Description of the problem
When anonymizing a raw object with func anonymize(), the subject info and meas_date are still accessible through raw._raw_extras attribute as shown in the below example.
Regarding subject info, it would be easier in my opinion to remove field "subject_info" from both raw.info and *raw._raw_extras (or simply empty unecessary fields, keeping "his_id" only for instance) as it is not expected to be able to retrieve the information afterwards by definition.
Steps to reproduce
import mne
data_path = mne.datasets.testing.data_path(download=False)
fpath = data_path / "EDF" / "test_edf_stim_resamp.edf"
raw = mne.io.read_raw_edf(fpath, verbose=False)
print("Subject info before anonymization")
print("\t", raw.info["subject_info"])
print("\t", raw.info["meas_date"])
print("\t", raw._raw_extras[0]["subject_info"])
print("\t", raw._raw_extras[0]["meas_date"])
raw.anonymize()
print("\t", "\nSubject info after anonymization")
print("\t", raw.info["subject_info"])
print("\t", raw.info["meas_date"])
print("\t", raw._raw_extras[0]["subject_info"])
print("\t", raw._raw_extras[0]["meas_date"])
print(raw._raw_extras[0]["subject_info"])Expected results
Subject info before anonymization
{'his_id': 'X', 'sex': 1, 'last_name': 'X', 'birthday': (1967, 10, 9)}
2016-02-14 10:33:35+00:00
{'id': 'X', 'sex': 'M', 'birthday': datetime.datetime(1967, 10, 9, 0, 0), 'name': 'X'}
2016-02-14 10:33:35+00:00
Subject info after anonymization
{'his_id': '0', 'sex': 0, 'last_name': 'mne_anonymize', 'birthday': (1951, 8, 25)}
2000-01-01 00:00:00+00:00
{'id': '0', 'sex': 'unknown', 'birthday': datetime.datetime(1951, 8, 25, 0, 0), 'name': 'mne_anonymize'}
2000-01-01 00:00:00+00:00
or alternatively
Subject info before anonymization
{'his_id': 'X', 'sex': 1, 'last_name': 'X', 'birthday': (1967, 10, 9)}
2016-02-14 10:33:35+00:00
{'id': 'X', 'sex': 'M', 'birthday': datetime.datetime(1967, 10, 9, 0, 0), 'name': 'X'}
2016-02-14 10:33:35+00:00
Subject info after anonymization
{'his_id': 'X'}
2000-01-01 00:00:00+00:00
{}
2000-01-01 00:00:00+00:00
Actual results
Subject info before anonymization
{'his_id': 'X', 'sex': 1, 'last_name': 'X', 'birthday': (1967, 10, 9)}
2016-02-14 10:33:35+00:00
{'id': 'X', 'sex': 'M', 'birthday': datetime.datetime(1967, 10, 9, 0, 0), 'name': 'X'}
2016-02-14 10:33:35+00:00
Subject info after anonymization
{'his_id': '0', 'sex': 0, 'last_name': 'mne_anonymize', 'birthday': (1951, 8, 25)}
2000-01-01 00:00:00+00:00
{'id': 'X', 'sex': 'M', 'birthday': datetime.datetime(1967, 10, 9, 0, 0), 'name': 'X'}
2016-02-14 10:33:35+00:00
Additional information
The above code uses the latest release (mne==1.5.1).
Description of the problem
When anonymizing a
rawobject with funcanonymize(), the subject info andmeas_dateare still accessible throughraw._raw_extrasattribute as shown in the below example.Regarding subject info, it would be easier in my opinion to remove field "subject_info" from both
raw.infoand*raw._raw_extras(or simply empty unecessary fields, keeping "his_id" only for instance) as it is not expected to be able to retrieve the information afterwards by definition.Steps to reproduce
Expected results
or alternatively
Actual results
Additional information
The above code uses the latest release (
mne==1.5.1).