app.py

import gradio as gr
import yfinance as yf
import pandas as pd
import numpy as np
import plotly.graph_objects as go
from plotly.subplots import make_subplots
from sklearn.preprocessing import MinMaxScaler
from sklearn.linear_model import LinearRegression
from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor
from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score
import warnings
warnings.filterwarnings("ignore")

# ── Theme colours (dark terminal-finance aesthetic) ──────────────────────────
BG        = "#0d1117"
SURFACE   = "#161b22"
BORDER    = "#30363d"
GREEN     = "#3fb950"
RED       = "#f85149"
BLUE      = "#58a6ff"
YELLOW    = "#d29922"
TEXT      = "#e6edf3"
MUTED     = "#8b949e"

POPULAR_TICKERS = [
    "AAPL", "MSFT", "GOOGL", "AMZN", "TSLA",
    "META", "NVDA", "JPM", "BRK-B", "V",
    "NFLX", "DIS", "BABA", "AMD", "INTC",
    "UBER", "SPOT", "PYPL", "SQ", "SNAP",
]

MODELS = {
    "Linear Regression":        LinearRegression(),
    "Random Forest":            RandomForestRegressor(n_estimators=100, random_state=42),
    "Gradient Boosting":        GradientBoostingRegressor(n_estimators=100, random_state=42),
}

PERIODS = {
    "6 Months":  "6mo",
    "1 Year":    "1y",
    "2 Years":   "2y",
    "5 Years":   "5y",
    "10 Years":  "10y",
}

INTERVALS = {
    "Daily":   "1d",
    "Weekly":  "1wk",
}

# ── Feature engineering ───────────────────────────────────────────────────────
def make_features(df: pd.DataFrame) -> pd.DataFrame:
    df = df.copy()
    df["MA7"]   = df["Close"].rolling(7).mean()
    df["MA21"]  = df["Close"].rolling(21).mean()
    df["MA50"]  = df["Close"].rolling(50).mean()
    df["EMA12"] = df["Close"].ewm(span=12, adjust=False).mean()
    df["EMA26"] = df["Close"].ewm(span=26, adjust=False).mean()
    df["MACD"]  = df["EMA12"] - df["EMA26"]
    df["Vol_MA"] = df["Volume"].rolling(7).mean()
    df["Return1"] = df["Close"].pct_change(1)
    df["Return5"] = df["Close"].pct_change(5)
    df["High_Low"]    = df["High"] - df["Low"]
    df["Close_Open"]  = df["Close"] - df["Open"]
    delta = df["Close"].diff()
    gain  = delta.clip(lower=0).rolling(14).mean()
    loss  = (-delta.clip(upper=0)).rolling(14).mean()
    rs    = gain / loss.replace(0, np.nan)
    df["RSI"] = 100 - (100 / (1 + rs))
    df["Target"] = df["Close"].shift(-1)
    return df.dropna()

# ── Core prediction logic ─────────────────────────────────────────────────────
def predict_stock(ticker, period_label, interval_label, model_name, future_days):
    ticker  = ticker.strip().upper()
    period  = PERIODS[period_label]
    interval = INTERVALS[interval_label]

    try:
        raw = yf.download(ticker, period=period, interval=interval, progress=False)
        if raw.empty:
            return None, None, f"❌  No data found for **{ticker}**. Check the ticker symbol."
        if isinstance(raw.columns, pd.MultiIndex):
            raw.columns = raw.columns.get_level_values(0)
        raw = raw[["Open","High","Low","Close","Volume"]].dropna()
    except Exception as e:
        return None, None, f"❌  Data fetch error: {e}"

    if len(raw) < 60:
        return None, None, f"❌  Not enough data ({len(raw)} rows). Try a longer period."

    df = make_features(raw)
    feature_cols = [
        "MA7","MA21","MA50","EMA12","EMA26","MACD",
        "Vol_MA","Return1","Return5","High_Low","Close_Open","RSI",
        "Open","High","Low","Volume",
    ]

    X = df[feature_cols].values
    y = df["Target"].values

    split = int(len(X) * 0.8)
    X_train, X_test = X[:split], X[split:]
    y_train, y_test = y[:split], y[split:]

    scaler = MinMaxScaler()
    X_train_s = scaler.fit_transform(X_train)
    X_test_s  = scaler.transform(X_test)

    model = MODELS[model_name]
    model.fit(X_train_s, y_train)
    y_pred = model.predict(X_test_s)

    mae  = mean_absolute_error(y_test, y_pred)
    rmse = np.sqrt(mean_squared_error(y_test, y_pred))
    r2   = r2_score(y_test, y_pred)
    acc  = max(0.0, r2) * 100

