Basically, gh::gh_token() is not optional but a requirement for piggyback::pb_download() to successfully download the data.
I also found that when piggyback::pb_download() failed, a corrupted rds object was saved in cache.
That means in future downloads piggyback::pb_download() would not download as the corrupted file is found.
Because the download is very fast, I set overwrite = TRUE
Realated to this, I decided to to remove the if statement checking if hte tmp file exists.
In addition, please note that gh::gh_token() makes use of gitcreds::gitcreds_get() to retireve the githb_pat -> https://gh.r-lib.org/reference/gh_token.html
I tried to find there in the DockerFile you install the R deps but I was not lucky.
A modification in the Docker image must be done so that gitcreds R package is installed.
I found the reason of #131
Basically, gh::gh_token() is not optional but a requirement for piggyback::pb_download() to successfully download the data.
I also found that when piggyback::pb_download() failed, a corrupted rds object was saved in cache. That means in future downloads piggyback::pb_download() would not download as the corrupted file is found. Because the download is very fast, I set overwrite = TRUE
Realated to this, I decided to to remove the if statement checking if hte tmp file exists.
In addition, please note that gh::gh_token() makes use of gitcreds::gitcreds_get() to retireve the githb_pat -> https://gh.r-lib.org/reference/gh_token.html I tried to find there in the DockerFile you install the R deps but I was not lucky. A modification in the Docker image must be done so that gitcreds R package is installed.