def do_sample(data_path, output_dir):
"""Generate MCMC samples given a user input directory.
This function creates a new directory in output_dir with a name starting
with "maud_output". It first copies the directory at data_path into the new
this directory at new_dir/user_input, then runs the running_stan.sample
function to write samples in new_dir/samples. Finally it prints the results
of cmdstanpy's diagnose and summary methods.
"""
mi = load_maud_input(data_path)
now = datetime.now().strftime("%Y%m%d%H%M%S")
output_name = f"maud_output-{mi.config.name}-{now}"
output_path = os.path.join(output_dir, output_name)
samples_path = os.path.join(output_path, "samples")
ui_dir = os.path.join(output_path, "user_input")
print("Creating output directory: " + output_path)
os.mkdir(output_path)
os.mkdir(samples_path)
print(f"Copying user input from {data_path} to {ui_dir}")
shutil.copytree(data_path, ui_dir)
stanfit = sample(mi, samples_path)
print(stanfit.diagnose())
print(stanfit.summary())
idata = get_idata(stanfit.runset.csv_files, mi, "train")
idata.to_netcdf(os.path.join(output_path, "idata.nc"))
return output_path
def do_predict(data_path: str):
"""Generate MCMC samples given a Maud output folder at train_path.
This function creates a new directory in output_dir with a name starting
with "maud-predict-output". It first copies the testing directory at
train_path into the new this directory at new_dir/user_input, then runs the
running_stan.predict_out_of_sample function to write samples in
new_dir/oos_samples.
The trained output is stored in the new_dir/trained_samples folder along
with the user input required to generate the trained samples.
"""
idata_train = az.from_netcdf(os.path.join(data_path, "idata.nc"))
mi = load_maud_input(os.path.join(data_path, "user_input"))
now = datetime.now().strftime("%Y%m%d%H%M%S")
output_name = f"maud-predict_output-{mi.config.name}-{now}"
output_path = os.path.join(data_path, output_name)
test_samples_path = os.path.join(output_path, "test_samples")
print("Creating output directory: " + output_path)
os.mkdir(output_path)
os.mkdir(test_samples_path)
idata_predict = predict(mi, output_path, idata_train)
# delete attrs hack to make netcdf save work:
# https://github.com/arviz-devs/arviz/issues/1554
idata_predict.sample_stats.attrs = {} # type: ignore
idata_predict.posterior.attrs = {} # type: ignore
idata_predict.to_netcdf(os.path.join(output_path, "idata_predict.nc"))
def do_simulate(data_path, output_dir, n):
"""Generate draws from the initial values."""
mi = load_maud_input(data_path=data_path)
now = datetime.now().strftime("%Y%m%d%H%M%S")
output_name = f"maud_output_sim-{mi.config.name}-{now}"
output_path = os.path.join(output_dir, output_name)
samples_path = os.path.join(output_path, "samples")
ui_dir = os.path.join(output_path, "user_input")
print("Creating output directory: " + output_path)
os.mkdir(output_path)
os.mkdir(samples_path)
print(f"Copying user input from {data_path} to {ui_dir}")
stanfit = simulate(mi, samples_path, n)
idata = get_idata(stanfit.runset.csv_files, mi, "train")
idata.to_netcdf(os.path.join(output_path, "idata.nc"))
print("\n\nSimulated concentrations:")
print(idata.posterior["conc"].mean(
dim=["chain", "draw"]).to_series().unstack().T)
print("\n\nSimulated fluxes:")
print(idata.posterior["flux"].mean(
dim=["chain", "draw"]).to_series().unstack().T)
print("\n\nSimulated enzyme concentrations:")
print(idata.posterior["conc_enzyme"].mean(
dim=["chain", "draw"]).to_series().unstack().T)
print("\n\nSimulated reaction delta Gs:")
print(idata.posterior["dgrs"].mean(dim=["chain", "draw"]).to_series())
print("\n\nSimulated measurements:")
print(idata.posterior["yconc_sim"].mean(dim=["chain", "draw"]).to_series())
print(idata.posterior["yflux_sim"].mean(dim=["chain", "draw"]).to_series())
print("\n\nSimulated log likelihoods:")
print(idata.posterior["log_lik_conc"].mean(
dim=["chain", "draw"]).to_series())
print(idata.posterior["log_lik_flux"].mean(
dim=["chain", "draw"]).to_series())
print("\n\nSimulated allostery terms:")
print(idata.posterior["allostery"].mean(
dim=["chain", "draw"]).to_series().unstack().T)
print("\n\nSimulated reversibility terms:")
print(idata.posterior["reversibility"].mean(
dim=["chain", "draw"]).to_series().unstack().T)
print("\n\nSimulated saturation terms:")
print(idata.posterior["saturation"].mean(
dim=["chain", "draw"]).to_series().unstack().T)
if mi.kinetic_model.phosphorylations is not None:
print("\n\nSimulated phosphorylation terms:")
print(idata.posterior["phosphorylation"].mean(
dim=["chain", "draw"]).to_series().unstack().T)
print("\n\nSimulated membrane potential:")
print(idata.posterior["psi"].mean(dim=["chain", "draw"]).to_series().T)
return output_path
def test_load_maud_input():
"""Test that the function load_maud_input behaves as expected."""
expected_var_ids = {
"dgf": [["M1", "M2"]],
"conc_unbalanced": [["condition1", "condition2"], ["M1_e", "M2_e"]],
"conc_pme": [["condition1", "condition2"], []],
"dissociation_constant": [["r1_M2_c", "r2_M1_c"]],
}
mi = load_maud_input(data_path=LINEAR_PATH)
r1 = next(r for r in mi.kinetic_model.reactions if r.id == "r1")
assert r1.stoichiometry == {"M1_e": -1, "M1_c": 1}
assert "r1_r1" in mi.priors.kcat.location.index
for var_name, expected in expected_var_ids.items():
assert var_name in mi.stan_variable_set.__dataclass_fields__
assert getattr(mi.stan_variable_set, var_name).ids == expected
actual_stan_input = mi.stan_input_train.stan_input_dict
with open(EXPECTED_STAN_INPUT_PATH, "r") as f:
expected_stan_input = json.load(f)
assert set(actual_stan_input.keys()) == set(expected_stan_input.keys())
for k, v in actual_stan_input.items():
actual = v.tolist() if isinstance(v, np.ndarray) else v
expected = expected_stan_input[k]
assert_equal(actual, expected, err_msg=f"{k} different from expected.")