You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
96 lines
3.1 KiB
96 lines
3.1 KiB
from model.editable import Editable
|
|
from model.translation import from_container_list
|
|
|
|
|
|
class Example(Editable):
|
|
@staticmethod
|
|
def from_xml(example_xml):
|
|
ce_xml = example_xml.querySelector("corpusExample")
|
|
if ce_xml is not None:
|
|
return CorpusExample(example_xml)
|
|
else:
|
|
return MultiwordExample(example_xml)
|
|
|
|
|
|
@staticmethod
|
|
def add_clusters(entry):
|
|
nocluster_examples = []
|
|
taken_clusters = []
|
|
|
|
# gahter all taken cluster numbers and all examples without clusters
|
|
for sense in entry.senses:
|
|
for example in sense.examples:
|
|
cluster = example.get_cluster()
|
|
if cluster == -1:
|
|
nocluster_examples.append(example)
|
|
elif cluster is not None:
|
|
taken_clusters.append(cluster)
|
|
|
|
cnum = 1
|
|
for example in nocluster_examples:
|
|
while cnum in taken_clusters:
|
|
cnum += 1
|
|
taken_clusters.append(cnum)
|
|
|
|
example.cluster = cnum
|
|
|
|
|
|
def __init__(self, example_xml):
|
|
self.translations = from_container_list(example_xml.querySelectorAll("translationContainer"))
|
|
self.components = [ComponentLexeme(el) for el in example_xml.querySelectorAll("comp")]
|
|
|
|
|
|
def get_cluster(self):
|
|
return None
|
|
|
|
def get_valid_cluster(self):
|
|
return None
|
|
|
|
def text(self):
|
|
return " ".join([comp.text for comp in self.components])
|
|
|
|
|
|
class CorpusExample(Example):
|
|
def __init__(self, example_xml):
|
|
super().__init__(example_xml)
|
|
xml = example_xml.querySelector("corpusExample")
|
|
|
|
self.other_attributes = {}
|
|
for oth_attr in ["example_id", "modified", "lexical_unit_id", "audio"]:
|
|
if xml.hasAttribute(oth_attr):
|
|
self.other_attributes[oth_attr] = xml.getAttribute(oth_attr)
|
|
|
|
|
|
class MultiwordExample(Example):
|
|
def __init__(self, example_xml):
|
|
super().__init__(example_xml)
|
|
xml = example_xml.querySelector("multiwordExample")
|
|
|
|
self.other_attributes = {}
|
|
for oth_attr in ["type", "lexical_unit_id", "structure_id", "structureName", "audio", "frequency", "logDice"]:
|
|
if xml.hasAttribute(oth_attr):
|
|
self.other_attributes[oth_attr] = xml.getAttribute(oth_attr)
|
|
|
|
self.cluster_valid = False
|
|
self.cluster = -1
|
|
if xml.hasAttribute("cluster"):
|
|
self.cluster_valid = True
|
|
self.cluster = int(xml.getAttribute("cluster"))
|
|
|
|
def get_cluster(self):
|
|
return self.cluster
|
|
|
|
def get_valid_cluster(self):
|
|
return self.cluster if self.cluster_valid else None
|
|
|
|
|
|
class ComponentLexeme:
|
|
def __init__(self, xml):
|
|
self.text = xml.textContent
|
|
self.role = xml.getAttribute("role")
|
|
|
|
self.other_attributes = {}
|
|
for oth_attr in ["lexical_unit_lexeme_id", "slolex", "kol"]:
|
|
if xml.hasAttribute(oth_attr):
|
|
self.other_attributes[oth_attr] = xml.getAttribute(oth_attr)
|