search

LSSTDESC / descwl-coadd-task

DM Task to run coaddition in cells
BSD 3-Clause "New" or "Revised" License
1 stars 0 forks source link

readme

descwl-coadd-task

Task to run coaddition in cells

tests codecov

Setting up the environment at USDF (for Gen3 butler)

Load a custom conda environment that has the Rubin Science Pipelines and other relevant packages installed.

For example,

conda activate /sdf/home/e/esheldon/miniconda3/envs/stack

Regularly update the stackvana package in the environment.

Similarly, you must setup this package or manually add the parent directory your PYTHONPATH.

cd descwl-coadd-task
setup -k -r .

Setup some convenient environment variables:

export REPO=/sdf/data/rubin/repo/dc2
export INPUT_COLLECTION=2.2i/runs/test-med-1/w_2023_49/DM-42037

Finally, invoke the pipetask run command. To produce the coadd for tract=3828, patch=19 and band="i", "r", run:

pipetask run -p pipeline.yaml -b $REPO -i $INPUT_COLLECTION -o u/$USER/coaddTest -d "skymap='DC2_cells_v1' AND tract=3828 AND patch=19 AND band IN ('i', 'r')"

If the output dataset has never been registered with the butler before, then the first time you run this command (and only for the first time), you will need to include --register-dataset-types option to the pipetask run command.

Notes on the pipeline file

A minimal pipeline file is provided in pipeline.yaml. This file currently calls the PipelineTasks to generate warps and to generate coadds from the warps. While bare tasks can be run using the --task invocation instead of -p, it is recommended to run the pipeline file instead. The pipeline file enforces 'contracts' that ensure that the different Tasks are configured in a consistent manner. For instance, attempting to warp the PSF with a different kernel than that was used to warp the images using

pipetask run -p pipeline.yaml -b $REPO -i $INPUT_COLLECTION -o u/$USER/coaddTest \
-d "skymap='DC2_cells_v1' AND tract=3828 AND patch=19 AND band IN ('i', 'r') \
-c "assembleCoadd:psf_warper.warpingKernelName='lanczos5'"

would result in an error.

Notes on configurability of the PipelineTasks

The PipelineTasks are highly configurable and can be configured in multiple places.

  1. Default values specified in the definition of the Config field.
  2. Default values overridden in the setDefaults method of the ConfigClass.
  3. Instrument-specific values in the obs_* package.
  4. Values specified in the pipeline file.
  5. Values specified as a config file with an invocation of -C.
  6. Values specified on the command-line with an invocation of -c.

Because the values can be overridden multiple times, the final set of configuration values that the Tasks were run with is stored as an output itself (which can be reused with the -C option).