gradio_app.py

import gradio as gr
import pandas as pd
import openai
import io

def setup_openai(api_key):
    openai.api_key = api_key
    try:
        # Test API key with a simple request
        response = openai.Completion.create(
            model="gpt-3.5-turbo",
            prompt="Tell me a joke",
            max_tokens=50
        )
        return response.choices[0].text.strip(), None
    except Exception as e:
        return None, f"Error with API key: {str(e)}"

def analyze_data_with_llm(df):
    # Convert the dataframe to a string format to send to the LLM
    data_str = df.head(10).to_string()  # For efficiency, limit to first 10 rows

    # Prepare prompt for the LLM
    prompt = f"Given the following dataset:\n{data_str}\n\nWhat AI/ML features or techniques could benefit from this data?"

    # Query the LLM for suggestions
    response = openai.Completion.create(
        model="text-davinci-003",
        prompt=prompt,
        max_tokens=150,
        temperature=0.7
    )

    return response.choices[0].text.strip()

def process_files(api_key, *files):
    if not api_key:
        return "API Key is required to proceed.", None

    # Set up OpenAI API
    joke, error = setup_openai(api_key)
    if error:
        return error, None

    outputs = []
    outputs.append(f"API Key Test Joke: {joke}\n")

    for file in files:
        if file is None:
            continue
        output = f"📁 File: {file.name}\n"
        try:
            if file.name.endswith(".xlsx"):
                df = pd.read_excel(io.BytesIO(file.read()))
            else:
                df = pd.read_csv(io.StringIO(file.read().decode('utf-8')))

            output += f"Shape: {df.shape}\n\n"
            output += "Preview:\n" + df.head().to_string() + "\n\n"

            # Analyze the data and get suggestions from the LLM
            suggestions = analyze_data_with_llm(df)
            output += "AI/ML Feature Suggestions:\n" + suggestions + "\n\n"

        except Exception as e:
            output += f"Error processing file: {str(e)}\n\n"

        outputs.append(output)

    return "\n".join(outputs), None

# Create Gradio interface
with gr.Blocks(title="📊 CSV/XLSX Data Inspector with AI/ML Feature Suggestions") as iface:
    gr.Markdown("# 📊 CSV/XLSX Data Inspector with AI/ML Feature Suggestions")
    
    api_key_input = gr.Textbox(
        label="Enter your OpenAI API key",
        type="password",
        placeholder="sk-..."
    )
    
    file_input = gr.File(
        label="Upload CSV or XLSX files",
        file_types=[".csv", ".xlsx"],
        file_count="multiple"
    )
    
    output_text = gr.Textbox(label="Results")
    
    submit_btn = gr.Button("Analyze Files")
    
    submit_btn.click(
        fn=process_files,
        inputs=[api_key_input, file_input],
        outputs=[output_text]
    )

iface.launch()