def test_from_too_large(self, caplog):
df = pd.DataFrame(
{
"a": np.arange(10),
"b": np.arange(10),
"c": np.arange(10),
"entity": ["CO2"] * 10,
"unit": ["Gg"] * 10,
"2001": np.arange(10),
}
)
df.attrs = {
"attrs": {},
"dimensions": {"CO2": ["a", "b", "c"]},
"time_format": "%Y",
}
# projected array size should be 1000 > 100
with pytest.raises(ValueError, match="Resulting array too large"):
pm2io.from_interchange_format(df, max_array_size=100)
assert "ERROR" in caplog.text
assert (
"Set with 1 entities and a total of 3 dimensions will have a size of 1,000"
in caplog.text
)
def test_from_add_coord_non_unique(self, caplog):
df = pd.DataFrame(
{
"a": np.arange(3),
"b": np.arange(3),
"c": np.arange(3),
"entity": ["CO2"] * 3,
"entity_name": ["Carbondioxide", "Carbondioxide", "Methane"],
"unit": ["Gg"] * 3,
"2001": np.arange(3),
}
)
df.attrs = {
"attrs": {},
"dimensions": {"CO2": ["a", "b", "c"]},
"time_format": "%Y",
"additional_coordinates": {"entity_name": "entity"},
}
with pytest.raises(
ValueError,
match="Different secondary coordinate values "
"for given first coordinate value for "
"entity_name.",
):
pm2io.from_interchange_format(df)
assert "ERROR" in caplog.text
assert (
"Different secondary coordinate values for given first coordinate "
"value for entity_name." in caplog.text
)
def test_inharmonic_units(minimal_ds, tmp_path):
path = tmp_path / "if"
pm2io.write_interchange_format(path, minimal_ds.pr.to_interchange_format())
df = pd.read_csv(path.with_suffix(".csv"))
df.loc[3, "unit"] = "m"
df.to_csv(path.with_suffix(".csv"),
index=False,
quoting=csv.QUOTE_NONNUMERIC)
with pytest.raises(ValueError, match="More than one unit"):
pm2io.from_interchange_format(pm2io.read_interchange_format(path))
def test_from(self):
file_input = DATA_PATH / "test_read_wide_csv_file_output.csv"
file_expected = DATA_PATH / "test_from_interchange_format_output.nc"
ds_expected = primap2.open_dataset(file_expected)
df_input = pd.read_csv(file_input, index_col=0)
dims = [
"area (ISO3)",
"category (IPCC2006)",
"scenario (general)",
"Class (class)",
"Type (type)",
"unit",
"entity",
"source",
]
attrs = {
"attrs": {
"area": "area (ISO3)",
"cat": "category (IPCC2006)",
"scen": "scenario (general)",
"sec_cats": ["Class (class)", "Type (type)"],
},
"time_format": "%Y",
"dimensions": {"*": dims},
}
ds_result = pm2io.from_interchange_format(df_input, attrs)
assert_ds_aligned_equal(ds_result, ds_expected, equal_nan=True)
def test_round_trip(any_ds: xr.Dataset, tmp_path):
path = tmp_path / "if"
pm2io.write_interchange_format(path, any_ds.pr.to_interchange_format())
with path.with_suffix(".yaml").open() as fd:
print(fd.read())
actual = pm2io.from_interchange_format(pm2io.read_interchange_format(path))
utils.assert_ds_aligned_equal(any_ds, actual)
def test_read_published_data(self):
actual = pm2io.from_interchange_format(
pm2io.read_interchange_format(
DATA_PATH / "Guetschow-et-al-2021-PRIMAP-crf96_2021-v1"
)
)
expected = primap2.open_dataset(
DATA_PATH / "Guetschow-et-al-2021-PRIMAP-crf96_2021-v1.nc"
)
assert_ds_aligned_equal(actual, expected, equal_nan=True)
def test_missing_file(minimal_ds, tmp_path):
path = tmp_path / "if"
pm2io.write_interchange_format(path, minimal_ds.pr.to_interchange_format())
with path.with_suffix(".yaml").open() as fd:
content = fd.readlines()
with path.with_suffix(".yaml").open("w") as fd:
for line in content:
if "data_file" in line:
continue
fd.write(line)
# first test automatic discovery
actual = pm2io.from_interchange_format(pm2io.read_interchange_format(path))
utils.assert_ds_aligned_equal(minimal_ds, actual)
# now test without csv file
path.with_suffix(".csv").unlink()
with pytest.raises(FileNotFoundError, match="Data file not found at"):
pm2io.read_interchange_format(path)