"""
config.py -- Project-wide configuration for hedonic pricing analysis.

Centralises every tunable constant, file path, spatial reference,
modelling parameter, and visual-style setting so that downstream
modules never hard-code values.
"""

from __future__ import annotations

import os
from dataclasses import dataclass, field
from pathlib import Path
from typing import Any, Dict, List, Tuple

# ---------------------------------------------------------------------------
# 1. Directory layout
# ---------------------------------------------------------------------------

PROJECT_ROOT: Path = Path(__file__).resolve().parent.parent

# Data directories
DATA_DIR: Path = PROJECT_ROOT / "data"
DATA_RAW: Path = DATA_DIR / "raw"
DATA_INTERIM: Path = DATA_DIR / "interim"
DATA_PROCESSED: Path = DATA_DIR / "processed"
DATA_DUCKDB: Path = DATA_DIR / "duckdb"

# Output directories
OUTPUTS_DIR: Path = PROJECT_ROOT / "results"
OUTPUTS_TABLES: Path = OUTPUTS_DIR / "tables"
OUTPUTS_FIGURES: Path = OUTPUTS_DIR / "figures"
OUTPUTS_MODELS: Path = OUTPUTS_DIR / "models"

# Paper / LaTeX directory
PAPER_DIR: Path = PROJECT_ROOT / "paper"

# Ensure all directories exist at import time (idempotent)
for _d in (
    DATA_RAW, DATA_INTERIM, DATA_PROCESSED, DATA_DUCKDB,
    OUTPUTS_TABLES, OUTPUTS_FIGURES, OUTPUTS_MODELS, PAPER_DIR,
):
    _d.mkdir(parents=True, exist_ok=True)

# ---------------------------------------------------------------------------
# 2. Raw data file paths
# ---------------------------------------------------------------------------

# Primary cleaned sample (always present)
CLEANED_SAMPLE_CSV: Path = PROJECT_ROOT / "data" / "raw" / "cleaned_dataset_sample.csv"

# Inside Airbnb full extracts (may or may not exist)
LISTINGS_CSV_GZ: Path = DATA_RAW / "listings.csv.gz"
CALENDAR_CSV_GZ: Path = DATA_RAW / "calendar.csv.gz"
REVIEWS_CSV_GZ: Path = DATA_RAW / "reviews.csv.gz"
NEIGHBOURHOODS_CSV: Path = DATA_RAW / "neighbourhoods.csv"
NEIGHBOURHOODS_GEOJSON: Path = DATA_RAW / "neighbourhoods.geojson"

# DuckDB database
DUCKDB_PATH: Path = DATA_DUCKDB / "airbnb.duckdb"

# ---------------------------------------------------------------------------
# 3. Spatial / geographic constants
# ---------------------------------------------------------------------------

@dataclass(frozen=True)
class CityCenter:
    """WGS-84 centroid of a Quebec city."""
    name: str
    lat: float
    lon: float


CITY_CENTERS: Dict[str, CityCenter] = {
    "Montréal":        CityCenter("Montréal",        45.5017, -73.5673),
    "Montreal":        CityCenter("Montréal",        45.5017, -73.5673),
    "Québec":          CityCenter("Québec",           46.8139, -71.2080),
    "Quebec":          CityCenter("Québec",           46.8139, -71.2080),
    "Gatineau":        CityCenter("Gatineau",         45.4765, -75.7013),
    "Sherbrooke":      CityCenter("Sherbrooke",       45.4042, -71.8929),
    "Laval":           CityCenter("Laval",            45.6066, -73.7124),
    "Trois-Rivières":  CityCenter("Trois-Rivières",   46.3432, -72.5432),
    "Trois-Rivieres":  CityCenter("Trois-Rivières",   46.3432, -72.5432),
    "Saguenay":        CityCenter("Saguenay",         48.4279, -71.0548),
    "Lévis":           CityCenter("Lévis",            46.8032, -71.1780),
    "Levis":           CityCenter("Lévis",            46.8032, -71.1780),
    "Terrebonne":      CityCenter("Terrebonne",       45.7005, -73.6474),
    "Longueuil":       CityCenter("Longueuil",        45.5312, -73.5185),
    "Rimouski":        CityCenter("Rimouski",         48.4489, -68.5234),
    "Saint-Hyacinthe": CityCenter("Saint-Hyacinthe",  45.6307, -72.9570),
    "Drummondville":   CityCenter("Drummondville",    45.8838, -72.4843),
    "Granby":          CityCenter("Granby",            45.4001, -72.7329),
    "Mont-Tremblant":  CityCenter("Mont-Tremblant",   46.2088, -74.5849),
    "Magog":           CityCenter("Magog",             45.2668, -72.1419),
    "Charlevoix":      CityCenter("Charlevoix",       47.4389, -70.5025),
}

