def checkerboard_svc(data,
metadata,
stimulus_list,
lab_notebook,
plot_directory,
nsamples,
n_draws=30,
px_per_deg=10.453):
sizes = glia.get_stimulus_parameters(stimulus_list, "CHECKERBOARD", 'size')
name = metadata["name"]
if name == 'checkerboard-contrast':
training_data = glia.bin_100ms(data["training_data"])
validation_data = glia.bin_100ms(data["validation_data"])
training_target = data["training_target"]
validation_target = data["validation_target"]
conditions = glia.get_checkerboard_contrasts(stimulus_list)
condition_name = "contrast"
elif name == "checkerboard-durations":
training_data = glia.bin_100ms(data["training_data"])
validation_data = glia.bin_100ms(data["validation_data"])
training_target = data["training_target"]
validation_target = data["validation_target"]
conditions = glia.get_stimulus_parameters(stimulus_list,
"CHECKERBOARD", 'lifespan')
condition_name = "durations"
elif name == "checkerboard":
training_100ms = glia.bin_100ms(data["training_data"])[0]
training_sum = glia.bin_sum(data["training_data"])[0]
training_data = [training_100ms, training_sum]
validation_100ms = glia.bin_100ms(data["validation_data"])[0]
validation_sum = glia.bin_sum(data["validation_data"])[0]
validation_data = [validation_100ms, validation_sum]
tt = data["training_target"][0]
training_target = [tt, tt]
vt = data["validation_target"][0]
validation_target = [vt, vt]
conditions = ['100ms bins', 'spike count']
condition_name = None
else:
raise (ValueError(f'Unknown experiment {metadata["name"]}'))
if nsamples > 0:
plot_diff_nsamples(data, stimulus_list, plot_directory, "checkerboard",
sizes, conditions, condition_name)
else:
acuity(training_data,
training_target,
validation_data,
validation_target,
stimulus_list,
plot_directory,
"checkerboard",
sizes,
conditions,
condition_name,
n_draws,
units="cpd",
px_per_deg=px_per_deg)
def tiled_letter_svc(data, metadata, stimulus_list, lab_notebook,
plot_directory, nsamples):
print("Classifying Letters - warning not using latest acuity function")
# TODO
sizes = glia.get_stimulus_parameters(stimulus_list, "TILED_LETTER", 'size')
name = metadata["name"]
# n_sizes, n_training, n_steps, n_x, n_y, n_units = data["training_data"].shape
logger.debug(data["training_data"].shape)
# add nconditions dim
training_100ms = glia.bin_100ms(np.expand_dims(data["training_data"], 0))
validation_100ms = glia.bin_100ms(
np.expand_dims(data["validation_data"], 0))
logger.debug(f'training_100ms shape {training_100ms.shape}')
logger.debug(f'sizes {sizes}')
for i, size in enumerate(sizes):
print(f'SVC for size {size}')
# note: no expand dims, hardcoded 1 ncondition
training_target = data["training_target"][i]
validation_target = data["validation_target"][i]
logger.debug(np.size(training_target))
svr = svm.SVC()
parameters = {'C': [1, 10, 100, 1000], 'gamma': [0.001, 0.0001]},
clf = GridSearchCV(svr, parameters, n_jobs=12)
report, confusion = glia.classifier_helper(
clf, (training_100ms[0, i], training_target),
(validation_100ms[0, i], validation_target))
with open(f"{plot_directory}/letter-{size}.txt", "w") as f:
f.write(report + '\n')
f.write(str(confusion))
def grating_svc(data,
metadata,
stimulus_list,
lab_notebook,
plot_directory,
nsamples,
n_draws=30,
sinusoid=False,
px_per_deg=10.453):
if sinusoid:
stimulus_type = "SINUSOIDAL_GRATING"
plot_name = "Sinusoidal grating"
else:
stimulus_type = 'GRATING'
plot_name = "grating"
sizes = glia.get_stimulus_parameters(stimulus_list, stimulus_type, "width")
if metadata["name"] in \
['grating-contrast', 'grating-sinusoidal-contrast']:
training_data = glia.bin_100ms(data["training_data"])
validation_data = glia.bin_100ms(data["validation_data"])
training_target = data["training_target"]
validation_target = data["validation_target"]
conditions = glia.get_grating_contrasts(stimulus_list, stimulus_type)
condition_name = "contrast"
elif metadata["name"] in \
["grating-durations", "grating-sinusoidal-durations"]:
training_data = glia.bin_100ms(data["training_data"])
validation_data = glia.bin_100ms(data["validation_data"])
training_target = data["training_target"]
validation_target = data["validation_target"]
conditions = glia.get_stimulus_parameters(stimulus_list, stimulus_type,
'lifespan')
condition_name = "durations"
elif metadata["name"] in \
["grating-speeds", "grating-sinusoidal-speeds"]:
training_data = glia.bin_100ms(data["training_data"])
validation_data = glia.bin_100ms(data["validation_data"])
training_target = data["training_target"]
validation_target = data["validation_target"]
conditions = glia.get_stimulus_parameters(stimulus_list, stimulus_type,
'speed')
condition_name = "speeds"
elif metadata["name"] in \
["grating", "grating-sinusoidal"]:
training_100ms = glia.bin_100ms(data["training_data"])[0]
training_sum = glia.bin_sum(data["training_data"])[0]
training_data = [training_100ms, training_sum]
validation_100ms = glia.bin_100ms(data["validation_data"])[0]
validation_sum = glia.bin_sum(data["validation_data"])[0]
validation_data = [validation_100ms, validation_sum]
tt = data["training_target"][0]
training_target = [tt, tt]
vt = data["validation_target"][0]
validation_target = [vt, vt]
conditions = ['100ms bins', 'spike count']
condition_name = None
else:
raise (ValueError(f'Unknown experiment {metadata["name"]}'))
if nsamples > 0:
plot_diff_nsamples(data, stimulus_list, plot_directory, plot_name,
sizes, conditions, condition_name)
else:
acuity(training_data,
training_target,
validation_data,
validation_target,
stimulus_list,
plot_directory,
plot_name,
sizes,
conditions,
condition_name,
n_draws,
units="cpd",
px_per_deg=px_per_deg)