remove unused functions (units, get_calendar)

src/xsdba/adjustment.py

@@ -14,10 +14,9 @@
 from scipy import stats
 from xarray.core.dataarray import DataArray
 
-from xsdba.base import get_calendar
 from xsdba.formatting import gen_call_string, update_history
 from xsdba.options import OPTIONS, XSDBA_EXTRA_OUTPUT, set_options
-from xsdba.units import convert_units_to, pint2str, units2str
+from xsdba.units import convert_units_to
 from xsdba.utils import uses_dask
 
 from ._adjustment import (
@@ -94,7 +93,7 @@ def _check_inputs(cls, *inputs, group):
                 )
 
         # All calendars used by the inputs
-        calendars = {get_calendar(inda, group.dim) for inda in inputs}
+        calendars = {inda.time.dt.calendar for inda in inputs}
         if not cls._allow_diff_calendars and len(calendars) > 1:
             raise ValueError(
                 "Inputs are defined on different calendars,"
@@ -246,7 +245,7 @@ def train(cls, ref: DataArray, hist: DataArray, **kwargs) -> TrainAdjust:
         ds, params = cls._train(ref, hist, **kwargs)
         obj = cls(
             _trained=True,
-            hist_calendar=get_calendar(hist),
+            hist_calendar=hist.time.dt.calendar,
             train_units=train_units,
             **params,
         )

src/xsdba/base.py

@@ -200,7 +200,7 @@ def get_coordinate(self, ds: xr.Dataset | None = None) -> xr.DataArray:
             )
         if self.prop == "dayofyear":
             if ds is not None:
-                cal = get_calendar(ds, dim=self.dim)
+                cal = ds.time.dt.calendar
                 mdoy = max(
                     xr.coding.calendar_ops._days_in_year(yr, cal)
                     for yr in np.unique(ds[self.dim].dt.year)
@@ -1081,49 +1081,6 @@ def compare_offsets(freqA: str, op: str, freqB: str) -> bool:
     return get_op(op)(t_a, t_b)
 
 
-# XC: calendar
-def get_calendar(obj: Any, dim: str = "time") -> str:
-    """Return the calendar of an object.
-
-    Parameters
-    ----------
-    obj : Any
-        An object defining some date.
-        If `obj` is an array/dataset with a datetime coordinate, use `dim` to specify its name.
-        Values must have either a datetime64 dtype or a cftime dtype.
-        `obj` can also be a python datetime.datetime, a cftime object or a pandas Timestamp
-        or an iterable of those, in which case the calendar is inferred from the first value.
-    dim : str
-        Name of the coordinate to check (if `obj` is a DataArray or Dataset).
-
-    Raises
-    ------
-    ValueError
-        If no calendar could be inferred.
-
-    Returns
-    -------
-    str
-        The Climate and Forecasting (CF) calendar name.
-        Will always return "standard" instead of "gregorian", following CF conventions 1.9.
-    """
-    if isinstance(obj, xr.DataArray | xr.Dataset):
-        return obj[dim].dt.calendar
-    if isinstance(obj, xr.CFTimeIndex):
-        obj = obj.values[0]
-    else:
-        obj = np.take(obj, 0)
-        # Take zeroth element, overcome cases when arrays or lists are passed.
-    if isinstance(obj, pydt.datetime):  # Also covers pandas Timestamp
-        return "standard"
-    if isinstance(obj, cftime.datetime):
-        if obj.calendar == "gregorian":
-            return "standard"
-        return obj.calendar
-
-    raise ValueError(f"Calendar could not be inferred from object of type {type(obj)}.")
-
-
 # XC: calendar
 def construct_offset(mult: int, base: str, start_anchored: bool, anchor: str | None):
     """Reconstruct an offset string from its parts.
@@ -1251,9 +1208,9 @@ def stack_periods(
         The coordinate of `period` is the first timestep of each window.
     """
     # Import in function to avoid cyclical imports
-    from xclim.core.units import (  # pylint: disable=import-outside-toplevel
-        ensure_cf_units,
+    from xsdba.units import (  # pylint: disable=import-outside-toplevel
         infer_sampling_units,
+        units2str,
     )
 
     stride = stride or window
@@ -1357,7 +1314,7 @@ def stack_periods(
     # Length as a pint-ready array : with proper units, but values are not usable as indexes anymore
     m, u = infer_sampling_units(da)
     lengths = lengths * m
-    lengths.attrs["units"] = ensure_cf_units(u)
+    lengths.attrs["units"] = units2str(u)
     # Start points for each period and remember parameters for unstacking
     starts = xr.DataArray(
         [da.time[slc.start].item() for slc in periods],
@@ -1429,7 +1386,7 @@ def unstack_periods(da: xr.DataArray | xr.Dataset, dim: str = "period"):
          0   o   o   o   x   x
         === === === === === === === ===
     """
-    from xclim.core.units import (  # pylint: disable=import-outside-toplevel
+    from xsdba.units import (  # pylint: disable=import-outside-toplevel
         infer_sampling_units,
     )
 

src/xsdba/processing.py

@@ -19,7 +19,7 @@
 from ._processing import _adapt_freq, _normalize, _reordering
 from .base import Grouper, parse_offset, uses_dask
 from .nbutils import _escore
-from .units import convert_units_to, harmonize_units, pint2str
+from .units import convert_units_to, harmonize_units
 from .utils import ADDITIVE, copy_all_attrs
 
