def test_ge_metrics_urn():
urn = "urn:great_expectations:metrics:20200403T1234.324Z:my_suite:expect_something.observed_value:column=mycol"
res = ge_urn.parseString(urn)
assert res["urn_type"] == "metrics"
assert res["run_id"] == "20200403T1234.324Z"
assert res["expectation_suite_name"] == "my_suite"
assert res["metric_name"] == "expect_something.observed_value"
kwargs_dict = parse_qs(res["metric_kwargs"])
assert kwargs_dict == {"column": ["mycol"]}
# No kwargs is ok
urn = "urn:great_expectations:metrics:20200403T1234.324Z:my_suite:expect_something.observed_value"
res = ge_urn.parseString(urn)
assert res["urn_type"] == "metrics"
assert res["run_id"] == "20200403T1234.324Z"
assert res["expectation_suite_name"] == "my_suite"
assert res["metric_name"] == "expect_something.observed_value"
assert "kwargs_dict" not in res
def test_ge_stores_urn():
urn = "urn:great_expectations:stores:my_store:mymetric:kw=param"
res = ge_urn.parseString(urn)
assert res["urn_type"] == "stores"
assert res["store_name"] == "my_store"
assert res["metric_name"] == "mymetric"
kwargs_dict = parse_qs(res["metric_kwargs"])
assert kwargs_dict == {
"kw": ["param"],
}
# No kwargs is ok
urn = "urn:great_expectations:stores:my_store:mymetric"
res = ge_urn.parseString(urn)
assert res["urn_type"] == "stores"
assert res["store_name"] == "my_store"
assert res["metric_name"] == "mymetric"
assert "metric_kwargs" not in res
def test_ge_validations_urn():
# We should be able to parse validations urns
urn = (
"urn:great_expectations:validations:my_suite:expect_something.observed_value:query=s%20tring&query="
"string3&query2=string2")
res = ge_urn.parseString(urn)
assert res["urn_type"] == "validations"
assert res["expectation_suite_name"] == "my_suite"
assert res["metric_name"] == "expect_something.observed_value"
kwargs_dict = parse_qs(res["metric_kwargs"])
assert kwargs_dict == {
"query": ["s tring", "string3"],
"query2": ["string2"]
}
# no kwargs is ok
urn = "urn:great_expectations:validations:my_suite:expect_something.observed_value"
res = ge_urn.parseString(urn)
assert res["urn_type"] == "validations"
assert res["expectation_suite_name"] == "my_suite"
assert res["metric_name"] == "expect_something.observed_value"
assert "metric_kwargs" not in res
def test_invalid_urn():
# Must start with "urn:great_expectations"
with pytest.raises(ParseException) as e:
ge_urn.parseString("not_a_ge_urn")
assert "not_a_ge_urn" in e.value.line
# Must have one of the recognized types
with pytest.raises(ParseException) as e:
ge_urn.parseString("urn:great_expectations:foo:bar:baz:bin:barg")
assert "urn:great_expectations:foo:bar:baz:bin:barg" in e.value.line
# Cannot have too many parts
with pytest.raises(ParseException) as e:
ge_urn.parseString(
"urn:great_expectations:validations:foo:bar:baz:bin:barg:boo")
assert "urn:great_expectations:validations:foo:bar:baz:bin:barg:boo" in e.value.line
def parse_evaluation_parameter(parameter_expression,
evaluation_parameters=None,
data_context=None):
"""Use the provided evaluation_parameters dict to parse a given parameter expression.
Args:
parameter_expression (str): A string, potentially containing basic arithmetic operations and functions,
and variables to be substituted
evaluation_parameters (dict): A dictionary of name-value pairs consisting of values to substitute
data_context (DataContext): A data context to use to obtain metrics, if necessary
The parser will allow arithmetic operations +, -, /, *, as well as basic functions, including trunc() and round() to
obtain integer values when needed for certain expectations (e.g. expect_column_value_length_to_be_between).
Valid variables must begin with an alphabetic character and may contain alphanumeric characters plus '_' and '$',
EXCEPT if they begin with the string "urn:great_expectations" in which case they may also include additional
characters to support inclusion of GE URLs (see :ref:`evaluation_parameters` for more information).
"""
if evaluation_parameters is None:
evaluation_parameters = {}
# Calling get_parser clears the stack
parser = expr.get_parser()
try:
L = parser.parseString(parameter_expression, parseAll=True)
except ParseException as err:
L = [
"Parse Failure", parameter_expression,
(str(err), err.line, err.column)
]
if len(L) == 1 and L[0] not in evaluation_parameters:
# In this special case there were no operations to find, so only one value, but we don't have something to
# substitute for that value
try:
res = ge_urn.parseString(L[0])
if res["urn_type"] == "stores":
store = data_context.stores.get(res["store_name"])
return store.get_query_result(res["metric_name"],
res.get("metric_kwargs", {}))
else:
logger.error(
"Unrecognized urn_type in ge_urn: must be 'stores' to use a metric store."
)
raise EvaluationParameterError(
"No value found for $PARAMETER " + str(L[0]))
except ParseException:
raise EvaluationParameterError("No value found for $PARAMETER " +
str(L[0]))
except AttributeError:
logger.warning(
"Unable to get store for store-type valuation parameter.")
raise EvaluationParameterError("No value found for $PARAMETER " +
str(L[0]))
elif len(L) == 1:
# In this case, we *do* have a substitution for a single type. We treat this specially because in this
# case, we allow complex type substitutions (i.e. do not coerce to string as part of parsing)
return evaluation_parameters[L[0]]
elif len(L) == 0 or L[0] != "Parse Failure":
for i, ob in enumerate(expr.exprStack):
if isinstance(ob, str) and ob in evaluation_parameters:
expr.exprStack[i] = str(evaluation_parameters[ob])
else:
err_str, err_line, err_col = L[-1]
raise EvaluationParameterError(
f"Parse Failure: {err_str}\nStatement: {err_line}\nColumn: {err_col}"
)
try:
result = expr.evaluate_stack(expr.exprStack)
except Exception as e:
exception_traceback = traceback.format_exc()
exception_message = (
f'{type(e).__name__}: "{str(e)}". Traceback: "{exception_traceback}".'
)
logger.debug(exception_message, e, exc_info=True)
raise EvaluationParameterError(
"Error while evaluating evaluation parameter expression: " +
str(e))
return result