bknisbac/dalmatian

dalmatian is a collection of high-level companion functions for Firecloud and FISS.

dalmatian

FISS' faithful companion.

dalmatian is a collection of high-level functions for interacting with Firecloud via Pandas dataframes.

Install

pip install firecloud-dalmatian

Requirements

FireCloud uses the Google Cloud SDK (https://cloud.google.com/sdk/) to manage authorization. To use dalmatian, you must install the SDK and login locally with

gcloud auth application-default login

Examples

Dalmatian provides the WorkspaceManager class for interacting with FireCloud workspaces.

import dalmatian
wm = dalmatian.WorkspaceManager(namespace, workspace)

Creating and managing workspaces

Create the workspace:

wm.create_workspace()

Upload samples and sample attributes (e.g., BAM paths). The attributes must be provided as a pandas DataFrame, in the following form:

• the index must be named 'sample_id', and contain the sample IDs
• the dataframe must contain the column 'participant_id'
• if a 'sample_set_id' columns is provided, the corresponding sample sets will be generated

wm.upload_samples(attributes_df, add_participant_samples=True)

If add_participant_samples=True, all samples of a participant are stored in participant.samples_.

Add or update workspace attributes:

attr = {
    'attribute_name':'gs://attribute_path',
}
wm.update_attributes(attr)

Get attributes on samples, sample sets, participants:

samples_df = wm.get_samples()
sets_df = wm.get_sample_sets()
participants_df = wm.get_participants()

Create or update sets:

wm.update_sample_set('all_samples', samples_df.index)
wm.update_participant_set('all_participants', participant_df.index)

Copy/move data from workspace:

samples_df = wm.get_samples()
dalmatian.gs_copy(samples_df[attibute_name], dest_path)
dalmatian.gs_move(samples_df[attibute_name], dest_path)

Clone a workspace:

wm2 = dalmatian.WorkspaceManager(namespace2, workspace2)
wm2.create_workspace(wm)

Running jobs

Submit jobs:

wm.create_submission(config_namespace, config_name, sample_id, 'sample', use_callcache=True)
wm.create_submission(config_namespace, config_name, sample_set_id, 'sample_set', expression=this.samples, use_callcache=True)
wm.create_submission(config_namespace, config_name, participant_id, 'participant', expression=this.samples_, use_callcache=True)

Monitor jobs:

wm.get_submission_status()

Get runtime statistics (including cost estimates):

status_df = wm.get_sample_status(config_name)
workflow_status_df, task_dfs = wm.get_stats(status_df)

Re-run failed jobs (for a sample set):

status_df = wm.get_sample_set_status(config_name)
print(status_df['status'].value_counts())  # list sample statuses
wm.update_sample_set('reruns', status_df[status_df['status']=='Failed'].index)
wm.create_submission(config_namespace, config_name, sample_set_id, 'reruns', expression=this.samples, use_callcache=True)

Contents

Including additional FireCloud Tools (enumerated below)

workflow_time
create_workspace
delete_workspace
upload_samples
upload_participants
update_participant_samples
update_attributes
get_submission_status
get_storage
get_stats
publish_config
get_samples
get_sample_sets
update_sample_set
delete_sample_set
update_configuration
check_configuration
get_google_metadata
parse_google_stats
calculate_google_cost
list_methods
get_method
get_method_version
list_configs
get_config
get_config_version
print_methods
print_configs
get_wdl
compare_wdls
compare_wdl
redact_outdated_method_versions
update_method
get_vm_cost

Usage

Some functionality depends on the installed gsutil.

When using PY3 this creates a potential issue of requiring multiple accessible python installs.

Remediate this issue by defining an env variable for gsutil python

# replace path with path to local python 2.7 path.
# if using pyenv the following should work
# (assuming of course 2.7.12 is installed)
export CLOUDSDK_PYTHON=/usr/local/var/pyenv/versions/2.7.12/bin/python