app.py

# app.py
import os
import streamlit as st
from PIL import Image
from groq import Groq
import base64
import io

# Set GROQ API Key (put your key directly for Colab or use environment variables)
os.environ["GROQ_API_KEY"] = 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

# Initialize GROQ client
client = Groq(api_key=os.environ.get("GROQ_API_KEY"))

st.set_page_config(page_title="AI Trade Predictor", layout="wide")
st.markdown("""
    <style>
    .main {
        background-color: #0d1117;
        color: white;
    }
    .stButton>button {
        background-color: #1f6feb;
        color: white;
        font-weight: bold;
    }
    .stFileUploader label {
        color: #58a6ff;
    }
    </style>
""", unsafe_allow_html=True)

st.title("\U0001F4B0 AI Trade Predictor")
st.markdown("Upload a candlestick chart image and get a trading signal analysis using AI")

# Upload chart image
uploaded_file = st.file_uploader("Upload Candlestick Chart Image", type=["jpg", "png", "jpeg"])

if uploaded_file is not None:
    image = Image.open(uploaded_file)
    st.image(image, caption="Uploaded Chart", use_column_width=True)

    buffer = io.BytesIO()
    image.save(buffer, format="PNG")
    img_str = base64.b64encode(buffer.getvalue()).decode()

    if st.button("Analyze Chart \U0001F52C"):
        with st.spinner("Analyzing chart and generating predictions..."):
            prompt = f"""
            You are an expert trading analyst AI.
            Analyze the attached candlestick chart image (base64 below).
            Apply technical strategies like RSI, MACD, moving averages, support/resistance, candlestick patterns.
            Then tell:
            1. Whether to BUY or SELL.
            2. The confidence level in %.
            3. The best timeframe for this prediction.
            4. The risk level and how it might go wrong.
            5. Why this prediction was made.
            Base64 image: {img_str}
            """
            
            chat_completion = client.chat.completions.create(
                messages=[{"role": "user", "content": prompt}],
                model="llama-3.3-70b-versatile"
            )

            result = chat_completion.choices[0].message.content
            st.markdown("### \U0001F4C8 Prediction Result")
            st.markdown(result)

else:
    st.info("Please upload a candlestick chart image to begin analysis.")