Is your feature request related to a problem? Please describe.
Users want to download data for their own use and do not need a web plot / table.
This would also assist with performance.
Describe the solution you'd like
Have the default view be the "metadata" tab and when the user has not selected table or plot yet have it directly call the timeseries endpoint for streaming the CSV (when that is ready) or change the mimetype to application/csv so that it triggers a download from the API.
Currently we pull the data into memory, create an invisible a-tag and fill it with the base-64 encoded data to generate an in-browser CSV.
Describe alternatives you've considered
Do not load any of the tabs and have a simple message that says "click download to download or select a tab to visualize"
Additional context
If a user selects 15 timeseries IDs for full period of record this could drag down the plot currently. The table is now updated to virtualize and probably would handle it.
Also having it go straight to a file could assist with memory if we were dealing with very large datasets.
Is your feature request related to a problem? Please describe.
Users want to download data for their own use and do not need a web plot / table.
This would also assist with performance.
Describe the solution you'd like
Have the default view be the "metadata" tab and when the user has not selected table or plot yet have it directly call the timeseries endpoint for streaming the CSV (when that is ready) or change the mimetype to application/csv so that it triggers a download from the API.
Currently we pull the data into memory, create an invisible a-tag and fill it with the base-64 encoded data to generate an in-browser CSV.
Describe alternatives you've considered
Do not load any of the tabs and have a simple message that says "click download to download or select a tab to visualize"
Additional context
If a user selects 15 timeseries IDs for full period of record this could drag down the plot currently. The table is now updated to virtualize and probably would handle it.
Also having it go straight to a file could assist with memory if we were dealing with very large datasets.