Reduce class provides the underlying structure of the
command. This section describes and discusses the programmatic interface
available on the class
Reduce. This section is for users wanting to use
Reduce class programmatically.
reduce application introduced in the previous chapter is a user
interface script providing a command line access to the
reduce application parses the arguments and initialize the
class and its
runr method. It is possible to bypass the
command and sets attributes directly on an instance of
Reduce, as the
following discussion illustrates.
The Reduce class is defined in the
module. The Reduce class provides a set of attributes and one public method,
runr that launches a reduction. This is the only public method on the
5.1.1. Very Basic Usage¶
The most basic usage involves importing the class, instantiating it, assigning
a file to reduce and then launching the
>>> from recipe_system.reduction.coreReduce import Reduce >>> myreduce = Reduce() >>> myreduce.files.append('S20161025S0111.fits') >>> myreduce.runr()
5.1.2. Typical Usage for Reduction¶
A more typical usage for reducing data can involve setting other options and can include setting up a logger. When using the Gemini data reduction primitives, the logger is highly recommended.
Normal usage will also likely involve the use of the calibration database
caldb. We will ignore
caldb here and rather fully describe
it and its usage in a subsequent chapter, Calibration Service. See Full API Example
where we put it all together.
>>> from recipe_system.reduction.coreReduce import Reduce >>> from gempy.utils import logutils >>> >>> logutils.config(file_name='example.log') >>> >>> inputfiles = ['S20161025S0200.fits', 'S20161025S0201.fits'] >>> myreduce = Reduce() >>> myreduce.files = inputfiles >>> myreduce.runr()
coreReduce nor the
Reduce class initializes any logging activity.
This is the responsibility of the programmer. The Recipe System does not
require a logger but the Gemini primitives do. The absence of a logger when
using the Gemini data reduction package leads to double the reporting on
the screen. More an annoyance than a problem, admittedly.
You are free to provide your own logger, or you can use the fully defined
logger provided in DRAGONS. It is recommended that you use the system
Reduce is tuned to use the DRAGONS logger.
Returning to the example above, we could also set the recipe to a custom recipe, override a primitive parameters, set a data reduction package, etc. The attributes that can be set are discussed in Public Attributes to Reduce below.
>>> myreduce.recipename = 'myrecipelib.myrecipe' >>> myreduce.uparms = [('stackFrames:operation', 'median')] >>> myreduce.dkpkg = 'thirdpartydr' >>> # rerun with the modified recipe and parameter >>> myreduce.runr()
A notable quirk is how to set the
adpkg that is defined in the
command line interface. The
Reduce class does not have an attribute for
it. Instead, the programmer must import any third party AstroData instrument
configuration files explicitely before launching
>>> import astrodata >>> import thirdparty_instruments >>> >>> myreduce.Reduce() >>> myreduce.drpkg = 'thirdpartydr' >>> myreduce.files.append('filename.fits') >>> myreduce.runr()
5.2. Public Attributes to
|Public Attribute||Python type||Default|
|files||<type ‘list’ of ‘str’>|||
|output_filenames||<type ‘list’ of ‘str’>||None|
|uparms||<type ‘list’ of ‘tuple’>||None|
|upload||<type ‘list’ of ‘str’>||None|
A list of input file names to reduce. Only the first file in the list will be used for the recipe and primitive selection.
A list of output file names. This cannot be set. It is a return value. It is used after the recipe has run to collect the names of the files that were created.
output_stack = myreduce.output_filenames
The reduction mode. The Gemini data reduction package currently supports ‘sq’ and ‘qa’, with ‘ql’ in the works. [‘sq’: Science Quality, ‘qa’: Quality Assessment, ‘ql’: Quick Look Reduction.]
myreduce.mode = 'qa'
The name of the recipe to use. If left to “_default”, the Recipe System will invoke the mappers and select the best matching recipe library and use its default recipe.
If only the name of a recipe is provided, the mappers will be invoked to find the best matching recipe library and use the named recipe rather than the default.
If a “module.recipe” string is provided, the user’s “module” will be imported and the user’s “recipe” will be used. No mapping will be done.
myreduce.recipename = 'myrecipelib.myrecipe'
If the name of a primitive is given, the Recipe System will find the best match primitive set and run the specified primitives from that set.
The suffix to add the final outputs of a recipe. In the Gemini primitives, default suffixes are assigned to each primitives. Setting
suffixwill override the default suffix of the last primitive in the recipe.
myreduce.suffix = '_flatBfilter'
The name of the data reduction package to use. The default is
geminidr. If using a third-party package, or during new instrument development, set this attributes to import the correct suite of recipes and primitives.
myreduce.drpkg = 'scorpiodr'
Set the processed calibration to be used. This overrides the automatic selection done by the calibration manager, if one is being used. This setting must be used if no calibration manager is used or available, or when, for example, the calibrations association rules are not yet implemented. It is also useful for testing and for getting full control of the calibrations being used.
The format for this attribute’s value needs to looks like this:
There must be one entry for each type of calibrations.
The recognized calibration types are currently:
normalize_ucalsfunction in the
recipe_system.utils.reduce_utilsmodule will convert a list of strings (in the format required for the command-line interface) into the appropriate dictionary and check that the specified calibrations files are readable and have the required tags. However, it is not necessary to make use of this function.
Here’s how to use
from recipe_system.utils.reduce_utils import normalize_ucals mycalibrations = ['processed_bias:/path/master_bias.fits', 'processed_flat:/path/master_Bflat.fits'] myreduce.ucals = normalize_ucals(mycalibrations)
Set primitive parameter values. This will override the primitive defaults. This is a list of tuples with the primitive name and parameter in the first element, and the value in the second one.
myreduce.uparms = [('stackFrames:operation', 'median')]
If the primitive name is omitted all parameters with that name, in any primitives will be reset. Be careful.
- Internal use only. Specify which types of product to upload to the Gemini internal database. Allowed values are “metrics”, “calibs”, and “science”.