Incorporate data in Releases instead of GDrive

[achafetz]

Leverage the use of GitHub releases to store the publicly available UNAIDS data in an accessible way.

Based on Tan Ho's presentation at RStudio Conf, posted updated data to releases (not being indexed and not counting against repo file size).

We can post data in different formats and download easily as well. Ideally posting in a compressed format for quick upload/download using piggyback

new_tag <- "v2022.07.27"
pb_new_release(tag = new_tag)

glamr::temp_folder()

readr::write_csv(df_est, file.path(folderpath_tmp, "UNAIDS_HIV_Estimates.csv"))
readr::write_csv(df_est, file.path(folderpath_tmp, "UNAIDS_HIV_Estimates.csv.gz"))
saveRDS((df_est, file.path(folderpath_tmp, "UNAIDS_HIV_Estimates.rds"))

readr::write_csv(df_tt, file.path(folderpath_tmp, "UNAIDS_TestTreat.csv"))
readr::write_csv(df_tt, file.path(folderpath_tmp, "UNAIDS_TestTreat.csv.gz"))
saveRDS((df_tt, file.path(folderpath_tmp, "UNAIDS_TestTreat.rds"))

#upload multiple files
list.files(folderpath_tmp, full.names = TRUE) %>% 
  pb_upload(tag = new_tag)

#download a specific file
pb_download(file = "UNAIDS_HIV_Estimates.csv.gz",
            dest = glamr::si_path("path_downloads"))

• [ ] Use tag for versioning with date of release.
• [ ] If we are just using piggyback to post, we can add it as Suggests not Imports in Description file. If not using piggyback for downloading (minimize dependencies), mimic something similar to raw_from_url() (if using csv/csv.gz)
• [ ] Need to test if included in a package, if we need to specify the repo in the download (USAID-OHA_SI)
• [ ] Probably want to store as a compressed csv (csv.gz) for accessibility? or post both csv.gz and rds (compressed)?

@karishmas26 @tessam30

[tessam30]

@achafetz - I think we'll still need a gdrive version of the UNAIDS data for the Tableau products. These use hit the google sheet as a data source.

[achafetz]

I think that's completely reasonable. Continue leveraging the current model for Tableau (maybe in the future DDC 🙄) and we leverage GH releases for R.