Skip to content

Create Archive Items

If you are working on a challenge, you most likely need to upload cases to an archive on the platform. This can be done via the API and most easily using the convienence method: Client.add_cases_to_archive

First things first, we need to get started and initiate the client:

import gcapi
client = gcapi.Client(token="your-personal-token")

Example¤

For this particular example, the data on your machine would be structured as follows:

mainfolder/
├── patient1_folder
│       ├──file_for_a_single_series
├── patient2_folder
│       ├──file_for_a_single_series
├──patient3_folder
│       ├──file_for_a_single_series
          ...
├──patientN_folder
│       ├──file_for_a_single_series

Lets now list the files in the upload directory: 

# Specify the directory path which you want to upload from.
upload_from_dir = Path(r"path/on/your/machine/with/data/for/archive")

# Create list of files in the specified directory.
files = sorted(f for f in upload_from_dir.rglob("*.*") if f.is_file())
print("Found", len(files), "cases for upload")

Create archive items¤

An Archive contains Archive Items. An archive item consists of one or more socket values, which can be images, files or other data. To create archive items for the algorithms to predict on, you need to provide a socket slug for each socket value. Within an archive item, a socket slug needs to be unique. For example, one cannot have three generic-medical-images within a single archive item.

In this example two archive items are created and each archive item contains three sockets, with slugs: a ct-image, an airway-segmentation and some-score.

Challenge Sockets

For archives that are linked to challenges, the sockets will need to correspond to the sockets that have been configured as input for the challenge algorithms. For an overview of these sockets, visit the submit page(s) of your challenge.

archive_items = [
    {
        "ct-image": files[0],
        "airway-segmentation": files[1],
        "some-score": 2.1,
    },
    {
        "ct-image": files[2],
        "airway-segmentation": files[3],
        "some-score": 7.9,
    },
    ...
]

To upload cases and create the archive items, you will need to know the slug of the particular archive you will work with.

# Specify the "slug" of the archive you want to upload your data to.
upload_archive_slug = "my-archive-slug"

archive_item_pks = client.add_cases_to_archive(
  archive=upload_reader_study_slug,
  archive_items=archive_items
)
print("Uploaded cases and created archive items: " + archive_item_pks)