Examples¶

This page provides some brief, minimal examples for common tasks with the SDK.

Groups¶

For managing permissions, see Permissions.

Add Group¶

Create a new group, with an id of my_group and a label of My Test Group:

group_id = fw.add_group(flywheel.Group('my_group', 'My Test Group'))

List Groups¶

List all groups that I have access to:

for group in fw.groups():
  print('%s: %s' % (group.id, group.label))

Modify Group Tags¶

Add tags named Control and Study to a group.

group = fw.get(group_id)
group.add_tag('Control')
group.add_tag('Study')

group = group.reload()
print(', '.join(group.tags))

List Projects in Group¶

List all of the projects belonging to group.

for project in group.projects():
  print('%s: %s' % (project.id, project.label))

Projects¶

For managing tags, files, notes or info, see Containers.

For managing permissions, see Permissions.

Add Project¶

Create a new project that belongs to group_id with a label of My Test Project.

project = group.add_project(label='My Test Project')

List Projects¶

List all of the projects that I have access to:

for project in fw.projects():
  print('%s: %s' % (project.id, project.label))

List Subjects in Project¶

List all subjects belonging to project.

for subject in project.subjects():
        print('%s: %s' % (subject.id, subject.label))

List Sessions in Project¶

List all of the sessions belonging to project.

for session in project.sessions():
        print('%s: %s' % (session.id, session.label))

Subjects¶

For managing tags, files, notes or info, see Containers.

Add Subject¶

Create a new subject with a label of Subject 01

subject = project.add_subject(label='Subject 01')

List Subjects¶

List all of the subjects that I have access to:

for subject in fw.subjects():
        print('%s: %s' % (subject.id, subject.label))

List Sessions in Subject¶

List all of the sessions belonging to subject.

for session in subject.sessions():
        print('%s: %s' % (session.id, session.label))

Modify Subject¶

Update the details of a subject

subject.update(
        firstname='John',
        lastname='Doe',
        cohort='Study',
        type='human',
        sex='male',
        race='Unknown or Not Reported'
)

Sessions¶

For managing tags, files, notes or info, see Containers.

Add Session¶

Create a new session with a label of Session 01.

session = subject.add_session(label='Session 01')

List Sessions¶

List all of the sessions that I have access to:

for session in fw.sessions():
        print('%s: %s' % (session.id, session.label))

List Acquisitions in Session¶

List all of the acquisitions belonging to session.

for acquisition in session.acquisitions():
  print('%s: %s' % (acquisition.id, acquisition.label))

Acquisitions¶

For managing tags, files, notes or info, see Containers.

For uploading and downloading files, see Dealing with Files.

Add Acquisition¶

Create a new acquisition with a label of Localizer, and upload a file.

acquisition = session.add_acquisition(label='Localizer')

acquisition.upload_file('localizer.nii.gz')

List Acquisitions¶

List all of the acquisitions that I have access to:

for acquisition in fw.acquisitions.iter():
  print('%s: %s' % (acquisition.id, acquisition.label))

List Files in Acquisition¶

List all of the files on an acquisition.

for file in acquisition.files:
  print(file.name)

Collections¶

Collection let you group subjects, sessions or acquisitions across multiple projects to meet your unique needs.

Add Acquisition¶

Create a new collection with a label of My Collection

collection_id = fw.add_collection({'label': 'My Collection'})

Find a collection¶

Find your collection per label or collection.id

collection = fw.collections.find_first('label=My Collection')

collection = fw.get_collection(collection_id)

List Collections¶

List all of the collections that I have access to:

for collection in fw.collections.iter():
  print('%s: %s' % (collection.id, collection.label))

Add sessions or acquisitions to a collection¶

You can add sessions or acquisitions to a collection

acquisitions = fw.acquisitions.find('label=T1w')  # all acquisitions with label=T1w
    for acquisition in acquisitions.iter():
    collection.add_acquisition(acquisition.id)

sessions = fw.sessions.find('label=baseline')  # all sessions with label=baseline
    for session in sessions.iter():
    collection.add_session(session.id)

Analyses¶

NOTE: Analyses are available on Projects, Subjects, Sessions and Acquisitions.

For managing tags, files, notes or info, see Containers.

For uploading and downloading files, see Dealing with Files.

Add Analysis¶

Create a new analysis on session referencing an input file from an acquisition, then upload a file.

file_ref = acquisition.get_file('localizer.nii.gz').ref()
analysis = session.add_analysis(label='Localizer Analysis', inputs=[file_ref])

analysis.upload_output('my-analysis.csv')

List Session Analyses¶

List all of the analyses belonging to session.

for analysis in session.analyses():
        print('%s: %s' % (analysis.id, analysis.label))

Archive Downloads¶

Occasionally it’s desirable to download all files of a given type from one or more containers. Flywheel provides this capability in the form of tarfile downloads. An archive can be downloaded from a single container, or a list of containers, and can include or exclude given file types.

