2. Data Reduction¶
This chapter will guide you on reducing GMOS imaging data using command line tools. In this example we reduce a GMOS observation star field. The observation is a simple dither-on-target sequence. Just open a terminal to get started.
While the example cannot possibly cover all situations, it will help you get acquainted with the reduction of GMOS data with DRAGONS. We encourage you to look at the Tips and Tricks and Issues and Limitations chapters to learn more about GMOS data reduction.
DRAGONS installation comes with a set of scripts that are used to reduce astronomical data. The most important script is called “reduce”, which is extensively explained in the Recipe System Users Manual. It is through that command that a DRAGONS reduction is launched.
For this tutorial, we will be also using other Supplemental Tools, like:
2.1. The dataset¶
If you have not already, download and unpack the tutorial’s data package. Refer to Downloading the tutorial datasets for the links and simple instructions.
The dataset specific to this example is described in:
Here is a copy of the table for quick reference.
10 s, i-band
40 to 16 s, i-band
2.2. Set up the Local Calibration Manager¶
DRAGONS comes with a local calibration manager that uses the same calibration
association rules as the Gemini Observatory Archive. This allows
to make requests to a local light-weight database for matching processed
calibrations when needed to reduce a dataset.
Let’s set up the local calibration manager for this session.
~/.geminidr/, create or edit the configuration file
[calibs] standalone = True database_dir = /path_to_my_data/gmosimg_tutorial/playground
This simply tells the system where to put the calibration database, the database that will keep track of the processed calibrations we are going to send to it.
The tilde (
~) in the path above refers to your home directory.
Also, mind the dot in
Then initialize the calibration database:
That’s it! It is ready to use! You can check the configuration and confirm
the setting with
You can add processed calibrations with
caldb add <filename> (we will
later), list the database content with
caldb list, and
caldb remove <filename> to remove a file only from the database
(it will not remove the file on disk). For more the details, check the
“caldb” documentation in the
Recipe System: User’s Manual.
2.3. Check files¶
For this example, all the raw files we need are in the same directory called
../playdata/. Let us learn a bit about the data we have.
Ensure that you are in the
playground directory and that the
environment that includes DRAGONS has been activated.
Let us call the command tool “typewalk”:
$ typewalk -d ../playdata/ directory: /data/workspace/gmosimg_tutorial/playdata N20170613S0180.fits ............... (AT_ZENITH) (AZEL_TARGET) (BIAS) (CAL) (GEMINI) (GMOS) (NON_SIDEREAL) (NORTH) (RAW) (UNPREPARED) ... N20170614S0201.fits ............... (GEMINI) (GMOS) (IMAGE) (NORTH) (RAW) (SIDEREAL) (UNPREPARED) ... N20170615S0534.fits ............... (AT_ZENITH) (AZEL_TARGET) (BIAS) (CAL) (GEMINI) (GMOS) (NON_SIDEREAL) (NORTH) (RAW) (UNPREPARED) ... N20170702S0182.fits ............... (CAL) (FLAT) (GEMINI) (GMOS) (IMAGE) (NORTH) (RAW) (SIDEREAL) (TWILIGHT) (UNPREPARED) Done DataSpider.typewalk(..)
This command will open every FITS file within the folder passed after the
flag (recursively) and will print an unsorted table with the file names and the
associated tags. For example, calibration files will always have the
tag. Flat images will always have the
FLAT tag. This means that we can start
getting to know a bit more about our data set just by looking the tags. The
output above was trimmed for presentation.
2.4. Create File lists¶
This data set contains science and calibration frames. For some programs, it could have different observed targets and different exposure times depending on how you like to organize your raw data.
The DRAGONS data reduction pipeline does not organize the data for you. You have to do it. DRAGONS provides tools to help you with that.
2.4.1. List of Biases¶
The bias files are selected with “dataselect”:
$ dataselect --tags BIAS ../playdata/*.fits -o list_of_bias.txt
2.4.2. List of Flats¶
Now we can do the same with the FLAT files:
$ dataselect --tags FLAT ../playdata/*.fits -o list_of_flats.txt
If your dataset has flats obtained with more than one filter, you can add the
--expr 'filter_name=="i"' expression to get only the flats obtained within
the i-band. For example:
$ dataselect --tags FLAT --expr 'filter_name=="i"' ../playdata/*.fits -o list_of_flats.txt
2.4.3. List for science data¶
The rest is the data with your science target. The simplest way, in this case, of creating a list of science frames is excluding everything that is a calibration:
$ dataselect --xtags CAL ../playdata/*.fits -o list_of_science.txt
This will work for our dataset because we know that a single target was observed with a single filter and with the same exposure time. But what if we don’t know that?
$ dataselect --expr 'observation_class=="science"' ../playdata/*.fits | showd -d object,exposure_time ----------------------------------------------------------- filename object exposure_time ----------------------------------------------------------- ../playdata/N20170614S0201.fits starfield 10.0 ../playdata/N20170614S0202.fits starfield 10.0 ../playdata/N20170614S0203.fits starfield 10.0 ../playdata/N20170614S0204.fits starfield 10.0 ../playdata/N20170614S0205.fits starfield 10.0
To select on target name and exposure time, specify the criteria in the
expr field of “dataselect”:
$ dataselect --expr '(object=="starfield" and exposure_time==10.)' ../playdata/*.fits -o list_of_science.txt
We have our input lists and we have initialized the calibration database, we are ready to reduce the data.
Please make sure that you are still in the
2.5. Create a Master Bias¶
We start the data reduction by creating a master bias for the science data. It can be created and added to the calibration database using the commands below:
$ reduce @list_of_bias.txt $ caldb add N20170613S0180_bias.fits
@ character before the name of the input file is the “at-file” syntax.
More details can be found in the "at-file" Facility documentation.
To check that the master bias was added to the database, use
The master bias will be saved in the same folder where “reduce” was
called and inside the
./calibrations/processed_bias folder. The latter
location is to cache a copy of the file. This applies to all the processed
Some people might prefer adding the copy in the
as it is safe from a
rm *, for example.
$ caldb add ./calibrations/processed_bias/N20170613S0180_bias.fits
“reduce” uses the first filename in the input list as basename and adds
_bias as a suffix to it. So if your first filename is, for example,
N20170613S0180.fits, the output will be N20170613S0180_bias.fits`.
2.6. Create a Master Flat Field¶
Twilight flats images are used to produce an imaging master flat and the result is added to the calibration database.
$ reduce @list_of_flats.txt $ caldb add N20170702S0178_flat.fits
Note “reduce” will query the local calibration manager for the master bias and use it in the data reduction.
Once finished you will have the master flat in the current work directory and
./calibrations/processed_flat. It will have a
2.7. Create Master Fringe Frame¶
2.8. Reduce Science Images¶
Once we have our calibration files processed and added to the database, we can
reduce on our science data:
$ reduce @list_of_science.txt
This command will generate bias and flat corrected files and will stack them.
If a fringe frames is needed this command will apply the correction. The stacked
image will have the
The output stack units are in electrons (header keyword BUNIT=electrons). The output stack is stored in a multi-extension FITS (MEF) file. The science signal is in the “SCI” extension, the variance is in the “VAR” extension, and the data quality plane (mask) is in the “DQ” extension.
Depending on your version of Astropy, you might see a lot of Astropy warnings about headers and coordinates system. You can safely ignore them.
Below are one of the raw images and the final stack: