inputs["input_ids"], max_length=50, num_beams=4, early_stopping=True)
    
    predicted_job_title = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return predicted_job_title

# ===== Flask Routes =====

@app.route('/')
def home():
    return render_template('index.html')  # Default home page

@app.route('/predict', methods=['POST'])
def predict():
    if 'resume' not in request.files:
        return jsonify({'error': 'No file uploaded'}), 400
    
    file = request.files['resume']
    if file.filename == '':
        return jsonify({'error': 'No file selected'}), 400
    
    if not file.filename.endswith('.pdf'):
        return jsonify({'error': 'Please upload a PDF file'}), 400

    try:
        # Step 1: Extract text from PDF
        resume_text = extract_text_from_pdf(file)
        
        # Step 2: Extract entities using Custom NER
        extracted_data = extract_entities(resume_text)
        
        # Step 3: Prepare input for T5 prediction
        skills = extracted_data.get("SKILL", "")
        area = extracted_data.get("AREA", "")
        roles = extracted_data.get("ROLE", "")
        location = extracted_data.get("LOCATION", "")
        industry = extracted_data.get("INDUSTRY", "")
        # Step 4: Predict job title
        predicted_title = predict_job_title(skills, area,roles,location,industry)
        
        # Step 5: Return response
        response = {
            'success': True,
            'predicted_title': predicted_title.split(";")[0],
            'extracted_skills': extracted_data.get("SKILL", ""),
            'roles': extracted_data.get("ROLE", ""),
            'locations': extracted_data.get("LOCATION", ""),
            'area': extracted_data.get("AREA", ""),
            'industry': extracted_data.get("INDUSTRY", "")
        }
        
        return jsonify(response)
    
    except Exception as e:
        return jsonify({'error': str(e)}), 500

if __name__ == '__main__':
    from waitress import serve
    print("Starting Flask app...")
    serve(app, host="0.0.0.0", port=7860)