def post(self, *args, **kwargs):
response = self.request.data['response']
label_mapping = self.request.data['label_mapping']
project = get_object_or_404(Project, pk=self.kwargs['project_id'])
task = TaskFactory.create(project.project_type)
labels = task.label_collection(response)
post_processor = PostProcessor(label_mapping)
labels = post_processor.transform(labels)
return Response(labels.dict(), status=status.HTTP_200_OK)
def post(self, *args, **kwargs):
response = self.request.data["response"]
task_type = self.request.data["task_type"]
label_mapping = self.request.data["label_mapping"]
label_collection = get_label_collection(task_type)
labels = label_collection(response)
post_processor = PostProcessor(label_mapping)
labels = post_processor.transform(labels)
return Response(labels.dict(), status=status.HTTP_200_OK)
def test_postprocessor():
labels = [{'label': 'PERSON'}, {'label': 'ORG'}, {'label': 'Facility'}]
labels = ClassificationLabels(labels)
mapping = {'Facility': 'ORG'}
processor = PostProcessor(mapping=mapping)
labels = processor.transform(labels).dict()
expected = [
{
'label': 'ORG'
},
]
assert labels == expected
def execute_pipeline(text: str, project_type: str, model_name: str,
model_attrs: dict, template: str, label_mapping: dict):
task = TaskFactory.create(project_type)
model = RequestModelFactory.create(model_name=model_name,
attributes=model_attrs)
template = MappingTemplate(label_collection=task.label_collection,
template=template)
post_processor = PostProcessor(label_mapping)
labels = pipeline(text=text,
request_model=model,
mapping_template=template,
post_processing=post_processor)
return labels.dict()
def execute_pipeline(data: str, config: AutoLabelingConfig):
label_collection = get_label_collection(config.task_type)
model = RequestModelFactory.create(model_name=config.model_name,
attributes=config.model_attrs)
template = MappingTemplate(label_collection=label_collection,
template=config.template)
post_processor = PostProcessor(config.label_mapping)
labels = pipeline(text=data,
request_model=model,
mapping_template=template,
post_processing=post_processor)
labels = create_labels(config.task_type, labels)
return labels
def test_gcp_label_detection_pipeline(data_path, cassettes_path):
with vcr.use_cassette(
str(cassettes_path / 'pipeline_gcp_label_detection.yaml'),
mode='once',
filter_query_parameters=['key']
):
model = GCPImageLabelDetectionRequestModel(key=os.environ.get('API_KEY_GCP', ''))
template = GCPImageLabelDetectionTemplate()
filepath = data_path / 'images/1500x500.jpeg'
post_processor = PostProcessor({})
labels = pipeline(
text=filepath,
request_model=model,
mapping_template=template,
post_processing=post_processor
)
labels = labels.dict()
assert isinstance(labels, list)
assert len(labels) == 1
assert 'label' in labels[0]
def test_amazon_pipeline(cassettes_path):
with vcr.use_cassette(str(cassettes_path /
'amazon_comprehend_sentiment.yaml'),
mode='once',
filter_headers=['authorization']):
model = AmazonComprehendSentimentRequestModel(
aws_access_key=os.environ.get('AWS_ACCESS_KEY', ''),
aws_secret_access_key=os.environ.get('AWS_SECRET_ACCESS_KEY', ''),
region_name='us-east-1',
language_code='en')
template = AmazonComprehendSentimentTemplate()
post_processor = PostProcessor({})
labels = pipeline(text='I am very sad.',
request_model=model,
mapping_template=template,
post_processing=post_processor)
labels = labels.dict()
assert isinstance(labels, list)
assert len(labels) == 1
assert 'label' in labels[0]
def test_load():
expected = {'Facility': 'ORG'}
processor = PostProcessor.load(mapping=expected)
actual = processor.mapping
assert actual == expected
def test_to_dict():
expected = {'Facility': 'ORG'}
processor = PostProcessor(mapping=expected)
actual = processor.to_dict()
assert actual == expected
