import os

# =========================
# Path Config
# =========================
BASE_DIR = os.path.dirname(os.path.abspath(__file__))

INPUT_DIR = os.path.join(BASE_DIR, "input")
OUTPUT_DIR = os.path.join(BASE_DIR, "output_minute")
VERSION = "_30m_from_1s_ticks_smoothed"

INPUT_CSV = os.path.join(INPUT_DIR, "gemini_24h_analysis.csv")

# =========================
# Data Parameters
# =========================
SAMPLE_EVERY_N_EVENTS = 1
PREDICTION_HORIZON = 180
TOP_K_LEVELS = 5
WINDOW_MS = 1000

# =========================
# Model Parameters
# =========================
N_SPLITS = 5
RANDOM_STATE = 42
VALIDATION_FRACTION = 0.2
EARLY_STOPPING_ROUNDS = 50

# =========================
# Label / Strategy Parameters
# =========================
TRANSACTION_FEE_RATE = 0.001
MILD_RETURN_THRESHOLD = 0.00001
COST_AWARE_LABEL_THRESHOLD_MULTIPLIER = 1.0
SIGNAL_SCORE_WINDOW_MINUTES = 60
SIGNAL_SCORE_BUY_QUANTILE = 0.9
SIGNAL_SCORE_CLOSE_QUANTILE = 0.1
SIGNAL_SMOOTHING_SECONDS = max(60, PREDICTION_HORIZON // 3)
MIN_HOLD_SECONDS = max(60, PREDICTION_HORIZON // 3)

LGB_PARAMS = {
    "objective": "multiclass",
    "num_class": 3,
    "n_estimators": 1500,
    "learning_rate": 0.05,
    "num_leaves": 15,
    "max_depth": 4,
    "min_child_samples": 60,
    "subsample": 0.8,
    "subsample_freq": 1,
    "colsample_bytree": 0.8,
    "reg_alpha": 0.01,
    "reg_lambda": 0.001,
    "class_weight": "balanced",
    "random_state": RANDOM_STATE,
    "n_jobs": -1,
}

# =========================
# Output Paths
# =========================
FEATURE_CSV = os.path.join(OUTPUT_DIR, "1s_tick_feature_2.0.csv")
DATASET_CSV = os.path.join(OUTPUT_DIR, f"dataset_with_targets{VERSION}.csv")
FOLD_METRICS_CSV = os.path.join(OUTPUT_DIR, f"fold_metrics{VERSION}.csv")
OVERALL_METRICS_CSV = os.path.join(OUTPUT_DIR, f"overall_metrics{VERSION}.csv")
FEATURE_IMPORTANCE_CSV = os.path.join(OUTPUT_DIR, f"feature_importances{VERSION}.csv")
CONFUSION_MATRIX_CSV = os.path.join(OUTPUT_DIR, f"confusion_matrix{VERSION}.csv")
OOS_PREDICTIONS_CSV = os.path.join(OUTPUT_DIR, f"oos_predictions{VERSION}.csv")
OOS_TRADES_CSV = os.path.join(OUTPUT_DIR, f"oos_trades{VERSION}.csv")
OOS_TRADE_EVENTS_CSV = os.path.join(OUTPUT_DIR, f"oos_trade_events{VERSION}.csv")
PARAMETER_TUNING_CSV = os.path.join(OUTPUT_DIR, f"parameter_tuning{VERSION}.csv")

PRED_VS_TRUE_PNG = os.path.join(OUTPUT_DIR, f"pred_vs_true{VERSION}.png")
SCATTER_PNG = os.path.join(OUTPUT_DIR, f"scatter{VERSION}.png")
CUMRET_PNG = os.path.join(OUTPUT_DIR, f"cumret{VERSION}.png")
FEATURE_IMPORTANCE_PNG = os.path.join(OUTPUT_DIR, f"feature_importance{VERSION}.png")
SIGNAL_TS_PNG = os.path.join(OUTPUT_DIR, f"signal_timeseries_with_thresholds{VERSION}.png")
POSITION_MIDPRICE_PNG = os.path.join(OUTPUT_DIR, f"position_midprice{VERSION}.png")