# Quebec bounding box (generous, covers all of southern Quebec + some northern)
QUEBEC_LAT_RANGE: Tuple[float, float] = (44.5, 55.5)
QUEBEC_LON_RANGE: Tuple[float, float] = (-80.0, -56.0)

# Spatial buffer distances in metres
SPATIAL_BUFFERS_M: List[int] = [250, 500, 1_000, 2_000]

# Earth radius for haversine (km)
EARTH_RADIUS_KM: float = 6_371.0

# ---------------------------------------------------------------------------
# 4. Price & winsorization
# ---------------------------------------------------------------------------

# Winsorization percentile pairs (lower, upper)
WINSORIZE_1_99: Tuple[float, float] = (0.01, 0.99)
WINSORIZE_2_98: Tuple[float, float] = (0.02, 0.98)
WINSORIZE_5_95: Tuple[float, float] = (0.05, 0.95)

# Default winsorization used in modelling
DEFAULT_WINSORIZE: Tuple[float, float] = WINSORIZE_1_99

# Minimum valid nightly price in CAD
PRICE_FLOOR: float = 10.0

# Maximum plausible nightly price in CAD
PRICE_CEILING: float = 25_000.0

# ---------------------------------------------------------------------------
# 5. Quantile regression thresholds
# ---------------------------------------------------------------------------

QUANTILE_THRESHOLDS: List[float] = [0.10, 0.25, 0.50, 0.75, 0.90]

# ---------------------------------------------------------------------------
# 6. Feature groups
# ---------------------------------------------------------------------------

FEATURE_GROUPS: Dict[str, List[str]] = {
    "structural": [
        "bedrooms", "bathrooms", "guests_count", "amenities_count",
    ],
    "location": [
        "lat", "long", "dist_city_center_km",
    ],
    "quality": [
        "rating", "quality_score", "num_reviews", "num_images",
    ],
    "host": [
        "is_superhost", "host_response_rate", "host_acceptance_rate",
        "host_listings_count",
    ],
    "property": [
        "property_type", "room_type",
    ],
    "amenity_flags": [
        "pets_allowed", "is_guest_favorite",
    ],
}

# Flat list of all feature names (order-preserving)
ALL_FEATURES: List[str] = [
    feat for group in FEATURE_GROUPS.values() for feat in group
]

# ---------------------------------------------------------------------------
# 7. Property / room type mappings
# ---------------------------------------------------------------------------

PROPERTY_TYPE_MAP: Dict[str, str] = {
    "Rental unit":   "Rental unit",
    "House":         "House",
    "Cabin/Chalet":  "Cabin/Chalet",
    "Condo":         "Condo",
    "Apartment":     "Apartment",
    "Other":         "Other",
    # Inside Airbnb variants
    "Entire rental unit":       "Rental unit",
    "Private room in rental unit": "Rental unit",
    "Entire home":              "House",
    "Entire house":             "House",
    "Entire cabin":             "Cabin/Chalet",
    "Entire chalet":            "Cabin/Chalet",
    "Entire condo":             "Condo",
    "Entire condominium (condo)": "Condo",
    "Entire loft":              "Apartment",
    "Entire guest suite":       "Apartment",
    "Entire guesthouse":        "House",
    "Entire townhouse":         "House",
    "Entire cottage":           "Cabin/Chalet",
    "Entire bungalow":          "House",
    "Entire villa":             "House",
    "Tiny home":                "Other",
    "Treehouse":                "Other",
    "Boat":                     "Other",
    "Camper/RV":                "Other",
    "Farm stay":                "Other",
    "Yurt":                     "Other",
    "Tent":                     "Other",
    "Castle":                   "Other",
    "Private room":             "Other",
    "Shared room":              "Other",
}

