def _send_output_signals(self, embeddings):
embedded_images, skipped_images, num_skipped =\
ImageEmbedder.prepare_output_data(self._input_data, embeddings)
self.Outputs.embeddings.send(embedded_images)
self.Outputs.skipped_images.send(skipped_images)
if num_skipped is not 0:
self.input_data_info.setText(
"Data with {:d} instances, {:d} images skipped.".format(
len(self._input_data), num_skipped))
def _send_output_signals(self, embeddings):
embedded_images, skipped_images, num_skipped =\
ImageEmbedder.prepare_output_data(self._input_data, embeddings)
self.Outputs.embeddings.send(embedded_images)
self.Outputs.skipped_images.send(skipped_images)
if num_skipped is not 0:
self.input_data_info.setText(
"Data with {:d} instances, {:d} images skipped.".format(
len(self._input_data), num_skipped))
def _send_output_signals(self, embeddings):
embedded_images, skipped_images, num_skipped =\
ImageEmbedder.prepare_output_data(self._input_data, embeddings)
self.send(_Output.SKIPPED_IMAGES, skipped_images)
self.send(_Output.EMBEDDINGS, embedded_images)
if num_skipped is not 0:
self.input_data_info.setText(
"Data with {:d} instances, {:d} images skipped.".format(
len(self._input_data), num_skipped))
mycol.update_one({"client_id":upload.cid2},{"$set": { "metadata.features": emb_up[0]}})
#retrieving the just stored featues again
myquery = {"client_id":upload.cid2}
mydoc = mycol.find(myquery)
for x in mydoc:
embeddings[u] = x["metadata"]
data_2 = data
#getting embeddings for each image as a key value pair
values = []
for dictionary in embeddings:
values.extend([v for k, v in dictionary.items()])
data2 = ImageEmbedder.prepare_output_data(data_2, values)
out_data = data2[0]
#selection of distance algorithm by user
d_algo = upload.distance_algo
# computing distance among images
from Orange.distance import Cosine, Euclidean
dist = data2[0]
if(d_algo == 'Cosine'):
dist_matrix = Cosine(dist)
else:
dist_matrix = Euclidean(dist)