def scaling_prediction():
params = STANDARD_PARAMS
val_keys = ['time']
params['stop_condition'] = models.mostCorrect
OGs = [.01 * i for i in range(1, 21)]
#OGs = [.1,1,10]
base_gamma = 1
val_keys = ['Convergence Time']
dataset = Dataset(val_keys, params)
prediction = []
for i in rng(OGs):
print("\rIteration %s of %s " % (i + 1, len(OGs)), end="")
params['gamma'] = base_gamma * OGs[i]
params['alpha'] = base_gamma * OGs[i]
params['delta'] = base_gamma * OGs[i]
run_sweep.one_param_one_data_instance(MODEL, dataset, params)
orig = dataset.vals['Convergence Time']['avg']
prediction = [orig[0]] + [
orig[0] * (OGs[0] / OGs[i]) for i in range(1, len(OGs))
]
run_sweep.pickle_it(params, dataset)
plot_sweep.over_param_2ds(OGs,
'gamma = alpha = delta',
'Convergence Time', ['only one'],
'only one',
dataset,
params,
write_params_on_img=True,
prediction=prediction)
def over_param_2d():
params = STANDARD_PARAMS
val_keys = ['time']
params['stop_condition'] = models.mostCorrect
OGs = [.001 * i for i in range(1, 21)]
#OGs = [.1,1,10]
base_gamma = 1
val_keys = ['Convergence Time']
dataset = Dataset(val_keys, params)
for i in rng(OGs):
print("\rIteration %s of %s " % (i + 1, len(OGs)), end="")
params['delta'] = base_gamma * OGs[i]
params['gamma'] = .01 + (.01 - base_gamma * OGs[i]) / 2
params['alpha'] = .01 + (.01 - base_gamma * OGs[i]) / 2
run_sweep.one_param_one_data_instance(MODEL, dataset, params)
orig = dataset.vals['Convergence Time']['avg']
run_sweep.pickle_it(params, dataset)
plot_sweep.over_param_2ds(OGs, 'Delta', 'Convergence Time', ['only one'],
'only one', dataset, params)
def time_to_stop_3d():
params = STANDARD_PARAMS
base_gamma = .001
base_delta = 1
params['delta'] = base_delta
#scale = [i*.2 for i in range(1,9)]
scale = [i for i in range(1, 9)]
# ofc high gamma, low delta might not stop at all
# ofc high delta -> faster convergence
# ofc higher alpha -> faster convergence
# ofc higher gamma -> slower convergence (but seems less relv than the others)
# if gamma and alpha both occur via conjugation, set to same rates?
val_keys = ['Convergence Time', 'Log Convergence Time']
dataset = Dataset(val_keys, params)
for i in rng(scale):
params['gamma'] = base_gamma * scale[i]
for j in rng(scale):
print("\rIteration %s, %s of %s " %
(i + 1, j + 1, len(scale)),
end="")
params['alpha'] = base_delta * scale[j]
params['delta'] = base_delta * scale[j]
run_sweep.one_param_one_data_instance(MODEL, dataset, params)
run_sweep.pickle_it(params, dataset)
z, z_key = dataset.vals['Convergence Time']['avg'], 'Convergence Time'
x, x_key = dataset.params['gamma'], 'Gamma'
#y, y_key = dataset.params['alpha'], 'Alpha'
y, y_key = dataset.params['delta'], 'Alpha = Delta'
plot_sweep.tri(x,
y,
z,
x_key,
y_key,
z_key,
params,
write_params_on_img=True)
z, z_key = dataset.vals['Log Convergence Time'][
'avg'], 'Log Convergence Time'
plot_sweep.tri(x,
y,
z,
x_key,
y_key,
z_key,
params,
write_params_on_img=True)
def Ypercent_3d():
# appears to be depd on gate rxn rate, but not either duplication rates
params = STANDARD_PARAMS
output_gammas = [2**i for i in range(-4, 5)]
input_gammas = [2**i for i in range(-4, 5)]
betas = [2**i for i in range(-4, 5)]
#scale = [i for i in range(1,21)]
