app.py

import streamlit as st
import random
from transformers import pipeline
import pandas as pd

generator = pipeline('text-generation', model='gpt2')
max_length = 50

prompts = {
    "Difficulty sleeping": [
        "Try keeping a consistent sleep schedule and avoid caffeine before bedtime.",
        "Make your bedroom a comfortable and calming environment.",
        "Avoid using electronic devices before bedtime.",
        "Try relaxation techniques like deep breathing or meditation.",
        "Consider talking to a healthcare provider if sleep problems persist."
    ],
    "Time management": [
        "Use a planner or time-tracking app to prioritize tasks and stay on schedule.",
        "Break down large tasks into smaller ones.",
        "Limit multitasking and focus on one task at a time.",
        "Delegate tasks to others when possible.",
        "Take regular breaks and avoid overworking yourself."
    ],
    "Stress management": [
        "Practice mindfulness techniques such as deep breathing or meditation.",
        "Get regular exercise to reduce stress and improve mood.",
        "Get enough sleep and practice good sleep habits.",
        "Take breaks throughout the day to reduce stress levels.",
        "Try to identify the sources of stress in your life and develop strategies to manage them."
    ]
}

def generate_prompt(prompt):
    solution = random.choice(prompts[prompt])
    prompt_text = f"What can I do to {prompt.lower()}? "
    output = generator(prompt_text, max_length=max_length, num_return_sequences=1, no_repeat_ngram_size=2, early_stopping=True)
    output_text = output[0]['generated_text'][len(prompt_text):].strip()
    return prompt_text, output_text, solution

st.title('ICL-LM Interface')
option = st.selectbox('Select a problem:', list(prompts.keys()))

if st.button('Generate Prompt and Solution'):
    results = []
    for _ in range(3):
        prompt_text, prompt, solution = generate_prompt(option)
        results.append([prompt_text, prompt, solution])
        
    with open('results.txt', 'a') as f:
        for result in results:
            f.write(f"{result[0]}\t{result[1]}\t{result[2]}\n")
    
    df = pd.read_csv('results.txt', sep='\t', header=None, names=['Input', 'Prompt', 'Solution'])
    st.write(df)