# app.py
from flask import Flask, render_template, jsonify, request
import pandas as pd
from longshort_core import load_mt5_csv, make_features, FEATURES, auto_thresholds, load_or_train_model, build_payload_for_chart

CSV_PATH = "data/XAUUSD_M15.csv"
SUMMARY_PATH = "data/wf_summary_longshort.csv"

app = Flask(__name__)

@app.route("/")
def index():
    return render_template("index.html")  # nến + signal

@app.route("/equity")
def equity():
    return render_template("equity.html")

@app.route("/api/candles")
def api_candles():
    n = int(request.args.get("n", 500))
    df = load_mt5_csv(CSV_PATH)
    feat = make_features(df)
    # y để train nhanh nếu chưa có model
    y = (feat['Close'].shift(-1)/feat['Close'] - 1.0 > 0).astype(int)
    data = feat.join(y.rename('y')).dropna()
    X = data[FEATURES]; y = data['y']
    thrL, thrS, dz = auto_thresholds(SUMMARY_PATH)
    model = load_or_train_model(X, y, model_path="models/live_lgb.txt")
    prob = pd.Series(model.predict(X), index=X.index)
    payload = build_payload_for_chart(df.loc[X.index], prob, thrL, thrS, dz, n_last=n)
    return jsonify(payload)

@app.route("/api/latest")
def api_latest():
    from datetime import datetime
    df = load_mt5_csv(CSV_PATH)
    feat = make_features(df)
    y = (feat['Close'].shift(-1)/feat['Close'] - 1.0 > 0).astype(int)
    data = feat.join(y.rename('y')).dropna()
    X = data[FEATURES]; y = data['y']
    thrL, thrS, dz = auto_thresholds(SUMMARY_PATH)
    model = load_or_train_model(X, y, model_path="models/live_lgb.txt")
    prob = model.predict(X.iloc[[-1]])[0]
    ts = X.index[-1].strftime("%Y-%m-%d %H:%M:%S")
    sig = 0
    if abs(prob-0.5) < dz: sig = 0
    elif prob > thrL: sig = 1
    elif prob < thrS: sig = -1
    return jsonify(dict(time=ts, close=float(data['Close'].iloc[-1]),
                        prob_up=float(prob), thrL=thrL, thrS=thrS, dz=dz, signal=int(sig)))

@app.route("/api/equity")
def api_equity():
    try:
        eq = pd.read_csv("data/wf_equity_curve_longshort.csv")
    except:
        eq = pd.read_csv("data/wf_equity_curve.csv")  # fallback
    return jsonify(dict(time=eq.iloc[:,0].tolist(), equity=eq.iloc[:,1].round(4).tolist()))

if __name__ == "__main__":
    app.run(debug=True)