- Initial order x-positions are now dynamic and stored in the db rather than being hard coded.
- Updated how we fit the profile center/width to better fit faint traces
- We now have prefer calibrations in the following order: same block, same proposal, any public calibration.
- If a block is still going, we delay the processing in case there is a calibration taken at at the end of the block that we can use for processing
- We now only use arcs and flats taken with the same slit width as the science data
- Simplified the fitting for refining the peak centers. We no longer try to fit them all simultaneously
- Updated the used the line list to remove a less isolated arc line
- Fixes to the quality of the reductions
- We now trim the edges of orders better to remove artifacts
- Added the ability to combine the extraction from both orders into a single spectrum
- Numerous fixes based on commissioning experience
- Added/refactored the necessary logic to re-extract the data via the UI
- Fixes based on Joey's comments
- Deployment fixes
- We now don't keep the filepath of the standards in the db. We assume they are in the archive or are in the install director
- Increased the memory limit on the containers to accomodate stacking
- Simplified the triggering flat stacking to make it more testable
- Deployment fixes 0.6.0
- Deployment fixes
- Helm value fixes to get the pipeline scheduled on nodes.
- Helm chart fixes.
- Full alpha release
- Added in test data from Siding Springs
- Added order detection functionality
- Initial Release