mriapp.retrieve package

mriapp.retrieve.BaseRetrieve module

class mriapp.retrieve.BaseRetrieve.BaseRetrieve

Bases: object

Base class to retrieve new solver jobs

retrieve_file(location)

Fetch a file and return a json/dict representation

retrieve_task()

Retrieve the next task and return a json/dict representation. A task can contain multiple directives and files to fetch. This is a generator

mriapp.retrieve.LocalRetrieve module

class mriapp.retrieve.LocalRetrieve.LocalRetrieve(task_record)

Bases: object

Retrieve new solver jobs from local filesystem

Parameters:task_record (string) – File on the local system containing a list of folders with tasks

__del__()

Make sure we close our file on object deletion

retrieve_file(location)

Basically the identity function - we don’t need to get any network data for local files, so we’ll just pass back the file handle

Returns:location – Local location of a file. Return type:string

retrieve_task()

Retrieve the next task and return a json/dict representation.

A task can contain a number of directives, but will generally contain the training task for Caffe, including the locations of model/solvers. This function is a generator and will allow iteration through the entire set of possible tasks.

Yields:task (dict) – A task located on the local drive