src/data/loaders/hydrometric.py

"""
Hydrometric data loader for Hub'Eau API v2 discharge and water level data.
Single Responsibility: Load and parse hydrometric observations.
"""
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
from pathlib import Path
from typing import Optional, List
from .base import BaseDataLoader


class HydrometricLoader(BaseDataLoader):
    """
    Loads Hub'Eau hydrometric data (discharge & water level).

    Handles actual Hub'Eau API v2 structure:
    - date_obs_elab: observation date
    - station_code: station ID
    - resultat_obs_elab: measured value
    - grandeur_hydro_elab: variable type (QmnJ, HIXnJ, etc.)
    - code_qualification: quality code (16=acceptable, 20=good)
    """

    def __init__(
        self,
        data_path: Path,
        min_quality: bool = True,
        station_ids: Optional[List[str]] = None,
        discharge_grandeur: str = "QmnJ",     # Fixed: Hub'Eau daily mean discharge
        waterlevel_grandeur: str = "HIXnJ"    # Fixed: Hub'Eau daily max water level
    ):
        """
        Initialize hydrometric loader.

        Args:
            data_path: Path to datasets/hydrometric directory
            min_quality: Only keep good quality data (quality_code >= 16)
            station_ids: Optional list of station IDs to filter
            discharge_grandeur: Hub'Eau code for discharge ('QmnJ', 'QIXnJ', etc.)
            waterlevel_grandeur: Hub'Eau code for water level ('HIXnJ', 'HIXM', etc.)
        """
        super().__init__(data_path)
        self.min_quality = min_quality
        self.station_ids = station_ids
        self.discharge_grandeur = discharge_grandeur
        self.waterlevel_grandeur = waterlevel_grandeur

    def load(self) -> pd.DataFrame:
        """
        Load hydrometric data from Hub'Eau CSV files.

        Returns:
            DataFrame with [date, station_code, discharge_m3s, waterlevel_mm]
        """
        data_dir = Path(self.data_path)

        discharge_file = data_dir / "discharge_observations.csv"
        waterlevel_file = data_dir / "waterlevel_observations.csv"

        dfs = []

        if discharge_file.exists():
            discharge_df = self._load_single_variable(
                file_path=discharge_file, 
                value_col='discharge_m3s',
                grandeur_code=self.discharge_grandeur
            )
            dfs.append(discharge_df)

        if waterlevel_file.exists():
            waterlevel_df = self._load_single_variable(
                file_path=waterlevel_file, 
                value_col='waterlevel_mm',
                grandeur_code=self.waterlevel_grandeur
            )
            dfs.append(waterlevel_df)

        if not dfs:
            raise FileNotFoundError(f"No hydrometric files found in {data_dir}")

        # Merge discharge and water level
        df = dfs[0] if len(dfs) == 1 else pd.merge(
            dfs[0], dfs[1],
            on=['date', 'station_code'],
            how='outer'
        )

        # Filter by station IDs
        if self.station_ids:
            df = df[df["station_code"].isin(self.station_ids)]

        return df.sort_values(['date', 'station_code']).reset_index(drop=True)

    def _load_single_variable(
        self, 
        file_path: Path, 
        value_col: str, 
        grandeur_code: Optional[str] = None
    ) -> pd.DataFrame:
        """Load and process a single Hub'Eau CSV file."""
        df = pd.read_csv(file_path)

        # Rename to standard columns
        df = df.rename(columns={
            'date_obs_elab': 'date',
            'resultat_obs_elab': 'value',
            'code_qualification': 'quality_code',
            'grandeur_hydro_elab': 'grandeur_code'
        })

        # Explicitly filter by Hub'Eau variable code
        if grandeur_code and 'grandeur_code' in df.columns:
            df = df[df['grandeur_code'] == grandeur_code].copy()

        # Convert date
        df['date'] = pd.to_datetime(df['date'])

        # Filter by quality (16=acceptable, 20=good)
        if self.min_quality and 'quality_code' in df.columns:
            df = df[df['quality_code'] >= 16].copy()

