import os
import pandas as pd
from kaggle.api.kaggle_api_extended import KaggleApi
import streamlit as st

# App title
st.title("Automated Investment Advisory App")

# Retrieve Kaggle credentials from Hugging Face secrets
api_key = os.getenv("Kaggle_API_Key")
username = os.getenv("Kaggle_Username")  # Make sure you add this to Hugging Face secrets

# Function to set up Kaggle API configuration
def setup_kaggle_api():
    # Check if both API key and username are provided
    if api_key and username:
        # Create the .kaggle directory if it doesn't exist
        kaggle_dir = os.path.expanduser("~/.kaggle")
        os.makedirs(kaggle_dir, exist_ok=True)
        
        # Write kaggle.json file with credentials
        kaggle_json_path = os.path.join(kaggle_dir, "kaggle.json")
        with open(kaggle_json_path, "w") as f:
            f.write(f'{{"username":"{username}", "key":"{api_key}"}}')
        
        # Set correct permissions for the kaggle.json file
        os.chmod(kaggle_json_path, 0o600)

        # Authenticate Kaggle API
        api = KaggleApi()
        api.authenticate()
        return api
    else:
        st.error("Kaggle API credentials not found. Please ensure 'Kaggle_API_Key' and 'Kaggle_Username' are set in Hugging Face secrets.")
        return None

# Set up and authenticate Kaggle API
api = setup_kaggle_api()

# Download and load dataset if API is configured
if api:
    dataset_name = "borismarjanovic/price-volume-data-for-all-us-stocks-etfs"
    try:
        # Download the dataset to the current directory (where app.py exists)
        api.dataset_download_files(dataset_name, path=".", unzip=True)
        
        # Load dataset (adjust the filename if needed)
        df = pd.read_csv("all_stocks_5yr.csv")

        # App UI for user inputs
        st.sidebar.header("Investment Advisory Options")
        
        # Input for investment goal and risk tolerance
        investment_goal = st.sidebar.number_input("Enter your desired annual return (%)", min_value=0.0, step=0.1)
        risk_tolerance = st.sidebar.selectbox("Risk Tolerance", ["Low", "Medium", "High"])
        
        # Filtering dataset based on user preferences
        # Assuming dataset has columns 'Name', 'Close', 'Volume', and others that reflect company details
        # Here we use a simplified filtering approach based on risk tolerance
        if risk_tolerance == "Low":
            suggested_companies = df[df['Volume'] < df['Volume'].quantile(0.33)]  # Example logic for lower volatility
        elif risk_tolerance == "Medium":
            suggested_companies = df[(df['Volume'] >= df['Volume'].quantile(0.33)) & (df['Volume'] < df['Volume'].quantile(0.66))]
        else:
            suggested_companies = df[df['Volume'] >= df['Volume'].quantile(0.66)]  # Higher volume for high risk
        
        # Display suggested companies
        st.write(f"Suggested companies for {risk_tolerance} risk tolerance:")
        st.dataframe(suggested_companies[['Name', 'Close', 'Volume']].head(10))

        # Show insights for a selected company
        selected_company = st.selectbox("Select a company for more details", suggested_companies['Name'].unique())
        company_data = suggested_companies[suggested_companies['Name'] == selected_company]
        
        st.write(f"Details for {selected_company}:")
        st.dataframe(company_data)

    except Exception as e:
        st.error(f"An error occurred: {e}")
else:
    st.warning("Unable to load data due to missing Kaggle API credentials.")