Closed lboeman closed 2 years ago
wholmgren
commented
2 years ago Hmm, that's tough. I see your point, but I also don't want to ask someone to click on 10 different links. How do we create the zips with the html/pdf reports and signature? And could we do both?
lboeman
commented
2 years ago How do we create the zips with the html/pdf reports and signature? And could we do both?
The full report zips are created by making a request to the dashboard: https://github.com/SolarArbiter/solarforecastarbiter-dashboard/blob/40fee97b807ad32185772fce9242869ef96181ac/sfa_dash/blueprints/reports.py#L205-L238
By both do you mean zip up the report and timeseries values together? We could do that, and have a separate endpoint that just zips the values that would be accessible via the current download link.
wholmgren
commented
2 years ago By both do you mean zip up the report and timeseries values together?
No but I guess that's another option. I meant could we allow users to download time series one by one or click a download all button. The download all could result in either 10 downloads or a single zip.
lboeman
commented
2 years ago Here's the next iteration on this. There's a collapsible <details> element that contains a table of timeseries names, and checkboxes. When the user clicks "download csvs" one csv is downloaded per checkbox.
wholmgren
commented
2 years ago Nice!
Should be "Download CSVs" (no capitalization on the last s). Or "Download CSV files". We could make it even more clear with something like "Download N CSVs" where N = number of checked boxes and is set by a small javascript function. Or are we overthinking a tiny detail?
lboeman
commented
2 years ago Should be "Download CSVs" (no capitalization on the last s). Or "Download CSV files". We could make it even more clear with something like "Download N CSVs" where N = number of checked boxes and is set by a small javascript function. Or are we overthinking a tiny detail?
I think that's a good idea so I've implemented it and set the "Download CSVs" button to disabled with no download boxes checked, and prompts the user to select CSVs like so:
None checked/ Default:
Some checked:
wholmgren
commented
2 years ago And what's new entries please
lboeman
commented
2 years ago Added the what's new, but I realized this will work poorly for distributions, so we should hold off on merging. I'll take a look at it tomorrow morning, I think it won't be too bad.
lboeman
commented
2 years ago Alrighty, Got the distribution downloads working and they look like so:
timestamp,Prob(f <= x) = 2.0%,Prob(f <= x) = 5.0%,Prob(f <= x) = 10.0%,Prob(f <= x) = 20.0%,Prob(f <= x) = 30.0%,Prob(f <= x) = 40.0%,Prob(f <= x) = 50.0%,Prob(f <= x) = 60.0%,Prob(f <= x) = 70.0%,Prob(f <= x) = 80.0%,Prob(f <= x) = 90.0%,Prob(f <= x) = 95.0%,Prob(f <= x) = 98.0%,Prob(f <= x) = 99.0%
2020-08-13T07:00:00-06:00,4.0665,7.9438,13.9053,25.5227,52.2997,61.8447,77.3958,89.075,90.595,93.1887,97.8092,102.633,107.499,109.121
docs/source/whatsnewfor all changes. Includes link to the GitHub Issue with:issue:`num`or this Pull Request with:pull:`num`. Includes contributor name and/or GitHub username (link with:ghuser:`user`).Removes the
<br>tags included in csv headers for report timeseries downloads, and separates each timeseries into a separate download. Triggering a bunch of subsequent data downloads is pretty jarring, so perhaps inserting links to download each one individually would be better? We could use something like JSZip to create a zip, but I that is complicated by the need to supply the library in downloaded copies and include it in dashboard rendered reports.@wholmgren What do you think about adding links to download each timeseries individually?