        # Rename value column
        df[value_col] = df['value']

        # Keep relevant columns
        df = df[['date', 'station_code', value_col]].copy()

        # Deduplicate safely
        df = df.sort_values('date').drop_duplicates(
            subset=['date', 'station_code'],
            keep='last'
        )

        return df

    def get_metadata(self) -> dict:
        """Get hydrometric data metadata."""
        meta = super().get_metadata()
        meta.update({
            "data_type": "hydrometric",
            "source": "Hub'Eau API v2",
            "quality_filter": self.min_quality,
            "filtered_stations": self.station_ids,
            "discharge_grandeur": self.discharge_grandeur,
            "waterlevel_grandeur": self.waterlevel_grandeur
        })
        return meta

    # ------------------------------------------------------------------
    # Plotting
    # ------------------------------------------------------------------

    def _plot_timeseries(
        self,
        df: pd.DataFrame,
        value_col: str,
        ylabel: str,
        title: str,
        stations: Optional[List[str]] = None,
        ax: Optional[plt.Axes] = None,
        figsize: tuple = (12, 5),
        save_path: Optional[Path] = None,
    ) -> plt.Axes:
        """Shared line-plot logic for a single variable, one line per station."""
        if value_col not in df.columns:
            raise ValueError(f"Column '{value_col}' not found in data")

        plot_df = df.dropna(subset=[value_col])
        if stations:
            plot_df = plot_df[plot_df["station_code"].isin(stations)]

        if plot_df.empty:
            raise ValueError("No data available to plot for the given stations/variable")

        standalone = ax is None
        if standalone:
            fig, ax = plt.subplots(figsize=figsize)

        for station_code, group in plot_df.groupby("station_code"):
            group = group.sort_values("date")
            ax.plot(group["date"], group[value_col], marker="o", markersize=2,
                     linewidth=1, label=station_code)

        ax.set_title(title)
        ax.set_xlabel("Date")
        ax.set_ylabel(ylabel)
        ax.xaxis.set_major_locator(mdates.AutoDateLocator())
        ax.xaxis.set_major_formatter(mdates.ConciseDateFormatter(ax.xaxis.get_major_locator()))
        ax.legend(title="Station", fontsize=8, loc="best")
        ax.grid(True, alpha=0.3)

        if standalone:
            plt.tight_layout()
            if save_path:
                plt.savefig(save_path, dpi=150, bbox_inches="tight")
                print(f"Plot saved to {save_path}")

        return ax

    def plot_waterlevel(
        self,
        df: Optional[pd.DataFrame] = None,
        stations: Optional[List[str]] = None,
        figsize: tuple = (12, 5),
        save_path: Optional[Path] = None,
    ) -> plt.Axes:
        """
        Plot water level time series, one line per station.

        Args:
            df: Optional pre-loaded DataFrame. If not provided, data is loaded from disk.
            stations: Optional list of station codes to plot. Defaults to all stations present.
            figsize: Figure size in inches.
            save_path: If provided, saves the figure to this path.

        Returns:
            The matplotlib Axes object.
        """
        if df is None:
            df = self.load()
        return self._plot_timeseries(
            df,
            value_col="waterlevel_mm",
            ylabel="Water level (mm)",
            title="Water Level Observations",
            stations=stations,
            figsize=figsize,
            save_path=save_path,
        )

    def plot_discharge(
        self,
        df: Optional[pd.DataFrame] = None,
        stations: Optional[List[str]] = None,
        figsize: tuple = (12, 5),
        save_path: Optional[Path] = None,
    ) -> plt.Axes:
        """
        Plot discharge time series, one line per station.

        Args:
            df: Optional pre-loaded DataFrame. If not provided, data is loaded from disk.
            stations: Optional list of station codes to plot. Defaults to all stations present.
            figsize: Figure size in inches.
            save_path: If provided, saves the figure to this path.

