3.0.2 and 3.0.3¶
Note that 3.0.2 was found to have one broken recipe, 3.0.3 fixes it.
- Continue without crashing when
traceAperturescannot identify a starting location for a trace.
- Fix issues with assignment of on-source/sky frames when the user specifies specific frames.
- Fix bug where
stackFramescrashed if using the
statsecparameter when scaling or zero-offsetting.
- In fringeCorrect,
do_cal=forcehas been reactivated.
- Better handling of infinites and NaN in the flat normalization.
- Added new primitive to the recipes to mask amplifier 5 in GMOS-S data obtained since January 28, 2022. GMOS-S amplifier 5 suffered a major failure and it is not usable.
- Ensure that the masks are used when calculating the statistics in scaleByIntensity.
- Added missing support for YPHOT filter.
- Support of the Flamingos 2 filters.
** geminidr **
write1DSpectrato configure the output
- Under-the-hood modification to distinguish data reduced in quicklook mode versus science mode.
- Internal Gemini catalog server URL updated.
- Various fixes to the documentation affecting formatting, not the content.
- Fix bug where
sectionstart/end comparison was made on string, not numeric, values.
- Fix bug that caused longslit spectra to have incorrect WCS, offset from true slit location.
use_snr=Falsethe default, and estimate noise from pixel-to-pixel variations, regardless of its value.
- Various fixes to the documentation.
This release includes new support for GMOS longslit data. Reduction of GMOS longslit data is offered only quicklook mode. It does not produce science quality outputs, yet.
- In imaging mode, the science recipes now include a call to
scaleByExposureTimebefore the stacking step. It is now possible to stack frames with different exposure times.
- Fix the GCALLAMP tag for NIR data to include the QH lamp.
- Remove incorrect logging in separateSky when object and/or sky files are specified.
- Improve algorithm for separating on-source and on-sky frames.
- Avoid upsampling OBJMASK from uint8 to uint16
- In near-IR imaging mode, frames that fail to be sky subtracted are removed from the main reduction stream to avoid contamination. The reduction continues with the “good” frames. If all frames fail the sky subtraction, then all frames will be passed to the next step of the reduction.
- Fix to the calculation of the CC-band used in nighttime sky quality assessment.
- Fix to the calculation of the BG-band used in nighttime sky quality assessment.
- Ensure NIRI skyflats satisfy calibration association requirements
- Fix a Python 3 compatibility issue.
- Quicklook (
--qlmode) reduction support for GMOS longslit data.
remove_firstparameter to removeFirstFrame primitive.
match_radiusparameter to adjustWCSToReference primitive.
- Add an IRAF compatibility primitive and recipe for Flamingos 2.
astrodata and recipe_system
- Provenance history stored with the data in tables named: PROVENANCE and PROVHISTORY.
do_flatinput parameters have been replaced with
do_calwith more options than True or False. Use
showparsto inspect the options.
- Python 2 support has been dropped. Starting with v3.0.0, DRAGONS requires Python 3. All tests were run on Python 3.7, and this version of Python now serves as the minimal required version.
- Improved the F2 processed products backward compatibility with Gemini IRAF.
- Fix various links in the documentation.
- Add examples and cross-reference to disco-stu usage documentation.
- New tutorial for the quicklook reduction of GMOS longslit data.
- Fix a crash when a section was used when stacking.
- Add missing third party adpkg and drpkg support to utility scripts dataselect, showpars, typewalk, and showrecipes.
- Fix to Jacobian calculation for non-affine transforms
- Make adcc more robust to missing connection to fitsstore.
- Add compatibility with sigma_clip for astropy v3.1+
- Add IRAF compatibility keywords on GMOS mosaiced data.
- Add compatibility with astroquery 0.4.
- Add compatibility with sigma_clip fro astropy v3.1+
- Add IRAF compatibility recipe.
- Various fixes to documentation and instruction manual following feedback from users.