Update app.py

app.py

@@ -1,69 +1,75 @@
-# app.py
 import streamlit as st
 from PIL import Image
-import base64
 import os
 from groq import Groq
-import io
+from dotenv import load_dotenv
 
-# --- PAGE CONFIG ---
-st.set_page_config(
-    page_title="AI Trade Predictor",
-    layout="wide",
-    initial_sidebar_state="expanded"
-)
+load_dotenv()
 
-st.title("📈 AI Trade Predictor")
-st.markdown("Upload a candlestick chart and let AI analyze trade signals with risk and timeframe guidance.")
+st.set_page_config(page_title="AI Trade Predictor", layout="wide")
+st.markdown("""
+    <style>
+        .main {
+            background-color: #f5f7fa;
+        }
+        h1 {
+            color: #3b3b3b;
+        }
+        .stButton > button {
+            color: white;
+            background-color: #4CAF50;
+            font-size: 16px;
+            border-radius: 10px;
+            padding: 10px 24px;
+        }
+    </style>
+""", unsafe_allow_html=True)
 
-# --- UPLOAD SECTION ---
-uploaded_file = st.file_uploader("Upload Candlestick Chart (PNG or JPG)", type=["png", "jpg", "jpeg"])
+st.title("📈 AI Trade Predictor")
 
-# --- GROQ SETUP ---
-groq_api_key = st.secrets["GROQ_API_KEY"] if "GROQ_API_KEY" in st.secrets else os.environ.get("GROQ_API_KEY")
-client = Groq(api_key=groq_api_key)
+uploaded_file = st.file_uploader("Upload a candlestick chart image", type=["png", "jpg", "jpeg"])
 
-# --- FUNCTION: Convert image to base64 string ---
-def image_to_base64_str(image_file):
-    img_bytes = image_file.read()
-    base64_str = base64.b64encode(img_bytes).decode("utf-8")
-    return base64_str
+if uploaded_file:
+    image = Image.open(uploaded_file)
+    st.image(image, caption='Uploaded Chart', use_column_width=True)
+    st.write("Analyzing chart using AI model...")
 
-# --- SHORTER PROMPT TEMPLATE ---
-prompt_template = """
-You're an AI financial assistant. Analyze the candlestick chart image and give:
-1. Signal (Buy/Sell/Hold)
-2. Confidence percentage
-3. Reasoning (simple language)
-4. Suggest best timeframes like 30min, 1hr, 4hr
-5. Explain key risks in a beginner-friendly way
+    # Convert image to base64 string
+    import base64
+    import io
+    buffered = io.BytesIO()
+    image.save(buffered, format="PNG")
+    img_str = base64.b64encode(buffered.getvalue()).decode("utf-8")
 
-Only reply in concise format.
-"""
+    # Connect to Groq and send the image analysis prompt
+    client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
 
-# --- MAIN PREDICTION LOGIC ---
-if uploaded_file is not None:
-    st.image(uploaded_file, caption="Uploaded Chart", use_column_width=True)
-    with st.spinner("Analyzing chart and generating prediction..."):
+    user_prompt = f"""
+    Analyze the following candlestick chart image (base64-encoded PNG) and provide a trading decision:
+    - Tell whether the action should be BUY, SELL, or HOLD.
+    - Give confidence level in percentage.
+    - Suggest timeframes (e.g., 30 min, 1 hour, 4 hour, 1 day) and what signal applies to each.
+    - List any risks or reasons the prediction may fail.
+    - Use clear language that a beginner can understand.
+    - Give a short summary at the end.
 
-        image_base64 = image_to_base64_str(uploaded_file)
+    Image (base64 PNG): {img_str}
+    """
 
-        # Compose final prompt with image reference (simulated for now)
-        full_prompt = prompt_template + "\n[This is a simulated chart image base64: {}]".format(image_base64[:100])
+    try:
+        chat_completion = client.chat.completions.create(
+            messages=[
+                {"role": "user", "content": user_prompt}
+            ],
+            model="meta-llama/llama-guard-4-12b"
+        )
+        response = chat_completion.choices[0].message.content
+        st.success("Prediction Complete")
+        st.markdown(response)
 
-        try:
-            chat_completion = client.chat.completions.create(
-                messages=[
-                    {"role": "user", "content": full_prompt}
-                ],
-                model="llama-3.3-70b-versatile"  # Smaller model to reduce token use
-            )
-            result = chat_completion.choices[0].message.content
-            st.success("Prediction Ready")
-            st.markdown(result)
+    except Exception as e:
+        st.error(f"Something went wrong: {e}")
 
-        except Exception as e:
-            st.error(f"Something went wrong: {e}")
 
 # --- FOOTER ---
 st.markdown("---")