Uploaded image for project: 'SAFe Program'
  1. SAFe Program
  2. SP-3073

Identify and demonstrate astronomy workflow that uses Dask

Change Owns to Parent OfsSet start and due date...
    XporterXMLWordPrintable

Details

    • SRCnet
    • Hide

      Dask Gateway deployment has been demoed, enabling users to create Dask clusters on demand. Dask is widely used in a range of data science fields, but in order to test this functionally, we would like to have some basic astronomy workflows that we can demonstrate running on such an ephemeral cluster. The resulting demo will be a useful resource to show astronomers unfamiliar with Dask, and the workflow (likely a Jupyter Notebook) will be a good starting point for functional testing deployments, as well as providing a link to the use case work ongoing in WG6.

      Show
      Dask Gateway deployment has been demoed, enabling users to create Dask clusters on demand. Dask is widely used in a range of data science fields, but in order to test this functionally, we would like to have some basic astronomy workflows that we can demonstrate running on such an ephemeral cluster. The resulting demo will be a useful resource to show astronomers unfamiliar with Dask, and the workflow (likely a Jupyter Notebook) will be a good starting point for functional testing deployments, as well as providing a link to the use case work ongoing in WG6.
    • Hide

      AC1: A demo and a workflow (ideally a Jupyter Notebook or Python script runnable on a JupyterHub platform) of an astronomical analysis which benefits from Dask acceleration running on an ephemeral Dask cluster.

      Show
      AC1: A demo and a workflow (ideally a Jupyter Notebook or Python script runnable on a JupyterHub platform) of an astronomical analysis which benefits from Dask acceleration running on an ephemeral Dask cluster.
    • 1.5
    • 1.5
    • 0
    • Team_CORAL
    • Sprint 5
    • Show
      https://confluence.skatelescope.org/display/SRCSC/COR-260%3A+Astronomy+workflow+Dask+-+Deploy+dependencies+and+pipeline+environment and https://confluence.skatelescope.org/display/SRCSC/COR-287%3A+Astronomy+workflow+Dask+-+Run+and+validate+pipeline+execution
    • 18.6
    • Stories Completed, Outcomes Reviewed, Demonstrated, Satisfies Acceptance Criteria, Accepted by FO
    • PI24 - UNCOVERED

    • PI18-PB SRC-SciPlat SRC-UseCase

    Description

      This feature follows previous work demonstrating functionality provided by Dask Gateway. We wish to now identify and demonstrate an astronomy workflow that utilises Dask to distribute processing across multiple nodes.

      This could be done with a standalone DaskHub (JupyterHub + Dask Gateway) deployment if to be picked up by a team new to Dask Gateway.

      Starting points: Arpan Das/others may have access to pre-existing examples, otherwise there may be an element of adapting an existing workflow to use Python/Dask data structures. Failing that the RASCIL library may be a good starting point (see related feature: https://jira.skatelescope.org/browse/SP-2889)

      Attachments

        Issue Links

          Structure

            Activity

              People

                j.collinson Collinson, James
                j.collinson Collinson, James
                Votes:
                0 Vote for this issue
                Watchers:
                3 Start watching this issue

                Feature Progress

                  Story Point Burn-up: (100.00%)

                  Feature Estimate: 1.5

                  IssuesStory Points
                  To Do00.0
                  In Progress   00.0
                  Complete1122.0
                  Total1122.0

                  Dates

                    Created:
                    Updated:
                    Resolved:

                    Structure Helper Panel