Communication through Fabric

Functions for using fabric

Abed uses Fabric to synchronize data to and from the compute cluster. The various functions that Abed needs are defined here. A helper class is defined in fab_util, which is used to define the Fabric context environment that is used for the commands.

abed.fab.build_remote()

Runs the build command remotely if the program requires compilation

abed.fab.deploy(push_data=False)

abed.fab.fab_pull()

abed.fab.fab_push()

abed.fab.fab_repull()

abed.fab.fab_setup()

abed.fab.get_files_from_glob(glob_path, glob, dest_dir)

abed.fab.get_results(basepath=None)

abed.fab.init_data()

Push the data to the remote server

This function is used to synchronize the DATADIR directory to the compute cluster. This is done by first locally compressing the entire directory as a tar file, then syncing it to a datasets directory in the remote project path, and unpacking it there. The path where the unpacked dataset is located is stored in the MyFabric class.

abed.fab.move_data()

Move the data from previous release

abed.fab.setup()

abed.fab.write_and_queue()