def get_face_data(person_data, emotion, frame_index):
filepath = housekeeping.build_filepath({
'person_data': person_data,
'emotion': emotion,
'frame_index': frame_index,
'data_type': 'face'
})
with open(filepath, 'r') as file:
data = [line[:-1].split(' ')[1:-1] for line in file.readlines()]
data = [{"x": float(pair[0]), "y": float(pair[1])} for pair in data]
return data
def get_embedding_data(specs):
specs['data_type'] = 'dissimilarity'
filepath = housekeeping.build_filepath(specs)
indices = housekeeping.get_emotion_indices(
specs.get('emotions'),
F001_key if specs.get('person_data') == 'F001' else M001_key)
print(indices)
dissimilarities = housekeeping.filter_data(
housekeeping.read_matrix_file(filepath, specs.get('non_metric')),
indices)
print(len(dissimilarities))
if specs.get('return_dissimilarity'):
return dissimilarities
embedding_data = housekeeping.embed({
'perplexity':
int(specs.get('perplexity')),
'key':
F001_key if specs.get('person_data') == 'F001' else M001_key,
'focal_emotion':
specs.get('focal_emotion'),
'frame_indices':
specs.get('frame_indices'),
'embedding_type':
specs.get('embedding_type'),
'dimension':
1 if specs.get('embedding_type') == 'reld' else 2,
'emotions':
specs.get('emotions'),
'dissimilarities':
dissimilarities
})
if specs.get('embedding_type') != 'reld':
embedding_data = embedding_data.fit_transform(
np.asmatrix(dissimilarities))
return housekeeping.prepare_data(
specs.get('emotions'), specs.get('embedding_type'), embedding_data,
1 if specs.get('embedding_type') == 'reld' else 2,
F001_key if specs.get('person_data') == 'F001' else M001_key)
def get_persistence_diagram(section_list, person_data, emotion, frameNumber,
nonMetric, twoD):
filepaths = housekeeping.build_filepath({
'section_list': section_list,
'person_data': person_data,
'emotion': emotion,
'frame_index': frameNumber,
'is_two_dimensional': twoD,
'non_metric': nonMetric,
'data_type': 'persistence'
})
return {
'selection_1':
housekeeping.read_persistence_file(filepaths['selection_1']),
'selection_2':
housekeeping.read_persistence_file(filepaths['selection_2']),
'selection_3':
housekeeping.read_persistence_file(filepaths['selection_3'])
}
def get_shepard_diagram():
spearman_file = housekeeping.build_filepath({
'person_data':
request.args.get('personData'),
'difference_metric':
request.args.get('differenceMetric'),
'non_metric':
request.args.get('nonMetric'),
'section_list':
housekeeping.get_selected_subsections(request.args),
'is_two_dimensional':
False if request.args.get('twoD') == '3d' else True,
'data_type':
'spearman',
'emotion':
int(request.args.get('focalEmotion')),
'embedding_type':
request.args.get('embeddingType'),
'emotions':
list(
filter(
lambda e: True if e != None else False,
[e if request.args.get(e) == '1' else None for e in emotions]))
})
if path.exists(spearman_file):
with open(spearman_file, 'r') as file:
spearmans = json.load(file)
return json.dumps(spearmans)
emotion_file = housekeeping.build_filepath({
'person_data':
request.args.get('personData'),
'difference_metric':
request.args.get('differenceMetric'),
'non_metric':
request.args.get('nonMetric'),
'section_list':
housekeeping.get_selected_subsections(request.args),
'is_two_dimensional':
False if request.args.get('twoD') == '3d' else True,
'data_type':
'embedding',
'emotion':
int(request.args.get('focalEmotion')),
'embedding_type':
request.args.get('embeddingType'),
'emotions': [
emotions[i] for i in range(len(emotions))
if request.args.get(emotions[i]) == '1'
]
})
before_file = housekeeping.build_filepath({
'person_data':
request.args.get('personData'),
'difference_metric':
request.args.get('differenceMetric'),
'non_metric':
request.args.get('nonMetric'),
'section_list':
housekeeping.get_selected_subsections(request.args),
'is_two_dimensional':
False if request.args.get('twoD') == '3d' else True,
'data_type':
'diss_before',
'emotion':
int(request.args.get('focalEmotion')),
'embedding_type':
request.args.get('embeddingType'),
'emotions': [
emotions[i] for i in range(len(emotions))
if request.args.get(emotions[i]) == '1'
]
})
with open(emotion_file, 'r') as file:
embedded_data = [
(m['x'], m['y']) for l in json.load(file) # iterate
if l != None # remove None values
for m in l # flatten
]
matrix_before = getData.get_embedding_data({
'person_data':
request.args.get('personData'),
'section_list':
housekeeping.get_selected_subsections(request.args),
'difference_metric':
request.args.get('differenceMetric'),
'embedding_type':
request.args.get('embeddingType'),
'emotions': [
emotions[i] for i in range(len(emotions))
if request.args.get(emotions[i]) == '1'
],
'focal_emotion':
int(request.args.get('focalEmotion')),
'frame_indices':
[request.args.get(slidevalues[i]) for i in range(len(slidevalues))],
'non_metric':
request.args.get('nonMetric'),
'perplexity':
request.args.get('perplexity'),
'return_dissimilarity':
True,
'is_two_dimensional':
False if request.args.get('twoD') == '3d' else True
})
matrix_after = [[distance.euclidean(d, e) for d in embedded_data]
for e in embedded_data]
spearmans = spearmanr([l for line in matrix_before for l in line],
[l for line in matrix_after for l in line])
with open(spearman_file, 'w') as file:
file.write(json.dumps(spearmans))
return json.dumps(spearmans)
def get_embedding_data():
emotionFile = housekeeping.build_filepath({
'person_data':
request.args.get('personData'),
'difference_metric':
request.args.get('differenceMetric'),
'non_metric':
request.args.get('nonMetric'),
'section_list':
housekeeping.get_selected_subsections(request.args),
'is_two_dimensional':
False if request.args.get('twoD') == '3d' else True,
'data_type':
'embedding',
'frame_indices':
[request.args.get(slidevalues[i]) for i in range(len(slidevalues))],
'focal_emotion':
int(request.args.get('focalEmotion')),
'emotion':
int(request.args.get('focalEmotion')),
'embedding_type':
request.args.get('embeddingType'),
'emotions':
list(
filter(
lambda e: True if e != None else False,
[e if request.args.get(e) == '1' else None for e in emotions]))
})
if not path.exists(emotionFile):
data = getData.get_embedding_data({
'person_data':
request.args.get('personData'),
'section_list':
housekeeping.get_selected_subsections(request.args),
'difference_metric':
request.args.get('differenceMetric'),
'embedding_type':
request.args.get('embeddingType'),
'emotions':
[e if request.args.get(e) == '1' else None for e in emotions],
'focal_emotion':
int(request.args.get('focalEmotion')),
'frame_indices': [
request.args.get(slidevalues[i])
for i in range(len(slidevalues))
],
'non_metric':
request.args.get('nonMetric'),
'perplexity':
request.args.get('perplexity'),
'return_dissimilarity':
False,
'is_two_dimensional':
False if request.args.get('twoD') == '3d' else True
})
with open(emotionFile, 'w') as file:
file.write(json.dumps(data))
else:
with open(emotionFile, 'r') as file:
data = json.load(file)
return json.dumps(data)