Minor improvements of manual_correction.py

[plbenveniste]

I made the two following changes:

• the script checks if the json sidecar is empty or not. If it is empty, it considers that there was no existing json sidecar
• I addded to possibility to change the value stored in the json sidecar "Name" by adding the -task-name flag.

Ready for review

[valosekj]

To track such workflow (automatic segmentation by a model + manual correction), we add two entries into the JSON file (one for the model, the second for the rater). See here.

{
    "SpatialReference": "orig",
    "GeneratedBy": [
        {
            "Name": "sct_deepseg_sc",
            "Version": "SCT v6.1"
        },
        {
            "Name": "Manual",
            "Author": "Nathan Molinier",
            "Date": "2023-07-14 13:43:10"
        }
    ]
}

[valosekj]

You can specify the model's details using the -json-metadata arg; see L246-L260
@plbenveniste, can you please try -json-metadata arg on some of your corrections? I believe the output JSON sidecars will contain all the information you want to track.

[NathanMolinier]

I agree with Jan, sometimes we want to add multiple information to the sidecar and I believe the -json-metadata flag accomplish that already

[valosekj]

Please see #102