        Returns:
            The matplotlib Axes object.
        """
        if df is None:
            df = self.load()
        return self._plot_timeseries(
            df,
            value_col="discharge_m3s",
            ylabel="Discharge (m³/s)",
            title="Discharge Observations",
            stations=stations,
            figsize=figsize,
            save_path=save_path,
        )

    def plot_station(
        self,
        station_id: str,
        df: Optional[pd.DataFrame] = None,
        figsize: tuple = (12, 8),
        save_path: Optional[Path] = None,
    ) -> "plt.Figure":
        """
        Plot discharge and water level for a single station, stacked on
        two subplots so their differing scales don't distort each other.

        Args:
            station_id: The station code to plot.
            df: Optional pre-loaded DataFrame. If not provided, data is loaded from disk.
            figsize: Figure size in inches.
            save_path: If provided, saves the figure to this path.

        Returns:
            The matplotlib Figure object.
        """
        if df is None:
            df = self.load()

        station_df = df[df["station_code"] == station_id]
        if station_df.empty:
            raise ValueError(f"No data found for station '{station_id}'")

        fig, (ax1, ax2) = plt.subplots(2, 1, figsize=figsize, sharex=True)

        if "discharge_m3s" in station_df.columns and station_df["discharge_m3s"].notna().any():
            self._plot_timeseries(
                station_df, "discharge_m3s", "Discharge (m³/s)",
                f"Discharge — {station_id}", ax=ax1,
            )
        else:
            ax1.set_title(f"Discharge — {station_id} (no data)")

        if "waterlevel_mm" in station_df.columns and station_df["waterlevel_mm"].notna().any():
            self._plot_timeseries(
                station_df, "waterlevel_mm", "Water level (mm)",
                f"Water Level — {station_id}", ax=ax2,
            )
        else:
            ax2.set_title(f"Water Level — {station_id} (no data)")

        for ax in (ax1, ax2):
            ax.get_legend().remove() if ax.get_legend() else None

        plt.tight_layout()
        if save_path:
            plt.savefig(save_path, dpi=150, bbox_inches="tight")
            print(f"Plot saved to {save_path}")

        return fig

    def plot_data_availability(
        self,
        df: Optional[pd.DataFrame] = None,
        figsize: tuple = (10, 6),
        save_path: Optional[Path] = None,
    ) -> plt.Axes:
        """
        Plot a bar chart of observation counts per station, split by variable.
        Useful for spotting stations with sparse or missing coverage.

        Args:
            df: Optional pre-loaded DataFrame. If not provided, data is loaded from disk.
            figsize: Figure size in inches.
            save_path: If provided, saves the figure to this path.

        Returns:
            The matplotlib Axes object.
        """
        if df is None:
            df = self.load()

        counts = pd.DataFrame({
            "discharge_m3s": df.groupby("station_code")["discharge_m3s"].count()
                if "discharge_m3s" in df.columns else 0,
            "waterlevel_mm": df.groupby("station_code")["waterlevel_mm"].count()
                if "waterlevel_mm" in df.columns else 0,
        }).fillna(0)

        fig, ax = plt.subplots(figsize=figsize)
        counts.plot(kind="bar", ax=ax, color=["steelblue", "darkorange"])

        ax.set_title("Observation Count by Station")
        ax.set_xlabel("Station")
        ax.set_ylabel("Number of observations")
        ax.legend(title="Variable")
        ax.grid(True, alpha=0.3, axis="y")
        plt.xticks(rotation=45, ha="right")
        plt.tight_layout()

        if save_path:
            plt.savefig(save_path, dpi=150, bbox_inches="tight")
            print(f"Plot saved to {save_path}")

        return ax

    def plot_rating_curve(
        self,
        station_id: str,
        df: Optional[pd.DataFrame] = None,
        figsize: tuple = (8, 8),
        save_path: Optional[Path] = None,
    ) -> plt.Axes:
        """
        Plot discharge vs. water level for a single station as a scatter
        (a simple rating-curve style view). Points are colored by year to
        help spot rating shifts (e.g. channel changes) over time.

