Getting started tutorial part 1: setting up the Butler data repository¶

This hands-on tutorial is intended for anyone getting started with using the LSST Science Pipelines for data processing. You’ll get a feel for setting up a Pipelines environment, working with data repositories, running command-line tasks, and working with the Pipelines’ Python APIs. Along the way we’ll point you to additional documentation.

The LSST Science Pipelines can process data from several telescopes using LSST’s algorithms. In this tutorial series you will calibrate and reduce Hyper Suprime-Cam (HSC) exposures into coadditions and catalogs of objects.

In this first part of the tutorial series you’ll set up the LSST Science Pipelines software, and collect the raw observations and calibration data needed for the tutorial. Along the way, you’ll be introduced to the Butler, which is the Pipelines’ interface for managing, reading, and writing datasets.

Install the LSST Science Pipelines¶

If you haven’t already, you’ll need to install the LSST Science Pipelines. We recommend that you install the pre-built binary packages by following the instructions at Install with newinstall.sh and eups distrib. This tutorial is intended to work with current tag of the lsst_distrib EUPS package.

When working with the LSST Science Pipelines, you need to remember to activate the installation and set up the package stack in each new shell session. Follow the instructions Setting up installed LSST Science Pipelines to do this.

To make sure the environment is set up properly, you can run:

eups list lsst_distrib

The line printed out should contain the word setup. If not, review the set up instructions. It may simply be that you’re working in a brand new shell.

Downloading the sample HSC data¶

Sample data for this tutorial comes from the testdata_ci_hsc package. testdata_ci_hsc contains a small set of Hyper Suprime-Cam (HSC) exposures. The Science Pipelines provides native integrations for many observatories, including HSC, CFHT/MegaCam, SDSS, and of course LSST.

testdata_ci_hsc is a Git LFS-backed package, so make sure you’ve installed and configured Git LFS for LSST.

Important

Even if you’ve used Git LFS before, you do need to configure it to work with LSST’s servers.

First, clone testdata_ci_hsc using Git:

git clone https://github.com/lsst/testdata_ci_hsc

Then setup the package to add it to the EUPS stack:

setup -j -r testdata_ci_hsc

Tip

The -r testdata_ci_hsc argument is the the package’s directory path (either absolute or relative). In this case

The -j argument means that we’re just setting up testdata_ci_hsc without affecting other packages.

Now run:

echo $TESTDATA_CI_HSC_DIR

The $TESTDATA_CI_HSC_DIR environment variable should be the testdata_ci_hsc directory’s path.

Creating a Butler repository for HSC data¶

In the LSST Science Pipelines you don’t directly manage data files. Instead, you access data through the Butler client. This gives you flexibility to work with data from different observatories without significantly changing your workflow.

The Butler manages data in repositories. Butler repositories can be remote (the data is on a server, across a network) or local (the data in on a local filesystem). In this tutorial you’ll create and use a local Butler repository, which is a simple directory.

Go ahead and create the local Butler repository as a directory called DATA:

mkdir DATA

Then add a _mapper file to the repository:

echo "lsst.obs.hsc.HscMapper" > DATA/_mapper

The Butler uses the mapper to find and organize data in a format specific to each camera. Here you’re using the lsst.obs.hsc.HscMapper mapper because you’re processing HSC data in this repository.

This is what your current working directory should look like right now:

testdata_ci_hsc/
DATA/

Ingesting raw data into the Butler repository¶

Next, populate the repository with data from testdata_ci_hsc. The Pipelines’ ingestImages.py command (called a command-line task) links raw images into a Butler repository, allowing the mapper to organize the data. Run:

ingestImages.py DATA $TESTDATA_CI_HSC_DIR/raw/*.fits --mode=link

Tip

Notice that the first argument to most command-line tasks is the Butler repository. In this case it’s the DATA directory.

Tip

You can learn about the arguments for command-line tasks with the -h flag. For example:

ingestImages.py -h

Install transmission curves¶

Run this command to install transmission curves corresponding to the raw data:

installTransmissionCurves.py DATA

Transmission calibrations, like this, are currently a special feature for HSC data implemented in the obs_subaru package.

Ingesting calibrations into the Butler repository¶

Next, add calibration images (such as dark, flat, and bias frames) associated with the raw data:

ln -s $TESTDATA_CI_HSC_DIR/CALIB/ DATA/CALIB

Note

In general, you can use the ingestCalibs.py command-line task to ingest calibrations into a Butler repository. For this tutorial, we’ve taken a shortcut by manually symlinking pre-structured calibrations from the testdata_ci_hsc package.

Ingesting a reference catalog into the Butler repository¶

The Pipelines use external stellar catalogs to refine the WCS and photometric calibration of images. testdata_ci_hsc includes a subset of the Pan-STARRS PS1 catalog that has been prepared as an astrometric and photometric reference catalog. Ingest that catalog into the Butler repository by creating a symlink:

mkdir -p DATA/ref_cats
ln -s $TESTDATA_CI_HSC_DIR/ps1_pv3_3pi_20170110 DATA/ref_cats/ps1_pv3_3pi_20170110

Wrap up¶

In this tutorial, you’ve set up a Butler repository with the data you’ll process in later steps. Here are some key takeaways:

The Butler is the interface between data and LSST Science Pipelines processing tasks.
Butler repositories can be hosted on different backends, both remote and local. In this case you created a local Butler repository on your computer’s filesystem.
Butler repositories contain raw data, calibrations, and reference catalogs. As you’ll see in future tutorials, the Butler repository also contains the outputs of processing tasks.
Command-line tasks like ingestImages.py and ingestCalibs.py help you seed data into Butler repositories.

In part 2 of this tutorial series you will process the HSC data in this newly-created Butler repository into calibrated exposures.

Navigation