search

Keck-DataReductionPipelines / KPF-Pipeline

KPF-Pipeline
https://kpf-pipeline.readthedocs.io/en/latest/
Other
11 stars 2 forks source link

Develop QLP time series #352

Closed awhoward closed 9 months ago

awhoward commented 1 year ago

We have long talked about having QLP time series to track instrument performance over days, weeks, months, and years. It's an open question how this should be displayed in Jump. It could be a set of static plots that are regularly updated. Or Jump might ingest some QLP metrics for individual observations and we could use a new plotly tab/view in Jump to plot these over time, allowing for filtering of data and arbitrary timescales.

Some metrics that we're going to want to track are:

Thoughts on the implementation of this (and what we should track), @bjfultn?

I'm tagging @fdai-planet because he has been the architect of the QLP to date, but he need not be the one who implements time series features.

awhoward commented 1 year ago

@shalverson, what else should we track?

bjfultn commented 1 year ago

I think that the plots tab on jump was built for this purpose. We just need to get all of the data into the DB. Header values are preferred, but I can pull from the file contents as well if needed.

awhoward commented 1 year ago

You are right, of course, @bjfultn! Should these metrics be part of the L2 files? Most seem analogous to the RVs in that they are distillations of the L1 spectra. The CCD characteristics are an atypical case in that L1 files aren't created for these files (I think) so could we go L0 -> 2D -> L2 for those?

bjfultn commented 1 year ago

The task here is to start ingesting data into jump where we have it and put more measurements into the products where needed.

awhoward commented 1 year ago

@shalverson, let's make a list of parameters to ingest.