Page Comparison

Even when an App runs locally, the parameters of the functions used are stored in a config file in JSON format (config.json.example). But this file doesn’t support all the Python types (for instance the numpy.nd.arrays and the tuples).

So we have to convert these values in the Python file.

Case of numpy nd.array

In the config.json.example, it is not possible to enter directly a numpy.nd.array, so we enter a list of floats instead:

{
"param_origin": [0., 0., 0.] 
}

And in the Python file this list will be converted into a numpy.nd.array:

if isinstance(config['param_origin'], list):
    config['param_origin'] = np.array(config['param_origin'])

The part when we test if config['param_origin'] is a list is here to make this condition only used when the App is run locally.

Case of a slice

Here again it is impossible to enter a slice in the config.json.example, so in this file we enter:

{
"param_picks_by_channel_indices": "0, 10, 2"
}

And in the Python file, this string will be converted into a slice. First we convert it into a list and then into a slice.

For instance:

if isinstance(picks, str) and picks.find(",") != -1 and picks.find("[") == -1 and picks is not None:
    picks = list(map(int, picks.split(', ')))
    if len(picks) == 2:
        config['param_picks_by_channel_indices'] = slice(picks[0], picks[1])
    elif len(picks) == 3:
        config['param_picks_by_channel_indices'] = slice(picks[0], picks[1], picks[2])
    else:
        value_error_message = f"If you want to select channels using a slice, you must give two or three elements."
        raise ValueError(value_error_message)

In this case, these lines of code are used to convert the string to a slice when the App runs locally but also on Brainlife.

Case of a tuple

It is also impossible to enter a tuple in a json file, so instead we enter:

{
"param_baseline": [null, 0]
}

And then we convert this list into a tuple in the Python file:

if config['param_baseline'] is not None:
    config['param_baseline'] = tuple(config['param_baseline'])

Case of a dictionary

It is possible to enter a dictionary in the config.json.example using a nested structure:

{    
"param_reject": {
                 "grad": 4000e-13,
                 "mag": 4e-12,
                 "eeg": 40e-6
                 }
}

When the config.json.example is parsed by the Python code, config['param_reject'] is directly a dictionary, so no conversion is required.

Case of a pandas.Dataframe

Some metadata can be stored in a pandas.Dataframe and given to a MNE function (for instance mne.Epochs). So either, the information is stored into a .csv and converted into a pandas.Dataframe using:

if isinstance(config['param_metadata'], str):
    config['param_metadata'] = pd.read_csv(config['param_metadata'])

Or the user enters a dictionary in the config.json.example and this dictionary will be then converted into a pandas.Dataframe using:

elif isinstance(config['param_metadata'], dict):
    config['param_metadata'] = pd.DataFrame(list(config['param_metadata'].items()))