Fixed Github Actions setup
Removed pandas version constraint and implemented workaround to be able to use latest pandas version.
Modified installation of dcr on travis and Github Actions
New `File Suffixes`_ section
Added more test code.
Cleaned .travis.yml and created special dcr installation script for travis.
Changed the way core.combine_data names new files.
Fixed version checker due to deprecation of access token as url parameter.
New core.identify_technique. Was developed in the web application context.
Created –skip-slit-trim argument to provide more control for certain use cases.
Removed python 3.5 because it will not be supported anymore.
Improved AEON Support:
Added values for OBSTYPE and required logic.
Serial and Parallel binning extraction from header was not working
Changed url for astroplan’s server
Made it compatible with Astropy 4.0
All versions are free except for Pandas [#314]
wavelength.WavelengthCalibration.__call__ can now return a json output.
core.setup_logging can now create a generic logger (same format).
Modified how master bias are named.
Removed bias overscan and trimming correction on master bias creation.
–max-targets was not being used, missed connection in MainApp.
Updated keyword values of all reference lamps in the library according to [#292]
Refactored wavelength.WavelengthCalibration class moving several methods to core [#300, #303]
Refactored wavelength.WavelengthCalibration to be instantiated without arguments.
Improved messages at critical stages of wavelength calibration.
Moved setup_logging call from main package __init__ to scripts or entry points, this allows to re use other master loggers.
Changed –background-threshold to multiply by detection limit instead of background level
Created standard JSON output for
Bias process was not fully ignored when –ignore-bias was used [#289].
pandas version was not specified in environment.yml [#288, #290]
Target extraction failed for low signal targets because background subtraction was being ignored at the step of actually identifying targets.
Install instructions updated [#290]
Moved static methods from ImageProcessor to core.
Added function to validate ccd regions using regular expressions.
Using lamps keywords to select reference lamps.
Replaced target_stddev by target_fwhm in function extract and extract_fractional.
Replaced nsigmas by nfwhm everywhere.
Added argument –background-threshold with default value 3.
Added argument –fit-targets-with with options moffat and gaussian.
If there was more than one lamp for a science target the lamp recorded as used was one of them only.
A percentage symbols was added to the help of
--saturationargument, this caused a crash when
redccd --helpwas used.
Numpy is fixed to the version
1.15.2until further notice.
Reference lamps now get the extraction window added to the end of the file name. This is to avoid overwritting the lamps when they were used for more than one target.
DCR install script is now more advanced and requires a virtual environment to work on.
Added SOAR Logo to ReadTheDocs page.
Changed install instruction with exact steps and commands instead of referencing documentation.
Improved method to detect saturated images. Added a table with the half full well for all the readout modes possible and created a method to easily retrieve the value. This is a big improvement since in earlier versions the saturation limit was set to 65000 ADU regardless the input data and the user had to set a different one using the argument
--saturationcommand line argument, now is used to define the percentage of pixels above the saturation level, which for simplicity is the value of half full well. A default value of 1 percent was set as default.
Added record information of target trace into the header and logs.
Added record of background extraction regions into the header and logs.
Made all plots full screen and the images using the
Trace information is printed in the logs and also is recorded in the image’s header
Added sigma clipping to target tracing functions
Version 1.1.2 is pip instalable
pip install goodman-pipeline
Project and package renamed to
goodman_pipelinethis is because the previous was too generic. Now we have this structure:
goodman_pipeline/ docs/ goodman_pipeline/ core/ images/ ..etc setup.py ..etc
DataFrameindex is unusable when partial parts are eliminated. Added
Data conversion from string to integer needed to be converted to float first.
For low SNR data there was confusion of noise with targets, added a median filter and increased the the
ordervalue of peak detection.
Created several new keywords:
Extraction window at the first column.
Used for extracted comparison lamps, contains the name of the file of science target that the lamp was extracted for.
For science targets, it records the name of the lamp used for the wavelength calibration.
“Sliding” cross correlation window (to trace non-linearity of wavelength solution) is set to the maximum value between the length of the lamp spectrum in pixels and four times the global cross correlation of the reference lamp to the new one.
Iterations in sigma clipping of differences between obtained wavelength values and laboratory values was increased from 1 to 3. This is for removing bad fitted lines and also RMS error calculation.
Gaussian Kernel size for reference lamp convolution is now dependent on slit size and binning
Added reference lamps for all gratings and their modes except
Increased code coverage
Noneelements in list of instances of
Improved several logging messages
In general, it informs more, when it does an action and when it does not. What files are discarded,
Debugging plots are more complete for
Created new argument
--debug-plotdedicated for graphical debugging, the old
--debugwill show additional messages but will not produce any graphical output.
Removed ability to process several folders in sequence, now the pipeline has to be run for each folder separately.
Added clean exit when pipeline is unable to determine
Conversion from string to integer not always works, added intermediate float conversion.
Abrupt exit when there were non-fits-compliant keywords. Now it attempts to fix them all automatically and warns the user. Also, it ends the execution and informs the user to try again.
Removed unused code and tools.
Implemented Authorized GitHub API access and added actual version check
Moved command line interface from
Specified version of
cythonto be able to build.
Added reference lamps for all usable modes for the grating 600 l/mm
Created method to use automatic keyword fix from
Improved help information of arguments
--keep-cosmic-filewould work for
dcrbut not for
Changed organization of ReadTheDocs information
Added references to external packages
This page is the single place to add changes information. CHANGES.md still exist but contains a link here.
Implemented astroscrappy’s LACosmic method
defaultmethod for cosmic ray rejection.
For binning 1x1 default is dcr
For binning 2x2 default is lacosmic
For binning 3x3 default is lacosmic
none can still be forced by using
programatically access to the version number did not work because it was based purely on
setup.pyhas a function that creates the file
pipeline.versionwhich is accessed by
File naming was making some file dissapear by being overwritten for files that contained more than one target the next file name would match the previous one. A differentiator was added.
goodman/pipeline/info.pyand placed all metadata in
Several updates to documentation
Added comment on how to organize data on
Added link to licence on footer.
User manual now is in ReadTheDocs and no longer available as a pdf.
Improved information on debug plots
GSP_FNAMvalue for reference lamps
Spectral limit calculation by including binning into the equation
Included binning in the calculation of the wavelength solution
Corrected messages and conditions under which the prefix for cosmic ray rejection is used
Image combination call and messages
Other additions + Added lookup table
dcr.parfile generator and found optimal parameters for Red camera and binning 2x2
Moved user manual from external repo to
Added version checker
Centralised metadata (
__licence__, etc) in
First production ready release