Updated instance_id utility function

tests/test_utils.py

@@ -1,7 +1,12 @@
 import pytest
 import xarray as xr
 
-from xmip.utils import cmip6_dataset_id, google_cmip_col, model_id_match
+from xmip.utils import (
+ cmip6_dataset_id,
+ google_cmip_col,
+ instance_id_from_dataset,
+ model_id_match,
+)
 
 
 def test_google_cmip_col():
@@ -68,3 +73,80 @@ def test_cmip6_dataset_id():
  cmip6_dataset_id(ds, id_attrs=["grid_label", "activity_id", "wrong_attrs"])
  == "gl.ai.none"
  )
+
+
+class Test_instance_id_from_dataset:
+ def test_default_cmip6(self):
+ ds = xr.Dataset(
+ attrs={
+ "mip_era": "a",
+ "grid_label": "b",
+ "version": "c",
+ "activity_id": "d",
+ "institution_id": "e",
+ "source_id": "f",
+ "experiment_id": "g",
+ "member_id": "h",
+ "table_id": "i",
+ "variable_id": "j",
+ }
+ )
+ assert instance_id_from_dataset(ds) == "a.d.e.f.g.h.i.j.b.c"
+
+ def test_custom_sep(self):
+ ds = xr.Dataset(
+ attrs={
+ "a": "a",
+ "b": "b",
+ }
+ )
+ iid = instance_id_from_dataset(ds, id_schema="a.b", sep="-")
+ assert iid == "a-b"
+
+ def test_custom_schema(self):
+ ds = xr.Dataset(attrs={"some": "thing", "totally": "unrelated"})
+ assert (
+ instance_id_from_dataset(ds, id_schema="some.totally") == "thing.unrelated"
+ )
+
+ @pytest.mark.parametrize("missing_value", ["none", "some"])
+ def test_missing_attrs_print_missing(self, missing_value):
+ ds = xr.Dataset(
+ attrs={
+ "a": "a",
+ "b": "b",
+ }
+ )
+ iid = instance_id_from_dataset(
+ ds, id_schema="a.b.c", print_missing=True, missing_value=missing_value
+ )
+ assert iid == f"a.b.{missing_value}"
+
+ def test_missing_attrs_omit(self):
+ ds = xr.Dataset(
+ attrs={
+ "a": "a",
+ "b": "b",
+ }
+ )
+ iid_omit = instance_id_from_dataset(ds, id_schema="a.b.c", print_missing=False)
+ assert iid_omit == "a.b"
+
+ def test_missing_attrs_warning(self):
+ ds = xr.Dataset(
+ attrs={
+ "mip_era": "a",
+ "activity_id": "d",
+ "institution_id": "e",
+ "source_id": "f",
+ "experiment_id": "g",
+ "member_id": "h",
+ "table_id": "i",
+ "variable_id": "j",
+ }
+ )
+ with pytest.warns(
+ UserWarning,
+ match=r"Could not find dataset attributes for facets: \['grid_label', 'version'\]",
+ ):
+ instance_id_from_dataset(ds)

xmip/utils.py

@@ -3,6 +3,10 @@
 except ImportError:
  intake = None
 
+import warnings
+
+import xarray as xr
+
 
 def google_cmip_col(catalog="main"):
  """A tiny utility function to point to the 'official' pangeo cmip6 cloud files."""
@@ -56,9 +60,38 @@ def model_id_match(match_list, id_tuple):
 
 
 def _key_from_attrs(ds, attrs, sep="."):
+ raise
  return sep.join([ds.attrs[i] if i in ds.attrs.keys() else "none" for i in attrs])
 
 
+cmip_instance_id_schema = "mip_era.activity_id.institution_id.source_id.experiment_id.member_id.table_id.variable_id.grid_label.version"
+
+
+def instance_id_from_dataset(
+ ds: xr.Dataset,
+ id_schema: str = None,
+ print_missing=True,
+ missing_value="none",
+ sep=".",
+) -> str:
+ """
+ Formats a CMIP6 compatible instance id from `ds` attributes according to `id_schema` (defaults to official CMIP naming schema).
+ If `print_missing` is true missing facets as replaced with `missing_value`, otherwise missing facets are omitted.
+ """
+ if id_schema is None:
+ id_schema = cmip_instance_id_schema
+ facets = id_schema.split(".")
+ facet_dict = {k: ds.attrs.get(k, missing_value) for k in facets}
+ if not print_missing:
+ facets = [f for f in facets if facet_dict[f] != missing_value]
+ missing_value_dict = {k: v for k, v in facet_dict.items() if v == missing_value}
+ if len(missing_value_dict.keys()) > 0:
+ warnings.warn(
+ f"Could not find dataset attributes for facets: {list(missing_value_dict.keys())}"
+ )
+ return sep.join([facet_dict[f] for f in facets])
+
+
 def cmip6_dataset_id(
  ds,
  sep=".",
@@ -88,6 +121,7 @@ def cmip6_dataset_id(
  str
  Concatenated
  """
+ raise
  return _key_from_attrs(ds, id_attrs, sep=sep)