def fuzzymatch(set, value):
result = set.get(value)
if result and result[0][0] > 0.2:
log.warning(f"Did you mean {result[0][1]} (matched from {value})?")
return result[0][1]
else:
return None
def lint(session, model):
events = []
for CIM_class in tqdm(model.schema.class_hierarchy("dfs"),
desc=f"Linting...",
leave=True):
query = session.query(CIM_class.class_)
for prop in CIM_class.props:
if not prop.optional and prop.used:
total = query.count()
objects = query.filter_by(**{prop.full_label: None}).count()
if objects:
events.append({
"Class": CIM_class.label,
"Property": prop.full_label,
"Total": total,
"Type": "Missing",
"Violations": objects,
"Unique": None
})
log.debug(
f"Missing mandatory property {prop.full_label} for "
f"{objects} instances of type {CIM_class.label}.")
if prop.range:
try:
if isinstance(prop.range, CIMClass):
col = getattr(CIM_class.class_,
prop.full_label + "_id")
validity = session.query(col).except_(
session.query(prop.range.class_.id))
elif isinstance(prop.range, CIMEnum):
col = getattr(CIM_class.class_,
prop.full_label + "_name")
validity = session.query(col).except_(
session.query(CIMEnumValue.name))
except AttributeError:
log.warning(
f"Couldn't determine validity of {prop.full_label} on "
f"{CIM_class.label}. The linter does not yet support "
f"many-to-many relationships.")
# ToDo: Association table errors are currently not caught
else:
count = validity.count()
# query.except() returns (None) if right hand side table is empty
if count > 1 or (count == 1 and tuple(
validity.one())[0] is not None):
non_unique = query.filter(
col.in_(val[0]
for val in validity.all())).count()
events.append({
"Class": CIM_class.label,
"Property": prop.full_label,
"Total": total,
"Type": "Invalid",
"Violations": non_unique,
"Unique": count
})
return pivot_table(DataFrame(events),
values=["Violations", "Unique"],
index=["Type", "Class", "Total", "Property"])
def parse_values(self, el, session):
if not self.parent:
argmap = {}
insertables = []
else:
argmap, insertables = self.parent.parse_values(el, session)
props = [prop for prop in self.props if prop.used]
for prop in props:
value = prop.xpath(el)
if prop.many_remote and prop.used and value:
_id = [el.attrib.values()[0]]
_remote_ids = []
if len(set(value)) > 1:
for raw_value in value:
_remote_ids = _remote_ids + [
v for v in raw_value.split("#") if len(v)
]
else:
_remote_ids = [v for v in value[0].split("#") if len(v)]
_ids = _id * len(_remote_ids)
# Insert tuples in chunks of 400 elements max
for chunk in chunks(list(zip(_ids, _remote_ids)), 400):
_ins = prop.association_table.insert([{
f"{prop.domain.label}_id":
_id,
f"{prop.range.label}_id":
_remote_id
} for (_id, _remote_id) in chunk])
insertables.append(_ins)
elif len(value) == 1 or len(set(value)) == 1:
value = value[0]
if isinstance(prop.range, CIMEnum):
argmap[prop.key] = shorten_namespace(value, self.nsmap)
else:
try:
t = prop.mapped_datatype
if t == "Float":
argmap[prop.key] = float(value)
elif t == "Boolean":
argmap[prop.key] = value.lower() == "true"
elif t == "Integer":
argmap[prop.key] = int(value)
elif len([v for v in value.split("#") if v]) > 1:
log.warning(
f"Ambiguous data values for {self.name}:{prop.key}: {len(set(value))} unique values. "
f"(Skipped)")
# If reference doesn't resolve value is set to None (Validation
# has to catch missing obligatory values)
else:
argmap[prop.key] = value.replace("#", "")
except ValueError:
argmap[prop.key] = value.replace("#", "")
elif len(value) > 1:
log.warning(
f"Ambiguous data values for {self.name}:{prop.key}: {len(set(value))} unique values. "
f"(Skipped)")
# If reference doesn't resolve value is set to None (Validation
# has to catch missing obligatory values)
return argmap, insertables
def _raw_property(self, property_identifier) -> Union[list, str, None]:
"""
Extract a property from the CIM entity
:param property_identifier: property name
:return: The CIM entity's property as a list, a string, or None
"""
if self.Attributes[property_identifier] is None:
xp = self.XPathMap
if property_identifier not in xp.keys():
raise KeyError(f"Invalid property_identifier name {property_identifier}.")
results = xp[property_identifier](self.description) # pylint: disable=unsubscriptable-object
if len(set(results)) == 1:
self.Attributes[property_identifier] = results[0]
elif not results:
self.Attributes[property_identifier] = None
else:
log.warning(f"Ambiguous class property_identifier ({property_identifier}) for {self.name}.")
self.Attributes[property_identifier] = [result for result in set(results)]
return self.Attributes[property_identifier]
def merge_sources(sources):
d_ = defaultdict(dict)
from lxml.etree import XPath
xp = {
"id": XPath("@rdf:ID", namespaces=get_nsmap(sources)),
"about": XPath("@rdf:about", namespaces=get_nsmap(sources))
}
for source in sources:
for element in source.tree.getroot():
try:
uuid = determine_uuid(element, xp)
classname = shorten_namespace(element.tag,
HDict(get_nsmap(sources)))
if classname not in d_ or uuid not in d_[classname].keys():
d_[classname][uuid] = element
else:
[d_[classname][uuid].append(sub) for sub in element] # pylint: disable=expression-not-assigned
except ValueError:
log.warning(f"Skipped element during merge: {element}.")
