Automated Image Retrieval (AIR) Download

This is a small wrapper interface to the AIR web API. This will help you to batch download radiology studies if you have this service available on your PACS system.

Installation & Usage

With container

If on Mac, use the Dockerfile to compile this package and run it in a container.
If on Linux, use the Singularity image to compile and run this package.

For Singularity, build the image (once), with:

singularity build air_download.sif Singularity.def

Then, you can use the handy script run_air_download.py to run the container with the appropriate arguments. You only have to set the top 3 arguments (AIR_API_URL, DEFAULT_PROJECT_ID, DEFAULT_ANONYMIZATION_PROFILE) once in the script itself. If you don't yet have a air_login.txt file, it will prompt you to create it for you. It only requires a basic python install (no special packages).

python run_air_download.py -h  # for help message
python run_air_download.py 11111111  # to download a single study to current directory

If python is not an option, you can run the container directly – but you will need to mount the appropriate directories and files:

singularity run air_download.sif -h

Without container (python package)

Installation

Install the air_download package directly from the git repository like:

pip install git+https://github.com/rauschecker-sugrue-labs/air_download

(modify URL if the repository lives somewhere other than github)

Usage

air_download -c /path/to/air_login.txt https://air.<domain>.edu/api/ 11111111

Note on credentials

AIR_USERNAME=username
AIR_PASSWORD=password

Please ensure this file is reasonably secure.

chmod 600 air_login.txt

Alternatively, credentials may be stored as environment variables like:

AIR_USERNAME=username
AIR_PASSWORD=password
export AIR_USERNAME AIR_PASSWORD

air_download https://air.<domain>.edu/api/ 11111111

Type air_download -h for the help text.

$ air_download -h
usage: air_download [-h] [-c CRED_PATH] [-p PROFILE] [-o OUTPUT] URL ACCESSION

positional arguments:
  URL                   URL for AIR API, e.g. https://air.<domain>.edu/api/
  ACCESSION             Accession # to download

optional arguments:
  -h, --help            show this help message and exit
  -c CRED_PATH, --cred_path CRED_PATH
                        Login credentials file. If not present, will look for
                        AIR_USERNAME and AIR_PASSWORD environment variables.
                        (default: None)
  -p PROFILE, --profile PROFILE
                        Anonymization Profile (default: -1)
  -o OUTPUT, --output OUTPUT
                        Output path (default: ./<Accession>.zip)
  -s SERIES_INCLUSION, --series_inclusion SERIES_INCLUSION
                        Comma-separated list of series inclusion patterns (case insensitive, 'or' logic). Example for T1 type series: 't1,spgr,bravo,mpr' (default: None)

From within python, you can also import the module directly, so that it may be integrated with other tools.

import air_download.air_download as air
import argparse

args = argparse.Namespace()
args.cred_path = '/path/to/air_login.txt'
args.URL       = 'https://air.<domain>.edu/api/'
args.acc       = '11111111'
args.profile   = -1
args.output    = '11111111.zip'

air.main(args)