flixopt 3.0.1__py3-none-any.whl → 6.0.0rc7__py3-none-any.whl

flixopt/stats_accessor.py

@@ -0,0 +1,75 @@
+"""Xarray accessor for statistics and transformations (``.fxstats``)."""
+
+from __future__ import annotations
+
+import numpy as np
+import xarray as xr
+
+
+@xr.register_dataset_accessor('fxstats')
+class DatasetStatsAccessor:
+    """Statistics/transformation accessor for any xr.Dataset. Access via ``dataset.fxstats``.
+
+    Provides data transformation methods that return new datasets.
+    Chain with ``.plotly`` for visualization.
+
+    Examples:
+        Duration curve::
+
+            ds.fxstats.to_duration_curve().plotly.line()
+    """
+
+    def __init__(self, xarray_obj: xr.Dataset) -> None:
+        self._ds = xarray_obj
+
+    def to_duration_curve(self, *, normalize: bool = True) -> xr.Dataset:
+        """Transform dataset to duration curve format (sorted values).
+
+        Values are sorted in descending order along the 'time' dimension.
+        The time coordinate is replaced with duration (percentage or index).
+
+        Args:
+            normalize: If True, x-axis shows percentage (0-100). If False, shows timestep index.
+
+        Returns:
+            Transformed xr.Dataset with duration coordinate instead of time.
+
+        Example:
+            >>> ds.fxstats.to_duration_curve().plotly.line(title='Duration Curve')
+        """
+        if 'time' not in self._ds.dims:
+            raise ValueError("Duration curve requires a 'time' dimension.")
+
+        # Sort each variable along time dimension (descending), preserving attributes
+        sorted_vars = {}
+        for var in self._ds.data_vars:
+            da = self._ds[var]
+            if 'time' not in da.dims:
+                # Keep variables without time dimension unchanged
+                sorted_vars[var] = da
+                continue
+            time_axis = da.dims.index('time')
+            sorted_values = np.flip(np.sort(da.values, axis=time_axis), axis=time_axis)
+            sorted_vars[var] = xr.DataArray(
+                sorted_values,
+                dims=da.dims,
+                coords={k: v for k, v in da.coords.items() if k != 'time'},
+                attrs=da.attrs,
+            )
+
+        # Preserve non-time coordinates from the original dataset
+        non_time_coords = {k: v for k, v in self._ds.coords.items() if k != 'time'}
+        sorted_ds = xr.Dataset(sorted_vars, coords=non_time_coords, attrs=self._ds.attrs)
+
+        # Replace time coordinate with duration
+        n_timesteps = sorted_ds.sizes['time']
+        if normalize:
+            duration_coord = np.linspace(0, 100, n_timesteps)
+            sorted_ds = sorted_ds.assign_coords({'time': duration_coord})
+            sorted_ds = sorted_ds.rename({'time': 'duration_pct'})
+        else:
+            duration_coord = np.arange(n_timesteps)
+            sorted_ds = sorted_ds.assign_coords({'time': duration_coord})
+            sorted_ds = sorted_ds.rename({'time': 'duration'})
+
+        return sorted_ds