For example:

project = fw.lookup('flywheel/Test Project')

# Download all NIfTI files in the project
project.download_tar('test-project.tar', include_types=['nifti'])

# Download all non-DICOM data from sessions created since 2018-10-31
sessions = project.sessions.find('created>2018-10-31')
fw.download_tar(sessions, 'session-files.tar', exclude_types=['dicom'])

Jobs And Analyses¶

Scheduling Jobs¶

Running a gear requires a few questions to be answered:

1. What gear to run?¶

A gear can be located by name and (if desired) version using the resolver. Calling print_details will print a textual description of the gear, including inputs and configuration values, and will help answer the remaining questions.

For example:

# Get the latest version of the example gear
gear = fw.lookup('gears/flywheel-example-gear')

# Get a specific version of the example gear
gear = fw.lookup('gears/flywheel-example-gear/0.0.4')

# Print details about the gear
gear.print_details()

Flywheel Example Gear

Sample gear to demonstrate a simple use case of outputting the name of each input file.

Name: flywheel-example-gear Version: 0.0.4 Category: converter Author: Flywheel <support@flywheel.io> Maintainer: Ryan Sanford <ryansanford@flywheel.io> URL: https://flywheel.io/ Source: https://github.com/flywheel-apps/example-gear

Inputs:

dicom (file, required)
Type: dicom Any dicom file.

file (file, required)
Any file.

text (file, optional)
Any test file that is 10 KB in size or less.

Configuration:

multiple (number, default: 20)
Any two-digit multiple of ten.

string (string, default: Example)
Any string.

number (number, default: 3.5)
Any number.

phone (string, default: 555-5555)
Any local phone number, no country or area code.

boolean (boolean, default: True)
Any boolean.

integer (integer, default: 7)
Any integer.

string2 (string, default: Example 2)
Any string from 2 to 15 characters long.

2. What type of job?¶

There are generally two types of gears: Utility and Analysis gears.

Utility gears generally perform basic data conversion and QA tasks. Often times they run within the context of a single container, taking input files and generating output files and/or metadata.

Analysis gears are a bit different in that they create a new Analysis object when they run. A destination is still specified, but rather than outputs being attached directly to the destination container, a new analysis is attached to that container, which contains any output files.

When executing analysis gears, an analysis label is required. The gear category (in the description above) determines whether or not a gear is an analysis gear.

3. What are the inputs?¶

Gears can specify one or more file inputs, and can designate whether those file inputs are optional or required. It’s not uncommon for the gear to also designate an input file type.

4. Where should outputs go?¶

In addition to the input files, a destination container for output files is required. In the case of analysis gears, the destination will be a new analysis object on the destination container.

5. What configuration is desired?¶

Finally, any configuration values that do not have default values, or desirable default values should be specified at job creation time.

Utility Job Example¶

The Gear object provides the ability to directly start a job:

# Get the Flywheel Example Utility gear
gear = fw.lookup('gears/flywheel-example-gear')

# Find the input files, acquisition will be the destination container
acquisition = fw.lookup('flywheel/Test Project/sub-1000/session1/Scan')
inputs = {
        'dicom': acquisition.get_file('scan.dicom.zip'),
        'file': acquisition.get_file('hello-world.txt')
}

# Override some configuration values, the rest will use defaults
config = {
        'number': 42,
        'string': 'Hello World!'
}

# Schedule the job, adding the "my-job" tag
job_id = gear.run(config=config, inputs=inputs, destination=acquisition, tags=['my-job'])

Analysis Job Example¶

The main difference when running an analysis is setting a label:

# Get the afq-demo gear
gear = fw.lookup('gears/afq-demo')

# Find the session, which will be the destination
session = fw.lookup('flywheel/Test Project/sub-1000/session1')

# Determine the input files, which are on the DTI acquisition
# Find the DTI acquisition, which contains the input files
inputs = {
        'diffusion': dti.get_file('8892_14_1_dti.nii.gz'),
        'bvec': dti.get_file('8892_14_1_dti.bvec'),
        'bval': dti.get_file('8892_14_1_dti.bval')
}

# Set config value
config = {
        'qmr_metadata_bvalue': 2000
}

# Schedule the job, which returns the analysis ID
analysis_id = gear.run(analysis_label='My AFQ Demo', config=config, inputs=inputs, destination=session)

Batch Scheduling Example¶

It is also possible to schedule batch jobs on a group of containers. In this case, a job will be scheduled for each container specified that has matching inputs. Batch scheduling happens via a proposal process.

You can also schedule batch analysis runs in this manner, by providing an analysis_label.

For example:

# Get the dcm2niix gear
gear = fw.lookup('gears/dcm2niix')

# Find matching acquisitions, using regular expression match on label
session = fw.lookup('flywheel/Test Project/sub-1000/session1')
t1_acquisitions = session.acquisitions.find('label=~^T1')

# Propose the batch
proposal = gear.propose_batch(t1_acquisitions, config={'merge2d': 'y'})

# Run the batch job
jobs = proposal.run()