def get_json_data(self, request, *args, **kwargs):
doc_class_name = kwargs.get('document_class')
doc_class_fields = DOCUMENT_FIELDS.get(doc_class_name)
if not doc_class_fields:
return []
return sorted([{
'code': f.field_code,
'name': f.name,
'type': f.field_type.value
} for f_name, f in doc_class_fields.items()], key=lambda k: k['name'])
def process(self, **kwargs):
self.log('Going to train field detector model based on the datasets stored in DB...')
document_class = kwargs.get('document_class')
task_id = kwargs.get('task_id')
if document_class:
TrainFieldDetectorModel.train_model_for_document_class(
document_class, task_id)
else:
for document_class, fields in DOCUMENT_FIELDS.items():
TrainFieldDetectorModel.train_model_for_document_class(
document_class, task_id)
class TrainFieldDetectorModelForm(forms.Form):
header = 'Train Field Detector Model'
_document_classes = list(DOCUMENT_FIELDS.keys())
_document_classes.sort()
_document_classes = [('', 'All')] + [(d, d) for d in _document_classes]
document_class = forms.ChoiceField(choices=_document_classes,
required=False)
def _post_clean(self):
super()._post_clean()
self.cleaned_data['module_name'] = MODULE_NAME
def process(self, **kwargs):
self.log_info('Going to train field detector model based on the datasets stored in DB...')
document_class = kwargs.get('document_class')
if document_class:
TrainFieldDetectorModel.local_train_model_for_document_class(self, document_class)
else:
train_model_for_document_class_args = []
for document_class, fields in DOCUMENT_FIELDS.items():
train_model_for_document_class_args.append((document_class,))
self.run_sub_tasks('Train Model For Each Document Class',
TrainFieldDetectorModel.train_model_for_document_class,
train_model_for_document_class_args)
def process(self, **kwargs):
self.log_info(
"Going to prepare datasets based on the pre-coded regexps and annotations"
"entered by users...")
document_class = kwargs.get('document_class')
document_ids = kwargs.get('document_ids')
if document_class:
self.build_sentences_to_fields_relations_dataset(
document_class, document_ids)
else:
for document_class, fields in DOCUMENT_FIELDS.items():
self.build_sentences_to_fields_relations_dataset(
document_class, document_ids)
def get_json_data(self, request, *args, **kwargs):
if request.method == 'PUT':
data = json.loads(request.body.decode('utf-8'))
field_code = data.get('field')
if not field_code:
return {
'error': None,
'value': data.get('value')
}
doc_class_name, field = field_code.split('__')
value = data['value']
doc_class_fields = DOCUMENT_FIELDS.get(doc_class_name)
if not doc_class_fields:
return {
'error': 'Unknown document class: {0}'.format(doc_class_name),
'value': None
}
field_config = doc_class_fields.get(field)
if not field_config:
return {
'error': 'Field {1} is not configured for document class: {0}'.format(
doc_class_name, field),
'value': None
}
doc_id = kwargs.get('document_id')
doc_class = field_config.document_class
doc = doc_class.objects.get(pk=doc_id)
if field_config:
val = field_config.set_value_from_selection(doc, value)
doc.save()
return {
'error': None,
'value': val
}
else:
return {
'error': 'No such field: {0}'.format(field),
'value': None
}
return None
def process(self, **kwargs):
self.task.log_info(
"Going to prepare datasets based on the pre-coded regexps and annotations"
"entered by users...")
document_class = kwargs.get('document_class')
document_ids = kwargs.get('document_ids')
task_count = 0
if document_class:
task_count += BuildFieldDetectorDataset.build_sentences_to_fields_relations_dataset(
document_class, document_ids, self.task)
else:
for document_class, fields in DOCUMENT_FIELDS.items():
task_count += BuildFieldDetectorDataset.build_sentences_to_fields_relations_dataset(
document_class, document_ids, self.task)
self.task.update_subtasks_total(task_count)
def _get_doc_class(doc_class_name: str):
field_configs = DOCUMENT_FIELDS.get(doc_class_name)
if not field_configs:
return None
for field_config in field_configs.values():
return field_config.document_class