return d_
def update_path(self, path):
if path is None:
out_dir = os.getcwd()
elif isinstance(path, list):
try:
out_dir = os.path.commonpath(
[os.path.abspath(path) for path in path])
except ValueError:
# Paths are on different drives - default to cwd.
log.warning(
f"Datasources have no common root. Database file will be saved to {os.getcwd()}"
)
out_dir = os.getcwd()
else:
out_dir = os.path.abspath(path)
if not os.path.isabs(self.path):
if os.path.isdir(out_dir):
db_path = os.path.join(out_dir, self.path)
else:
db_path = os.path.join(os.path.dirname(out_dir), "out.db")
else:
db_path = os.path.abspath(self.path)
self.path = db_path
def _generate(self):
xp_type_res = XPath(f"rdf:type/@rdf:resource",
namespaces=self.root.nsmap)
xp_stype_res = XPath(f"cims:stereotype/@rdf:resource",
namespaces=self.root.nsmap)
xp_stype_txt = XPath(f"cims:stereotype/text()",
namespaces=self.root.nsmap)
postponed = []
for element in self.root:
type_res = xp_type_res(element)
stype_res = xp_stype_res(element)
stype_txt = xp_stype_txt(element)
if Schema._isclass(type_res):
if Schema._isenum(stype_res):
obj = CIMEnum(element)
self.Elements["CIMEnum"][obj.name].append(obj)
elif Schema._isdt(stype_txt):
obj = CIMDT(element)
self.Elements["CIMDT"][obj.name].append(obj)
else:
obj = CIMClass(element)
self.Elements["CIMClass"][obj.name].append(obj)
elif Schema._isprop(type_res):
postponed.append(element)
elif Schema._ispackage(type_res):
obj = CIMPackage(element)
self.Elements["CIMPackage"][obj.name].append(obj)
elif type_res:
postponed.append(element)
else:
obj = SchemaElement(element)
log.warning(f"Element skipped: {obj.name}")
for element in postponed:
type_res = xp_type_res(element)
if Schema._isprop(type_res):
obj = CIMProp(element)
if obj._domain in self.Elements["CIMDT"].keys():
if obj.name.endswith(".unit"):
obj = CIMDTUnit(element)
self.Elements["CIMDTUnit"][obj.name].append(obj)
elif obj.name.endswith(".value"):
obj = CIMDTValue(element)
self.Elements["CIMDTValue"][obj.name].append(obj)
elif obj.name.endswith(".multiplier"):
obj = CIMDTMultiplier(element)
self.Elements["CIMDTMultiplier"][obj.name].append(obj)
elif obj.name.endswith(".denominatorUnit"):
obj = CIMDTDenominatorUnit(element)
self.Elements["CIMDTDenominatorUnit"][obj.name].append(
obj)
elif obj.name.endswith(".denominatorMultiplier"):
obj = CIMDTDenominatorMultiplier(element)
self.Elements["CIMDTDenominatorMultiplier"][
obj.name].append(obj)
else:
obj = CIMDTProperty(element)
self.Elements["CIMDTProperty"][obj.name].append(obj)
else:
self.Elements["CIMProp"][obj.name].append(obj)
continue
obj = CIMEnumValue(element)
if obj._enum_name and obj._enum_name in self.Elements[
"CIMEnum"].keys():
self.Elements["CIMEnumValue"][obj.name].append(obj)
else:
log.debug(f"Failed to identify purpose for {type_res}")
self._merge_elements()
for key, value in self.Elements.items():
if value:
log.debug(f"Generated {len(value)} {key}.")
def test_all(runslow=False):
if runslow:
pytest.main([os.path.join(_TESTROOT), "--runslow"])
else:
pytest.main([os.path.join(_TESTROOT)])
except ModuleNotFoundError:
pass
try:
# See if we already know a schemaroot
CONFIG["Paths"]["SCHEMAROOT"] = get_path("SCHEMAROOT")
if not os.path.isdir(CONFIG["Paths"]["SCHEMAROOT"]):
# Is schemaroot an actual directory?
log.warning(f"Invalid schema path in configuration.")
raise NotADirectoryError
except (KeyError, NotADirectoryError):
if os.path.isdir(os.path.join(_PACKAGEROOT, "res", "schemata")):
# Look in the default path
CONFIG["Paths"]["SCHEMAROOT"] = os.path.join(_PACKAGEROOT, "res", "schemata")
log.info(f"Found schemata in default location.")
else:
# Ask user to configure
log.warning(f"No schemata configured. Use cimpyorm.configure(path_to_schemata) to set-up.")
from cimpyorm.api import configure
try:
# See if we already know a datasetroot
CONFIG["Paths"]["DATASETROOT"] = get_path("DATASETROOT")
if not os.path.isdir(CONFIG["Paths"]["DATASETROOT"]):