        Args:
            station_id: The station code to plot.
            df: Optional pre-loaded DataFrame. If not provided, data is loaded from disk.
            figsize: Figure size in inches.
            save_path: If provided, saves the figure to this path.

        Returns:
            The matplotlib Axes object.
        """
        if df is None:
            df = self.load()

        station_df = df[df["station_code"] == station_id].dropna(
            subset=["discharge_m3s", "waterlevel_mm"]
        )
        if station_df.empty:
            raise ValueError(
                f"No overlapping discharge/water level data for station '{station_id}'"
            )

        fig, ax = plt.subplots(figsize=figsize)
        years = station_df["date"].dt.year
        scatter = ax.scatter(
            station_df["waterlevel_mm"], station_df["discharge_m3s"],
            c=years, cmap="viridis", s=15, alpha=0.7,
        )

        cbar = plt.colorbar(scatter, ax=ax)
        cbar.set_label("Year")

        ax.set_title(f"Rating Curve — {station_id}")
        ax.set_xlabel("Water level (mm)")
        ax.set_ylabel("Discharge (m³/s)")
        ax.grid(True, alpha=0.3)
        plt.tight_layout()

        if save_path:
            plt.savefig(save_path, dpi=150, bbox_inches="tight")
            print(f"Plot saved to {save_path}")

        return ax

    def plot_seasonal_climatology(
        self,
        variable: str = "discharge_m3s",
        df: Optional[pd.DataFrame] = None,
        stations: Optional[List[str]] = None,
        figsize: tuple = (10, 6),
        save_path: Optional[Path] = None,
    ) -> plt.Axes:
        """
        Plot the monthly climatology (median with 25th-75th percentile band)
        of a variable, one line per station. Shows the typical seasonal
        cycle and its spread across all years of record.

        Args:
            variable: Either 'discharge_m3s' or 'waterlevel_mm'.
            df: Optional pre-loaded DataFrame. If not provided, data is loaded from disk.
            stations: Optional list of station codes to include. Defaults to all.
            figsize: Figure size in inches.
            save_path: If provided, saves the figure to this path.

        Returns:
            The matplotlib Axes object.
        """
        if df is None:
            df = self.load()

        if variable not in df.columns:
            raise ValueError(f"Column '{variable}' not found in data")

        plot_df = df.dropna(subset=[variable]).copy()
        if stations:
            plot_df = plot_df[plot_df["station_code"].isin(stations)]
        if plot_df.empty:
            raise ValueError("No data available to plot for the given stations/variable")

        plot_df["month"] = plot_df["date"].dt.month

        fig, ax = plt.subplots(figsize=figsize)
        for station_code, group in plot_df.groupby("station_code"):
            stats = group.groupby("month")[variable].agg(
                median="median", q25=lambda x: x.quantile(0.25), q75=lambda x: x.quantile(0.75)
            )
            line, = ax.plot(stats.index, stats["median"], marker="o", label=station_code)
            ax.fill_between(stats.index, stats["q25"], stats["q75"],
                             color=line.get_color(), alpha=0.15)

        ylabel = "Discharge (m³/s)" if variable == "discharge_m3s" else "Water level (mm)"
        ax.set_title(f"Seasonal Climatology — {ylabel}")
        ax.set_xlabel("Month")
        ax.set_ylabel(ylabel)
        ax.set_xticks(range(1, 13))
        ax.set_xticklabels(["Jan","Feb","Mar","Apr","May","Jun",
                             "Jul","Aug","Sep","Oct","Nov","Dec"])
        ax.legend(title="Station", fontsize=8)
        ax.grid(True, alpha=0.3)
        plt.tight_layout()

        if save_path:
            plt.savefig(save_path, dpi=150, bbox_inches="tight")
            print(f"Plot saved to {save_path}")

        return ax