Closed nhurleywalker closed 6 months ago
In branch 23-pandas-fits
I have updated the download page so that all the downloads now produce fits files instead of .csv files.
The .fits files have correct data types for each column but don't (yet) have any VO information.
I have duplicated the links to the Survey status and Candidate ratings table pages. An example is below.
The plan here is to update the filtering of the tables on these pages and add a second button which downloads only the data that matches the given query parameters.
Exporting data directly via SAMP is problematic due to HTTP/HTTPS issues (see 2020ASPC..527...85T). As noted in the linked article:
Users of those services will just have to save files to local disk and reload them into a suitable local application rather than use a “send to TOPCAT” button or similar. This is a loss of convenience, but since more complex uses of Web SAMP such as exchanging coordinates between web and desktop clients are uncommon, it won’t have much impact on what users are actually able to do.
In branch 23-vo-approach
I tried to implement a vo service using CASDA_VO_tools as the documentation indicated that it would provide a ready-made but customisable way of providing VO compliant SIA/TAP services based on an existing database.
After much investigation and bug fixing reporting (eg https://github.com/csiro-rds/casda_vo_tools/issues/21), I managed to get the service built and running in a docker container. However I was not able to connect to the gleamx db and run queries on it.
In branch 23-daiquiri
I tried to use the (Daiquiri)[https://github.com/django-daiquiri/daiquiri] django app which provides an ADQL interface to an existing science database.
The setup of the app was simpler than the previous approach, however I was again not able to properly connect to the gleamx db and run queries on it.
The issue with both of these approaches was incomplete, incorrect, or just confusing documentation.
We need an improved export function so that we may more conveniently and quickly examine large numbers of candidates. Ideally, this would allow us to broadcast a selection to SAMP (to be picked up by TopCat) or exported as a FITS table. This will be more time-efficient than designing in-browser viewing options.