app.py

import gradio as gr
from transformers import GPT2LMHeadModel, GPT2Tokenizer
import spacy

# Load GPT-2 model and tokenizer from Hugging Face
model_name = "gpt2"  # You can experiment with different models, e.g., "t5-small"
model = GPT2LMHeadModel.from_pretrained(model_name)
tokenizer = GPT2Tokenizer.from_pretrained(model_name)

# Load SpaCy's small English model for text processing
nlp = spacy.load('en_core_web_sm')

# Preprocess the input text
def preprocess_text(text):
    # Process the text using SpaCy
    doc = nlp(text)
    processed_text = " ".join([token.text for token in doc if not token.is_stop and not token.is_punct])
    return processed_text

# Function to generate the cover letter based on input job description and user profile
def generate_cover_letter(job_description, user_profile):
    # Combine job description and user profile into a single prompt
    prompt = f"Job Description: {job_description}\n\nUser Profile: {user_profile}\n\nCover Letter:"

    # Encode the prompt
    inputs = tokenizer(prompt, return_tensors="pt")

    # Generate the cover letter text
    outputs = model.generate(inputs['input_ids'], max_length=500, num_return_sequences=1, no_repeat_ngram_size=2, top_p=0.95, temperature=0.7)

    # Decode the generated text
    cover_letter = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return cover_letter

# Create Gradio interface
def gradio_interface(job_description, user_profile):
    return generate_cover_letter(job_description, user_profile)

# Set up Gradio UI
iface = gr.Interface(
    fn=gradio_interface,
    inputs=[
        gr.Textbox(label="Job Description", placeholder="Enter job description here...", lines=5),
        gr.Textbox(label="User Profile (Qualifications, Experience, Certificates)", placeholder="Enter your qualifications, experience, and certificates...", lines=5)
    ],
    outputs="text",
    title="Cover Letter Generator",
    description="Enter a job description and your profile, and the bot will generate a personalized cover letter."
)

# Launch the interface (for Hugging Face deployment)
iface.launch(share=True)