interface.py

import streamlit as st
import numpy as np
import plotly.graph_objects as go
from app import SMCPredictionEngine

# ── Page Config ───────────────────────────────────────────────────────────────
st.set_page_config(page_title="AI-Powered SMC Suite", layout="wide")


# ── Lazy model loader (cached across reruns, NOT at module import time) ───────
@st.cache_resource(show_spinner="Loading AI engine — this may take ~30 s on first run…")
def load_ai_engine():
    return SMCPredictionEngine()


# ── Sidebar ───────────────────────────────────────────────────────────────────
st.sidebar.header("🎯 System Parameters")
trading_pair = st.sidebar.selectbox(
    "Market Asset Profile", ["BTC/USDT", "ETH/USDT", "SOL/USDT"]
)
timeframe = st.sidebar.selectbox(
    "Multi-Timeframe Horizon", ["5-Minute", "10-Minute", "1-Hour"]
)
prediction_horizon = st.sidebar.slider(
    "AI Forward Forecast Horizon (Candles)", 4, 24, 12
)

st.sidebar.markdown("---")
st.sidebar.header("🕹️ Generation Control")
market_condition = st.sidebar.radio(
    "Simulate Market Bias",
    [
        "Bullish Inflow (Order Block Expansion)",
        "Bearish Liquidity Sweep Collapse",
        "Ranging / Volatile Consolidation",
    ],
)

# ── Generate synthetic historical price data ──────────────────────────────────
np.random.seed(101)
base_price = (
    65000 if trading_pair == "BTC/USDT"
    else (3500 if trading_pair == "ETH/USDT" else 150)
)
steps = 100

if market_condition == "Bullish Inflow (Order Block Expansion)":
    noise = np.random.normal(5, 20, steps)
    trend = np.linspace(0, 400, steps)
elif market_condition == "Bearish Liquidity Sweep Collapse":
    noise = np.random.normal(-5, 20, steps)
    trend = np.linspace(0, -400, steps)
else:
    noise = np.random.normal(0, 35, steps)
    trend = np.zeros(steps)

historical_prices = base_price + trend + np.cumsum(noise)

# ── Header ────────────────────────────────────────────────────────────────────
st.title("🎛️ SMC Trading Suite — Version 4.0 AI Layer")
st.caption("Connected to Hugging Face Engine: Amazon Chronos-Bolt-Base")
st.markdown("---")

# ── Load model & run inference ────────────────────────────────────────────────
ai_brain = load_ai_engine()  # triggers spinner only on first load

with st.spinner("Executing time-series transformer inference…"):
    results = ai_brain.calculate_forecast(historical_prices, prediction_horizon)

# ── Metrics row ───────────────────────────────────────────────────────────────
col1, col2, col3 = st.columns(3)

with col1:
    st.metric("Latest Ticking Price", f"${historical_prices[-1]:,.2f}")

with col2:
    pct = results["projected_change_pct"]
    st.metric(
        label="AI Projected Momentum Drift",
        value=f"{pct:.2f}%",
        delta=f"{pct:.2f}% Directional Path",
    )

with col3:
    score = (
        "HIGH PROBABILITY ENTRY"
        if abs(results["projected_change_pct"]) > 0.5
        else "REJECTED / LOW VOLUMETRIC MOMENTUM"
    )
    st.info(f"Signal Optimizer Output: **{score}**")

# ── Plotly chart ──────────────────────────────────────────────────────────────
st.subheader("📊 Architectural Vector Forecast Visualization")

history_idx = list(range(len(historical_prices)))
future_idx  = list(range(len(historical_prices), len(historical_prices) + prediction_horizon))

fig = go.Figure()

fig.add_trace(go.Scatter(
    x=history_idx, y=historical_prices,
    mode="lines", name="Cleaned Candle Data Stream",
    line=dict(color="#1f77b4", width=2),
))

fig.add_trace(go.Scatter(
    x=future_idx, y=results["median"],
    mode="lines+markers", name="AI Median Structural Path",
    line=dict(color="#e377c2", width=2, dash="dash"),
))

# Upper boundary (invisible line — acts as ceiling for fill)
fig.add_trace(go.Scatter(
    x=future_idx, y=results["high"],
    mode="lines", line=dict(width=0),
    showlegend=False, name="Upper Volatility Boundary",
))

# Lower boundary filled back to upper → confidence band
fig.add_trace(go.Scatter(
    x=future_idx, y=results["low"],
    mode="lines", line=dict(width=0),
    fill="tonexty", fillcolor="rgba(227, 119, 194, 0.2)",
    showlegend=True, name="80% Structural Risk Envelope",
))

fig.update_layout(
    template="plotly_dark",
    xaxis_title="Historical & Predicted Timeline Sequence",
    yaxis_title="Asset Value Index ($)",
    margin=dict(l=20, r=20, t=20, b=20),
    height=550,
    legend=dict(yanchor="top", y=0.99, xanchor="left", x=0.01),
)

st.plotly_chart(fig, use_container_width=True)