def test_to_data_frame(self):
ds = Dataset(metadata=dict(fields="A,B,C"),
records=[[1, 2.2, 3.1], [3, 2.3, 3.2], [5, 2.4, 3.3]])
try:
import pandas
# noinspection PyUnresolvedReferences
df = ds.to_data_frame()
print(df)
except ImportError:
pass
def test_field_present_but_empty(self):
rule = MetaFieldOptionalRule("station", "@field_value_missing")
dataset = Dataset({"station": ""}, [])
issue = rule.eval(dataset, self._lib)
self.assertIsNotNone(issue)
self.assertEqual(ISSUE_TYPE_WARNING, issue.type)
self.assertEqual("@field_value_missing", issue.description)
def test_field_present(self):
rule = MetaFieldObsoleteRule("station", "@field_obsolete")
dataset = Dataset({"station": "whatever"}, [])
issue = rule.eval(dataset, self._lib)
self.assertIsNotNone(issue)
self.assertEqual(ISSUE_TYPE_WARNING, issue.type)
self.assertEqual("@field_obsolete", issue.description)
def test_field_empty(self):
rule = MetaFieldRequiredRule("investigator", "@required_field_missing")
dataset = Dataset({"investigator": ""}, [])
issue = rule.eval(dataset, self._lib)
self.assertIsNotNone(issue)
self.assertEqual(ISSUE_TYPE_ERROR, issue.type)
self.assertEqual("@required_field_missing", issue.description)
def test_float_larger_equal_success(self):
rule = MetaFieldCompareRule("north_lat",
"south_lat",
">=",
error="@south_north_mismatch",
data_type="number")
dataset = Dataset({"north_lat": "34.56", "south_lat": "28.33"}, [])
self.assertIsNone(rule.eval(dataset, self._lib))
def test_from_data_frame(self):
try:
import pandas
ds = Dataset.from_data_frame(
pandas.DataFrame(data=[[1, 2.2, 3.1], [3, 2.3, 3.2],
[5, 2.4, 3.3]],
columns=["A", "B", "C"]))
print(ds)
except ImportError:
pass
def test_date_smaller_equal_success(self):
rule = MetaFieldCompareRule("start_date",
"end_date",
"<=",
error="end date before start date",
data_type="date")
dataset = Dataset({
"start_date": "20080416",
"end_date": "20080416"
}, [])
self.assertIsNone(rule.eval(dataset, self._lib))
def test_int_less_than_error(self):
rule = MetaFieldCompareRule("int_1",
"int_2",
"<",
warning="should be smaller",
data_type="number")
dataset = Dataset({"int_1": "16", "int_2": "15"}, [])
issue = rule.eval(dataset, self._lib)
self.assertIsNotNone(issue)
self.assertEqual(ISSUE_TYPE_WARNING, issue.type)
self.assertEqual("should be smaller", issue.description)
def test_float_not_equal_warning(self):
rule = MetaFieldCompareRule("north_lat",
"south_lat",
"!=",
warning="should not be equal",
data_type="number")
dataset = Dataset({"north_lat": "28.33", "south_lat": "28.33"}, [])
issue = rule.eval(dataset, self._lib)
self.assertIsNotNone(issue)
self.assertEqual(ISSUE_TYPE_WARNING, issue.type)
self.assertEqual("should not be equal", issue.description)
def test_float_larger_equal_error(self):
rule = MetaFieldCompareRule("north_lat",
"south_lat",
">=",
error="@south_north_mismatch",
data_type="number")
dataset = Dataset({"north_lat": "11.56", "south_lat": "28.33"}, [])
issue = rule.eval(dataset, self._lib)
self.assertIsNotNone(issue)
self.assertEqual(ISSUE_TYPE_ERROR, issue.type)
self.assertEqual("@south_north_mismatch", issue.description)
def test_int_equal_invalid_field(self):
rule = MetaFieldCompareRule("int_1",
"int_2",
"==",
error="must be same",
data_type="number")
dataset = Dataset({"int_1": "16"}, [])
issue = rule.eval(dataset, self._lib)
self.assertIsNotNone(issue)
self.assertEqual(ISSUE_TYPE_ERROR, issue.type)
self.assertEqual("Requested field not contained in metadata: int_2",
issue.description)
def test_date_smaller_equal_failure(self):
rule = MetaFieldCompareRule("start_date",
"end_date",
"<",
error="end date before start date",
data_type="date")
dataset = Dataset({
"start_date": "20080416",
"end_date": "20080416"
}, [])
issue = rule.eval(dataset, self._lib)
self.assertIsNotNone(issue)
self.assertEqual("end date before start date", issue.description)
self.assertEqual("ERROR", issue.type)
def test_field_present(self):
rule = MetaFieldOptionalRule("station", "@field_value_missing")
dataset = Dataset({"station": "Svalbard"}, [])
self.assertIsNone(rule.eval(dataset, self._lib))
def test_field_present(self):
rule = MetaFieldRequiredRule("investigator", "@required_field_missing")
dataset = Dataset({"investigator": "Hans_Dampf"}, [])
self.assertIsNone(rule.eval(dataset, self._lib))
def test_field_missing(self):
rule = MetaFieldObsoleteRule("station", "@field_obsolete")
dataset = Dataset({}, [])
self.assertIsNone(rule.eval(dataset, self._lib))