sporada/nlp
1
0
Fork 0
mirror of https://gitlab.ub.uni-bielefeld.de/sfb1288inf/nlp.git synced 2024-12-26 08:34:17 +00:00
Code Issues Projects Releases Wiki Activity

optimizations

Browse Source
This commit is contained in:
Patrick Jentsch 2021-08-11 16:47:29 +02:00
parent 0ba0c14b72
commit 29ccfac4f6
1 changed files with 60 additions and 23 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

83
packages/stand-off-data-py/stand_off_data/models.py
View File

@ -2,12 +2,26 @@ from xml.sax.saxutils import escape
class StandOffData: class StandOffData:
def __init__(self, attrs): def __init__(self, attrs={}):
self.meta = attrs.get('meta', {}) self.meta = attrs.get('meta', {})
self.lookup = {tag_definition.id: tag_definition for tag_definition in self.lookup = {}
[TagDefinition(x) for x in attrs.get('tags', [])]} for x in attrs.get('tags', []):
self.annotations = [TagAnnotation(x, self.lookup) for x in self.add_tag_definition(x)
attrs.get('annotations', [])] self.annotations = [TagAnnotation(x, self.lookup)
for x in attrs.get('annotations', [])]
def add_tag_definition(self, attrs):
tag_definition = TagDefinition(attrs)
if tag_definition.id in self.lookup:
raise Exception('Tag id already in use: {}'.format(self.to_dict()))
self.lookup[tag_definition.id] = tag_definition
def to_dict(self):
return {
'meta': self.meta,
'lookup': {k: v.to_dict() for k, v in self.lookup.items()},
'annotations': [x.to_dict() for x in self.annotations]
}
def to_vrt(self, text): def to_vrt(self, text):
# Devide annotations into CWB's verticalized text format (.vrt) logic # Devide annotations into CWB's verticalized text format (.vrt) logic
@ -88,21 +102,23 @@ class TagAnnotation:
def __init__(self, attrs, lookup): def __init__(self, attrs, lookup):
self.lookup = lookup self.lookup = lookup
self.tag_id = attrs['tag_id'] self.tag_id = attrs['tag_id']
if self.tag_id not in self.lookup:
raise Exception('Unknown tag id: {}'.format(self.tag_id))
self.start = attrs['start'] self.start = attrs['start']
self.end = attrs['end'] self.end = attrs['end']
if self.start >= self.end:
raise Exception('start must be lower then end')
self.properties = [ self.properties = [
PropertyAnnotation({**x, 'tag_id': self.tag_id}, self.lookup) PropertyAnnotation(x, self.lookup[self.tag_id].properties)
for x in attrs.get('properties', []) for x in attrs.get('properties', [])
] ]
''' Sanity checks '''
if self.tag_id not in self.lookup:
raise Exception('Unknown tag: {}'.format(self.to_dict()))
if self.end < self.start:
raise Exception('Annotation end less then start: '
'{}'.format(self.to_dict()))
property_ids = [x.property_id for x in self.properties] property_ids = [x.property_id for x in self.properties]
for required_property_id in self.lookup[self.tag_id].required_properties: for required_property_id, required_property in self.lookup[self.tag_id].required_properties.items(): # noqa
if required_property_id not in property_ids: if required_property_id not in property_ids:
raise Exception( raise Exception('Missing required property: '
'Missing required property: {}'.format(required_property_id)) '{}'.format(required_property.to_dict()))
@property @property
def name(self): def name(self):
@ -151,15 +167,15 @@ class PropertyAnnotation:
def __init__(self, attrs, lookup): def __init__(self, attrs, lookup):
self.lookup = lookup self.lookup = lookup
self.property_id = attrs['property_id'] self.property_id = attrs['property_id']
self.tag_id = attrs['tag_id']
if self.property_id not in self.lookup[self.tag_id].properties:
raise Exception('Unknown property id: {}'.format(self.property_id))
self.value = attrs['value'] self.value = attrs['value']
# TODO: Process attrs['possibleValues'] as self.labels (no id?) # TODO: Process attrs['possibleValues'] as self.labels (no id?)
''' Sanity checks '''
if self.property_id not in self.lookup:
raise Exception('Unknown property: {}'.format(self.to_dict()))
@property @property
def name(self): def name(self):
return self.lookup[self.tag_id].properties[self.property_id].name return self.lookup[self.property_id].name
def to_dict(self): def to_dict(self):
return { return {
@ -174,18 +190,30 @@ class TagDefinition:
self.id = attrs['id'] self.id = attrs['id']
self.name = attrs['name'] self.name = attrs['name']
self.description = attrs.get('description', '') self.description = attrs.get('description', '')
self.properties = { self.properties = {}
property_definition.id: property_definition for x in attrs.get('properties', []):
for property_definition in [ self.add_property_definition(x)
PropertyDefinition(x) for x in attrs.get('properties', [])
] def add_property_definition(self, attrs):
} property_definition = PropertyDefinition(attrs)
if property_definition.id in self.properties:
raise Exception('Property id already in use: '
'{}'.format(property_definition.to_dict()))
self.properties[property_definition.id] = property_definition
@property @property
def required_properties(self): def required_properties(self):
return {property.id: property for property in self.properties.values() return {property.id: property for property in self.properties.values()
if property.is_required} if property.is_required}
def to_dict(self):
return {
'id': self.id,
'name': self.name,
'description': self.description,
'properties': {k: v.to_dict() for k, v in self.properties.items()}
}
class PropertyDefinition: class PropertyDefinition:
def __init__(self, attrs): def __init__(self, attrs):
@ -202,3 +230,12 @@ class PropertyDefinition:
@property @property
def has_multiple_values(self): def has_multiple_values(self):
return 'multiple' in self.flags return 'multiple' in self.flags
def to_dict(self):
return {
'id': self.id,
'name': self.name,
'description': self.description,
'flags': self.flags,
'labels': self.labels
}