Details
Description
Internal and external developers can process precursor datasets (HSC, DECam, CFHT-LS, LSST lab data), generating and analyzing data products at small scale, with comparable difficulty to other toolkits. Documentation exists that external developers can use to add new algorithms. Problem reports can be turned into automated tests.
Attachments
Issue Links
- is blocked by
-
DLP-378 Refactor Orchestration, including task deployment and execution
-
- Unplanned
-
-
DLP-384 Refactor pipeline construction toolkit with modifications to support I/O
-
- Unplanned
-
-
DLP-216 Provide basic APIs to use Firefly visualization components
-
- Resolved
-
-
DLP-217 Provide APIs for tabular data display using Firefly
-
- Resolved
-
-
DLP-218 Provide API for image display in Firefly
-
- Resolved
-
-
DLP-343 Native binary stack distribution releases implemented
-
- Resolved
-
-
DLP-564 Able to use some data-access APIs developed by database team
-
- Resolved
-
-
DLP-21 Working Butler Framework
-
- Closed
-
-
DLP-342 Publish updated introductory software guide to developing with the stack
-
- Closed
-
-
DLP-346 SQuaSH Prototype using existing architectural components deployed
-
- Closed
-
-
DLP-390 Hierarchy of logging domains
-
- Closed
-
-
DLP-391 Initial version of inter-process messaging framework, wrapping MPI and Event Services
-
- Closed
-
-
DLP-457 Support runs on real cameras data and ImSim data
-
- Closed
-
-
DLP-458 Ingest of bulk data from real cameras to ComCam scale
-
- Closed
-
-
DLP-459 Qserv capability to hold data reduced at NCSA
-
- Closed
-
- relates to
-
DM-7540 Two DLP milestones may be misplaced in PMCS under 02C.06
- Won't Fix
-
DM-7541 DLP-579 may be at the wrong level in PMCS
- Won't Fix
Precursor datasets can be processed reasonably well with pipe_tasks, pipe_drivers, and the DESDM framework (as evidenced by LDM-503-2) at small to moderate scale. Documentation has been improved substantially for tasks, task writing, and overall development. The pytest framework has improved unit testing, and coding problems can be turned into unit tests relatively easily. Scientific problems are discovered and monitored through regular runs.
While there is still much work to be done, I feel this is adequate for closing this long-delayed milestone.