app.py

import gradio as gr
import pandas as pd
import requests
from transformers import GPT2LMHeadModel, GPT2Tokenizer, AutoTokenizer, pipeline, AutoModelForCausalLM
from huggingface_hub import HfFolder
from io import StringIO
import os
import torch

# Access the Hugging Face API token from environment variables
hf_token = os.getenv('HF_API_TOKEN')

if not hf_token:
    raise ValueError("Hugging Face API token is not set. Please set the HF_API_TOKEN environment variable.")
HfFolder.save_token(hf_token)

# Set environment variable to avoid floating-point errors
os.environ['TF_ENABLE_ONEDNN_OPTS'] = '0'

# Load the tokenizer and model
tokenizer = GPT2Tokenizer.from_pretrained('gpt2')
model_gpt2 = GPT2LMHeadModel.from_pretrained('gpt2')

# Create a pipeline for text generation using GPT-2
text_generator = pipeline("text-generation", model=model_gpt2, tokenizer=tokenizer)

# Lazy loading function for Llama-3 model
model_llama = None
tokenizer_llama = None

def load_llama_model():
    global model_llama, tokenizer_llama
    if model_llama is None:
        model_name = "meta-llama/Meta-Llama-3.1-8B"
        model_llama = AutoModelForCausalLM.from_pretrained(
            model_name,
            torch_dtype=torch.float16,  # Use FP16 for reduced memory
            use_auth_token=hf_token
        )
        tokenizer_llama = AutoTokenizer.from_pretrained(model_name, token=hf_token)

# Define your prompt template
prompt_template = """\
You are an expert in generating synthetic data for machine learning models.
Your task is to generate a synthetic tabular dataset based on the description provided below.
Description: {description}
The dataset should include the following columns: {columns}
Please provide the data in CSV format with a minimum of 100 rows per generation.
Ensure that the data is realistic, does not contain any duplicate rows, and follows any specific conditions mentioned.
Example Description:
Generate a dataset for predicting house prices with columns: 'Size', 'Location', 'Number of Bedrooms', 'Price'
Example Output:
Size,Location,Number of Bedrooms,Price
1200,Suburban,3,250000
900,Urban,2,200000
1500,Rural,4,300000
...
Description:
{description}
Columns:
{columns}
Output: """

def preprocess_user_prompt(user_prompt):
    generated_text = text_generator(user_prompt, max_length=60, num_return_sequences=1)[0]["generated_text"]
    return generated_text

def format_prompt(description, columns):
    processed_description = preprocess_user_prompt(description)
    prompt = prompt_template.format(description=processed_description, columns=",".join(columns))
    return prompt

generation_params = {
    "top_p": 0.90,
    "temperature": 0.8,
    "max_new_tokens": 512,
    "return_full_text": False,
    "use_cache": False
}

def generate_synthetic_data(description, columns):
    try:
        # Load the Llama model only when generating data
        load_llama_model()
        
        # Prepare the input for the Llama model
        formatted_prompt = format_prompt(description, columns)
        
        # Tokenize the prompt
        inputs = tokenizer_llama(formatted_prompt, return_tensors="pt").to(model_llama.device)
        
        # Generate synthetic data
        with torch.no_grad():
            outputs = model_llama.generate(
                **inputs,
                max_length=512,
                top_p=generation_params["top_p"],
                temperature=generation_params["temperature"],
                num_return_sequences=1
            )
        
        # Decode the generated output
        generated_text = tokenizer_llama.decode(outputs[0], skip_special_tokens=True)
        
        # Return the generated synthetic data
        return generated_text
    except Exception as e:
        print(f"Error in generate_synthetic_data: {e}")
        return f"Error: {e}"


def generate_large_synthetic_data(description, columns, num_rows=1000, rows_per_generation=100):
    data_frames = []
    num_iterations = num_rows // rows_per_generation

    for _ in range(num_iterations):
        generated_data = generate_synthetic_data(description, columns)
        if "Error" in generated_data:
            return generated_data
        df_synthetic = process_generated_data(generated_data)
        data_frames.append(df_synthetic)
    
    return pd.concat(data_frames, ignore_index=True)

def process_generated_data(csv_data):
    data = StringIO(csv_data)
    df = pd.read_csv(data)
    return df

def main(description, columns):
    description = description.strip()
    columns = [col.strip() for col in columns.split(',')]
    df_synthetic = generate_large_synthetic_data(description, columns)
    if isinstance(df_synthetic, str) and "Error" in df_synthetic:
        return df_synthetic  # Return the error message if any
    return df_synthetic.to_csv(index=False)

iface = gr.Interface(
    fn=main,
    inputs=[
        gr.Textbox(label="Description", placeholder="e.g., Generate a dataset for predicting students' grades"),
        gr.Textbox(label="Columns (comma-separated)", placeholder="e.g., name, age, course, grade")
    ],
    outputs="text",
    title="Synthetic Data Generator",
    description="Generate synthetic tabular datasets based on a description and specified columns."
)

iface.api_name = "generate"

# Run the Gradio app
iface.launch(server_name="0.0.0.0", server_port=7860)