    # ── Future forecast ───────────────────────────────────────────────────────
    future_days = int(future_days)
    last_features = X[-1].reshape(1, -1)
    future_prices = []
    cur = last_features.copy()
    for _ in range(future_days):
        cur_s  = scaler.transform(cur)
        nxt    = model.predict(cur_s)[0]
        future_prices.append(float(nxt))
        cur[0, 0] = nxt          # crude: update Close proxy

    last_date  = df.index[-1]
    freq       = "B" if interval == "1d" else "W"
    fut_dates  = pd.date_range(last_date, periods=future_days + 1, freq=freq)[1:]

    # ── Candlestick + prediction chart ───────────────────────────────────────
    fig = make_subplots(
        rows=3, cols=1,
        shared_xaxes=True,
        row_heights=[0.55, 0.25, 0.20],
        vertical_spacing=0.04,
        subplot_titles=("Price & Prediction", "Volume", "RSI"),
    )

    test_dates = df.index[split:]

    # Candlestick (all history)
    fig.add_trace(go.Candlestick(
        x=raw.index, open=raw["Open"], high=raw["High"],
        low=raw["Low"],  close=raw["Close"],
        increasing_line_color=GREEN, decreasing_line_color=RED,
        name="Price", showlegend=False,
    ), row=1, col=1)

    # Moving averages
    for col, color, label in [("MA21", BLUE, "MA 21"), ("MA50", YELLOW, "MA 50")]:
        fig.add_trace(go.Scatter(
            x=df.index, y=df[col], line=dict(color=color, width=1.2),
            name=label, opacity=0.85,
        ), row=1, col=1)

    # Test-set predictions
    fig.add_trace(go.Scatter(
        x=test_dates, y=y_pred,
        line=dict(color="#a371f7", width=1.8, dash="dot"),
        name="Model (test)", opacity=0.9,
    ), row=1, col=1)

    # Future forecast
    fig.add_trace(go.Scatter(
        x=list(fut_dates), y=future_prices,
        line=dict(color="#f0883e", width=2),
        name=f"Forecast ({future_days}d)",
        mode="lines+markers",
        marker=dict(size=5),
    ), row=1, col=1)

    # Vertical "today" line
    fig.add_vline(x=str(last_date.date()), line_width=1,
                  line_dash="dash", line_color=MUTED, row=1, col=1)

    # Volume bars
    colors_v = [GREEN if c >= o else RED
                for c, o in zip(raw["Close"], raw["Open"])]
    fig.add_trace(go.Bar(
        x=raw.index, y=raw["Volume"],
        marker_color=colors_v, showlegend=False, name="Volume",
    ), row=2, col=1)

    # RSI
    fig.add_trace(go.Scatter(
        x=df.index, y=df["RSI"],
        line=dict(color=BLUE, width=1.5),
        name="RSI", showlegend=False,
    ), row=3, col=1)
    fig.add_hline(y=70, line_dash="dash", line_color=RED,   line_width=0.8, row=3, col=1)
    fig.add_hline(y=30, line_dash="dash", line_color=GREEN, line_width=0.8, row=3, col=1)

    fig.update_layout(
        paper_bgcolor=BG, plot_bgcolor=SURFACE,
        font=dict(family="'JetBrains Mono', monospace", color=TEXT, size=12),
        legend=dict(bgcolor=SURFACE, bordercolor=BORDER, borderwidth=1,
                    x=0.01, y=0.99, font=dict(size=11)),
        margin=dict(l=10, r=10, t=40, b=10),
        xaxis_rangeslider_visible=False,
        height=700,
    )
    for row in [1, 2, 3]:
        fig.update_xaxes(
            gridcolor=BORDER, showgrid=True,
            zeroline=False, row=row, col=1,
        )
        fig.update_yaxes(
            gridcolor=BORDER, showgrid=True,
            zeroline=False, row=row, col=1,
        )

    # ── Metrics card (markdown) ───────────────────────────────────────────────
    cur_price = float(raw["Close"].iloc[-1])
    fst_fcast = future_prices[0]  if future_prices else cur_price
    lst_fcast = future_prices[-1] if future_prices else cur_price
    delta_pct = (lst_fcast - cur_price) / cur_price * 100
    arrow     = "▲" if delta_pct >= 0 else "▼"
    clr_tag   = "🟢" if delta_pct >= 0 else "🔴"

    rsi_now = float(df["RSI"].iloc[-1])
    rsi_sig = ("Overbought ⚠️" if rsi_now > 70
               else "Oversold 💡"  if rsi_now < 30
               else "Neutral ✅")

    stats_md = f"""
## {ticker} — {model_name}

| Metric | Value |
|--------|-------|
| **Current Price** | `${cur_price:,.2f}` |
| **Next-Period Forecast** | `${fst_fcast:,.2f}` |
| **{future_days}-Day Forecast** | `${lst_fcast:,.2f}` |
| **Expected Change** | `{clr_tag} {arrow} {abs(delta_pct):.2f}%` |