val_keys = ['steps', 'Percent Y0', 'time']
dataset = Dataset(val_keys, params)
for i in rng(betas):
#params['beta'] = betas[i]
params['input_gamma'] = input_gammas[i]
print("\rIteration %s of %s " % (i, len(betas)), end="")
for j in rng(output_gammas):
params['output_gamma'] = output_gammas[j]
#params['delta'] = betas[j]
run_sweep.one_param_one_data_instance(MODEL, dataset, params)
run_sweep.pickle_it(params, dataset)
z, z_key = dataset.vals['Percent Y0']['avg'], 'Percent Y0'
#x, x_key = dataset.vals['steps']['avg'], '# Steps'
x, x_key = dataset.params['input_gamma'], 'Input Duplication Rate'
#x, x_key = dataset.params['beta'], 'Gate Reaction Rate'
y, y_key = dataset.params['output_gamma'], 'Amplifier Duplication Rate'
#y, y_key = dataset.params['delta'], 'Amplifier Death Rate'
plot_sweep.tri(x,
y,
z,
x_key,
y_key,
z_key,
params,
write_params_on_img=True)
def over_time_2d():
params = STANDARD_PARAMS
val_keys = ['Percent Y0', 'time']
params['stop_condition'] = models.stopTime
l, h = .001, .1
params['gamma'] = .01
#gad = [[l,l],[l,h],[h,l],[h,h]]
#gad = [[.001,.1],[.01,.01],[.1,.001]] #THIS
gad = [[.01, .01, .1], [.01, .1, .01], [.1, .01, .01]] # AND THIS
#gad = [[.01,.01,.1],[.01,.01,.2],[.01,.01,.25],[.01,.01,.28],[.01,.01,.3]]
dsets = []
for i in rng(gad):
print("\rIteration %s of %s " % (i, len(gad)), end="")
if False: # for variable err rate
err_rate = gad[i][2]
total = 100
min_concen = int(err_rate * total)
maj_concen = total - min_concen
params['A1_t0'] = params['B1_t0'] = maj_concen
params['A0_t0'] = params['B0_t0'] = min_concen
params['alpha'] = gad[i][0]
params['delta'] = gad[i][1]
params['gamma'] = gad[i][2]
dataset = Dataset(val_keys, params)
run_sweep.one_param_mult_data_instances(MODEL, dataset, params)
dsets += [dataset]
run_sweep.pickle_it(params, dataset)
legend_str = r'[$\alpha$, $\delta$, $\gamma$]'
plot_sweep.over_time_2ds('time', 'Percent Y0', gad, legend_str, dsets,
params)
def time_to_stop_3d():
params = STANDARD_PARAMS
base_gamma = 1
base_delta = 1
scale = [i * 100 for i in range(1, 11)]
val_keys = ['Convergence Time', 'Log Convergence Time']
dataset = Dataset(val_keys, params)
for i in rng(scale):
params['IA'] = base_gamma * scale[i]
for j in rng(scale):
print("\rIteration %s, %s of %s " %
(i + 1, j + 1, len(scale)),
end="")
params['IB'] = base_delta * scale[j]
run_sweep.one_param_one_data_instance(MODEL, dataset, params)
run_sweep.pickle_it(params, dataset)
z, z_key = dataset.vals['Convergence Time']['avg'], 'Convergence Time'
x, x_key = dataset.params['IA'], 'A_0'
#y, y_key = dataset.params['alpha'], 'Alpha'
y, y_key = dataset.params['IB'], 'B_0'
plot_sweep.tri(x,
y,
z,
x_key,
y_key,
z_key,
params,
write_params_on_img=True)
z, z_key = dataset.vals['Log Convergence Time'][
'avg'], 'Log Convergence Time'
plot_sweep.tri(x,
y,
z,
x_key,
y_key,
z_key,
params,
write_params_on_img=True)
z, z_key = dataset.vals['Convergence Time'][
'min_sd'], 'Convergence Time within 3 Std Devs'
plot_sweep.tri(x,
y,
z,
x_key,
y_key,
z_key,
params,
write_params_on_img=True)
z, z_key = dataset.vals['Log Convergence Time'][
'min_sd'], 'Log Convergence Time within 3 Std Devs'
plot_sweep.tri(x,
y,
z,
x_key,
y_key,
z_key,
params,
write_params_on_img=True)