 __all__ = [

src/xsdba/properties.py

@@ -23,9 +23,9 @@
 
 from xsdba.units import (
     convert_units_to,
+    infer_sampling_units,
     pint2cfattrs,
-    pint2str,
-    to_agg_units,
+    units,
     units2pint,
 )
 from xsdba.utils import uses_dask
@@ -130,11 +130,11 @@ def _mean(da: xr.DataArray, *, group: str | Grouper = "time") -> xr.DataArray:
     xr.DataArray, [same as input]
       Mean of the variable.
     """
-    units = da.units
+    u = da.units
     if group.prop != "group":
         da = da.groupby(group.name)
     out = da.mean(dim=group.dim)
-    return out.assign_attrs(units=units)
+    return out.assign_attrs(units=u)
 
 
 mean = StatisticalProperty(
@@ -164,12 +164,11 @@ def _var(da: xr.DataArray, *, group: str | Grouper = "time") -> xr.DataArray:
     xr.DataArray, [square of the input units]
         Variance of the variable.
     """
-    units = da.units
+    u = da.units
     if group.prop != "group":
         da = da.groupby(group.name)
     out = da.var(dim=group.dim)
-    u2 = units2pint(units) ** 2
-    out.attrs["units"] = pint2str(u2)
+    out.attrs["units"] = str((units(u) ** 2).units)
     return out
 
 
@@ -201,11 +200,11 @@ def _std(da: xr.DataArray, *, group: str | Grouper = "time") -> xr.DataArray:
     xr.DataArray,
         Standard deviation of the variable.
     """
-    units = da.units
+    u = da.units
     if group.prop != "group":
         da = da.groupby(group.name)
     out = da.std(dim=group.dim)
-    out.attrs["units"] = units
+    out.attrs["units"] = u
     return out
 
 
@@ -282,11 +281,11 @@ def _quantile(
     xr.DataArray, [same as input]
         Quantile {q} of the variable.
     """
-    units = da.units
+    u = da.units
     if group.prop != "group":
         da = da.groupby(group.name)
     out = da.quantile(q, dim=group.dim, keep_attrs=True).drop_vars("quantile")
-    return out.assign_attrs(units=units)
+    return out.assign_attrs(units=u)
 
 
 quantile = StatisticalProperty(
@@ -406,7 +405,12 @@ def _spell_stats(
         stat=stat,
         stat_resample=stat_resample or stat,
     ).out
-    return to_agg_units(out, da, op="count")
+    # in xclim this was managed by to_agg_units
+    # will hard-code this part for now
+    m, freq_u_raw = infer_sampling_units(da["time"])
+    with xr.set_options(keep_attrs=True):
+        out = (out * m).assign_attrs({"units": freq_u_raw})
+    return out
 
 
 spell_length_distribution = StatisticalProperty(
@@ -581,7 +585,7 @@ def _annual_cycle(
         {stat} of the annual cycle.
     """
     group = group if isinstance(group, Grouper) else Grouper(group)
-    units = da.units
+    u = da.units
 
     ac = da.groupby("time.dayofyear").mean()
     if window > 1:  # smooth the cycle
@@ -596,7 +600,7 @@ def _annual_cycle(
     # TODO: In April 2024, use a match-case.
     if stat == "absamp":
         out = ac.max("dayofyear") - ac.min("dayofyear")
-        out.attrs.update(pint2cfattrs(units2pint(units), is_difference=True))
+        out.attrs.update(pint2cfattrs(units2pint(u), is_difference=True))
     elif stat == "relamp":
         out = (ac.max("dayofyear") - ac.min("dayofyear")) * 100 / ac.mean("dayofyear")
         out.attrs["units"] = "%"
@@ -605,10 +609,10 @@ def _annual_cycle(
         out.attrs.update(units="", is_dayofyear=np.int32(1))
     elif stat == "min":
         out = ac.min("dayofyear")
-        out.attrs["units"] = units
+        out.attrs["units"] = u
     elif stat == "max":
         out = ac.max("dayofyear")
-        out.attrs["units"] = units
+        out.attrs["units"] = u
     elif stat == "asymmetry":
         out = (ac.idxmax("dayofyear") - ac.idxmin("dayofyear")) % 365 / 365
         out.attrs["units"] = "yr"
@@ -704,7 +708,7 @@ def _annual_statistic(
     xr.DataArray, [same units as input or dimensionless]
         Average annual {stat}.
     """
-    units = da.units
+    u = da.units
 
     if window > 1:
         da = da.rolling(time=window, center=True).mean()
@@ -713,7 +717,7 @@ def _annual_statistic(
 
     if stat == "absamp":
         out = yrs.max() - yrs.min()
-        out.attrs.update(pint2cfattrs(units2pint(units), is_difference=True))
+        out.attrs.update(pint2cfattrs(units2pint(u), is_difference=True))
     elif stat == "relamp":
         out = (yrs.max() - yrs.min()) * 100 / yrs.mean()
         out.attrs["units"] = "%"
@@ -971,7 +975,12 @@ def _bivariate_spell_stats(
         stat=stat,
         stat_resample=stat_resample or stat,
     ).out
-    return to_agg_units(out, da1, op="count")
+    # in xclim this was managed by to_agg_units
+    # will hard-code this part for now
+    m, freq_u_raw = infer_sampling_units(da["time"])
+    with xr.set_options(keep_attrs=True):
+        out = (out * m).assign_attrs({"units": freq_u_raw})
+    return out
 
 
 bivariate_spell_length_distribution = StatisticalProperty(
@@ -1229,7 +1238,7 @@ def _trend(
 
     numpy.polyfit
     """
-    units = da.units
+    u = da.units
 
     da = da.resample({group.dim: group.freq})  # separate all the {group}
     da_mean = da.mean(dim=group.dim)  # avg over all {group}
@@ -1250,7 +1259,7 @@ def modified_lr(
         vectorize=True,
         dask="parallelized",
     )
-    out.attrs["units"] = f"{units}/year"
+    out.attrs["units"] = f"{u}/year"
     return out