---

### Model Performance (test set)
| | |
|---|---|
| MAE  | `${mae:.4f}` |
| RMSE | `${rmse:.4f}` |
| R²   | `{r2:.4f}` |
| Accuracy proxy | `{acc:.1f}%` |

---

### Technical Signals
| Indicator | Value | Signal |
|-----------|-------|--------|
| RSI (14) | `{rsi_now:.1f}` | {rsi_sig} |
| MACD | `{float(df['MACD'].iloc[-1]):.4f}` | {'Bullish 📈' if float(df['MACD'].iloc[-1]) > 0 else 'Bearish 📉'} |
| MA21 vs MA50 | — | {'Golden Cross ✨' if float(df['MA21'].iloc[-1]) > float(df['MA50'].iloc[-1]) else 'Death Cross 💀'} |

> ⚠️ **Disclaimer:** This tool is for educational purposes only. Not financial advice.
"""

    return fig, stats_md, ""

# ── UI layout ─────────────────────────────────────────────────────────────────
css = f"""
@import url('https://fonts.googleapis.com/css2?family=JetBrains+Mono:wght@300;400;600&family=Inter:wght@400;600&display=swap');

* {{ box-sizing: border-box; }}

body, .gradio-container {{
    background: {BG} !important;
    color: {TEXT} !important;
    font-family: 'Inter', sans-serif !important;
}}

/* Header */
.app-header {{
    background: {SURFACE};
    border-bottom: 1px solid {BORDER};
    padding: 28px 32px 20px;
    margin-bottom: 24px;
}}
.app-header h1 {{
    font-family: 'JetBrains Mono', monospace;
    font-size: 2rem;
    font-weight: 600;
    color: {TEXT};
    margin: 0 0 4px;
    letter-spacing: -0.5px;
}}
.app-header p {{
    color: {MUTED};
    font-size: 0.9rem;
    margin: 0;
}}
.accent {{ color: {GREEN}; }}

/* Cards */
.card {{
    background: {SURFACE} !important;
    border: 1px solid {BORDER} !important;
    border-radius: 8px !important;
    padding: 16px !important;
}}

/* Controls */
label {{
    color: {MUTED} !important;
    font-size: 0.78rem !important;
    font-weight: 600 !important;
    text-transform: uppercase !important;
    letter-spacing: 0.08em !important;
    margin-bottom: 4px !important;
}}
input, select, .svelte-1gfkn6j {{
    background: {BG} !important;
    border: 1px solid {BORDER} !important;
    color: {TEXT} !important;
    border-radius: 6px !important;
    font-family: 'JetBrains Mono', monospace !important;
}}
input:focus {{ border-color: {BLUE} !important; outline: none !important; }}

/* Run button */
.run-btn button {{
    background: {GREEN} !important;
    color: #0d1117 !important;
    font-weight: 700 !important;
    font-size: 0.95rem !important;
    border: none !important;
    border-radius: 6px !important;
    padding: 12px 0 !important;
    width: 100% !important;
    cursor: pointer !important;
    font-family: 'JetBrains Mono', monospace !important;
    letter-spacing: 0.05em !important;
    transition: opacity .15s;
}}
.run-btn button:hover {{ opacity: 0.85; }}

/* Metrics markdown */
.stats-box {{
    background: {BG} !important;
    border: 1px solid {BORDER} !important;
    border-radius: 8px !important;
    padding: 20px !important;
    font-family: 'JetBrains Mono', monospace !important;
    font-size: 0.82rem !important;
}}
.stats-box table {{ width: 100%; border-collapse: collapse; }}
.stats-box td, .stats-box th {{
    padding: 6px 10px;
    border-bottom: 1px solid {BORDER};
    text-align: left;
}}
.stats-box th {{ color: {MUTED}; font-weight: 600; }}
.stats-box code {{
    background: {SURFACE};
    padding: 2px 6px;
    border-radius: 4px;
    color: {BLUE};
}}

