def load_session(directory: str, session_name: str) -> pd.DataFrame:
session = {}
experiment_fname = directory + "session" + session_name + ".mat"
params_fname = directory + "sessionParams" + session_name + ".mat"
experiment_dict = strip_mat_metadata(load_mat(experiment_fname, False))
for k, v in experiment_dict.items():
if "neuron" in k:
df = pd.DataFrame(v).transpose()
session[k] = df
params_data = strip_mat_metadata(load_mat(params_fname, False))["params"]
session["params"] = pd.DataFrame(params_data, columns=["StimOn", "SignalOn", "LeverUp", "Coh1", "Coh2"]).dropna()
session["params"]["LeverDelay"] = get_session_lever_delay(session["params"])
session["params"]["LeverSuccess"] = get_session_lever_success(session["params"]["LeverDelay"])
session["params"]["Session"] = pd.Series([session_name] * len(session["params"]))
session["params"].index.name = "Trial"
return session
def hierarchical_mat_to_dir(mat: str, out_prefix = "", show_debug = False, fmt="%.2f") -> bool:
data = strip_mat_metadata(load_mat(mat, show_debug))
if not out_prefix == "":
if out_prefix[-1] == "/":
if not os.path.exists(out_prefix):
os.makedirs(out_prefix)
for k, v in data.items():
df = pd.DataFrame(v)
df.to_csv(out_prefix + k + ".csv")
def hierarchical_mat_to_dir(mat: str,
out_prefix="",
show_debug=False,
fmt="%.2f") -> bool:
data = strip_mat_metadata(load_mat(mat, show_debug))
if not out_prefix == "":
if out_prefix[-1] == "/":
if not os.path.exists(out_prefix):
os.makedirs(out_prefix)
for k, v in data.items():
df = pd.DataFrame(v)
df.to_csv(out_prefix + k + ".csv")
def load_session_corr_signal(directory: str, session_name: str) -> pd.DataFrame:
scs = {}
session_fname = directory + "sessionCorrSignal" + session_name + ".mat"
session_dict = strip_mat_metadata(load_mat(session_fname))
for k, v in session_dict.items():
scs[k] = pd.DataFrame(v)
for col in scs["corrC"].columns:
scs["time-" + str(col)] = pd.DataFrame(
{"correlation": scs["corrC"][col], "mutual_information": scs["MIC"][col],
"label": [1] * len(scs["corrC"][col])})
scs["time-" + str(col)] = scs["time-" + str(col)].append(
pd.DataFrame({"correlation": scs["corrF"][col], "mutual_information": scs["MIF"][col],
"label": [0] * len(scs["corrF"][col])})).reset_index(drop=True)
return scs
def preprocess_directory_of_raw_mats(directory: str) -> None:
for f in os.listdir(directory):
if f[-3:] == "mat":
if "Params" in f:
out_prefix = f[:-4] + "/raw/"
out_prefix = out_prefix.replace("Params", "")
else:
out_prefix = f[:-4] + "/raw/"
data = strip_mat_metadata(load_mat(f, False))
if not os.path.exists(out_prefix):
os.makedirs(out_prefix)
for k, v in data.items():
df = pd.DataFrame(v)
if "neuron" in k:
df.transpose()
df.to_csv(out_prefix + k + ".csv")