ROOM_TYPE_MAP: Dict[str, str] = {
    "Entire home/apt":   "Entire home/apt",
    "Private room":      "Private room",
    "Shared room":       "Shared room",
    "Hotel room":        "Hotel room",
    # Normalised aliases
    "entire home":       "Entire home/apt",
    "entire place":      "Entire home/apt",
    "private":           "Private room",
    "shared":            "Shared room",
    "hotel":             "Hotel room",
}

# ---------------------------------------------------------------------------
# 8. Model catalogue
# ---------------------------------------------------------------------------

MODEL_NAMES: List[str] = [
    "ols_linear",
    "ols_log",
    "ols_semilog",
    "robust_huber",
    "robust_bisquare",
    "quantile_10",
    "quantile_25",
    "quantile_50",
    "quantile_75",
    "quantile_90",
    "spatial_lag",
    "spatial_error",
    "gwr",
    "rf_regressor",
    "gradient_boosting",
    "xgboost",
    "lasso",
    "ridge",
    "elastic_net",
]

# ---------------------------------------------------------------------------
# 9. Reproducibility
# ---------------------------------------------------------------------------

RANDOM_SEED: int = 42
TEST_SIZE: float = 0.20
CV_FOLDS: int = 5

# ---------------------------------------------------------------------------
# 10. Visualisation palettes & figure settings
# ---------------------------------------------------------------------------

# Colour palettes
PALETTE_SEQUENTIAL: str = "YlOrRd"
PALETTE_DIVERGING: str = "RdBu_r"
PALETTE_CATEGORICAL: List[str] = [
    "#1f77b4",  # muted blue
    "#ff7f0e",  # safety orange
    "#2ca02c",  # cooked-asparagus green
    "#d62728",  # brick red
    "#9467bd",  # muted purple
    "#8c564b",  # chestnut brown
    "#e377c2",  # raspberry-yoghurt pink
    "#7f7f7f",  # middle grey
    "#bcbd22",  # curry yellow-green
    "#17becf",  # blue-teal
]

PALETTE_PROPERTY_TYPE: Dict[str, str] = {
    "Rental unit":   "#1f77b4",
    "House":         "#2ca02c",
    "Cabin/Chalet":  "#8c564b",
    "Condo":         "#ff7f0e",
    "Apartment":     "#9467bd",
    "Other":         "#7f7f7f",
}

PALETTE_PRICE_CATEGORY: Dict[str, str] = {
    "Budget (<$100)":          "#2ca02c",
    "Moderate ($100-200)":     "#1f77b4",
    "Premium ($200-400)":      "#ff7f0e",
    "Luxury ($400-1000)":      "#d62728",
    "Ultra-Luxury (>$1000)":   "#9467bd",
    "Unknown":                 "#7f7f7f",
}

# Figure defaults
FIGURE_DPI: int = 300
FIGURE_FORMAT: str = "pdf"
FIGURE_SIZE_SINGLE: Tuple[float, float] = (6.5, 4.5)
FIGURE_SIZE_DOUBLE: Tuple[float, float] = (13.0, 4.5)
FIGURE_SIZE_FULL: Tuple[float, float] = (13.0, 9.0)
FIGURE_FONT_FAMILY: str = "serif"
FIGURE_FONT_SIZE: int = 11

# ---------------------------------------------------------------------------
# 11. LaTeX table formatting
# ---------------------------------------------------------------------------

@dataclass(frozen=True)
class LaTeXSettings:
    """Settings for stargazer-style LaTeX table output."""
    float_format: str = "%.4f"
    significance_levels: Tuple[float, ...] = (0.01, 0.05, 0.10)
    significance_symbols: Tuple[str, ...] = ("***", "**", "*")
    caption_position: str = "top"
    label_prefix: str = "tab:"
    font_size: str = "\\footnotesize"
    column_sep: str = "  &  "
    booktabs: bool = True
    threeparttable: bool = True
    note_width: str = "0.95\\linewidth"

LATEX: LaTeXSettings = LaTeXSettings()

# ---------------------------------------------------------------------------
# 12. Logging
# ---------------------------------------------------------------------------

LOG_LEVEL: str = os.environ.get("HEDONIC_LOG_LEVEL", "INFO")
LOG_FORMAT: str = "%(asctime)s | %(name)-20s | %(levelname)-8s | %(message)s"
LOG_DATE_FORMAT: str = "%Y-%m-%d %H:%M:%S"