/* Error box */
.error-box textarea {{
    background: transparent !important;
    color: {RED} !important;
    border: none !important;
    font-family: 'JetBrains Mono', monospace !important;
    font-size: 0.85rem !important;
}}

/* Ticker pills */
.ticker-pills {{
    display: flex; flex-wrap: wrap; gap: 6px;
    margin-top: 8px;
}}
.ticker-pills button {{
    background: {SURFACE} !important;
    border: 1px solid {BORDER} !important;
    color: {TEXT} !important;
    border-radius: 4px !important;
    padding: 3px 10px !important;
    font-size: 0.75rem !important;
    font-family: 'JetBrains Mono', monospace !important;
    cursor: pointer !important;
    transition: border-color .15s;
}}
.ticker-pills button:hover {{ border-color: {GREEN} !important; color: {GREEN} !important; }}

/* Plotly panel */
.plot-panel {{ border: 1px solid {BORDER}; border-radius: 8px; overflow: hidden; }}

/* Slider */
input[type=range] {{ accent-color: {BLUE}; }}

/* Footer */
.footer {{
    text-align: center;
    color: {MUTED};
    font-size: 0.75rem;
    margin-top: 32px;
    padding: 16px;
    border-top: 1px solid {BORDER};
}}
"""

# ── Build Gradio app ──────────────────────────────────────────────────────────
with gr.Blocks(css=css, title="StockSense — ML Stock Predictor") as demo:

    # Header
    gr.HTML("""
    <div class="app-header">
      <h1>📈 Stock<span class="accent">Sense</span></h1>
      <p>Machine-learning powered stock analysis &amp; price forecasting · Educational use only</p>
    </div>
    """)

    with gr.Row():
        # ── Left panel: controls ─────────────────────────────────────────────
        with gr.Column(scale=1, elem_classes="card"):

            gr.HTML("<div style='font-size:0.78rem;color:#8b949e;font-weight:600;text-transform:uppercase;letter-spacing:.08em;margin-bottom:8px'>Quick Pick</div>")
            ticker_pills_html = "".join(
                f'<button onclick="document.querySelector(\'#ticker_input input\').value=\'{t}\';'
                f'document.querySelector(\'#ticker_input input\').dispatchEvent(new Event(\'input\'))">{t}</button>'
                for t in POPULAR_TICKERS
            )
            gr.HTML(f'<div class="ticker-pills">{ticker_pills_html}</div>')

            ticker_input = gr.Textbox(
                label="Ticker Symbol",
                placeholder="e.g. AAPL, TSLA, MSFT …",
                value="AAPL",
                elem_id="ticker_input",
            )

            with gr.Row():
                period_input = gr.Dropdown(
                    label="History Period",
                    choices=list(PERIODS.keys()),
                    value="2 Years",
                )
                interval_input = gr.Dropdown(
                    label="Interval",
                    choices=list(INTERVALS.keys()),
                    value="Daily",
                )

            model_input = gr.Dropdown(
                label="ML Model",
                choices=list(MODELS.keys()),
                value="Random Forest",
            )

            future_input = gr.Slider(
                label="Forecast Horizon (days)",
                minimum=1, maximum=90, step=1, value=30,
            )

            run_btn = gr.Button("▶  Run Prediction", elem_classes="run-btn")
            error_box = gr.Textbox(
                visible=True, interactive=False,
                show_label=False, elem_classes="error-box",
            )

            # Stats card below button
            stats_md = gr.Markdown(
                value="*Run a prediction to see metrics here.*",
                elem_classes="stats-box",
            )

        # ── Right panel: chart ───────────────────────────────────────────────
        with gr.Column(scale=3):
            chart = gr.Plot(elem_classes="plot-panel", label="")

    # Footer
    gr.HTML(f"""
    <div class="footer">
        StockSense · Built with Gradio &amp; scikit-learn ·
        Data via Yahoo Finance ·
        <span style="color:{RED}">Not financial advice</span>
    </div>
    """)

    # ── Wire up ───────────────────────────────────────────────────────────────
    def run(ticker, period, interval, model, future):
        fig, stats, err = predict_stock(ticker, period, interval, model, future)
        return (
            fig if fig else go.Figure(),
            stats if stats else "",
            err,
        )

    run_btn.click(
        fn=run,
        inputs=[ticker_input, period_input, interval_input, model_input, future_input],
        outputs=[chart, stats_md, error_box],
    )

    # Auto-run on load
    demo.load(
        fn=run,
        inputs=[ticker_input, period_input, interval_input, model_input, future_input],
        outputs=[chart, stats_md, error_box],
    )

if __name__ == "__main__":
    demo.launch()