zandergordan / pub2s3 Goto Github PK

use_readmermd()

Or perhaps also copy it to the clipboard, though that is potentially dangerous if users have something important on their clipboard. Probably best to just print it to the console. Oh actually, rstudioapi::showDialog() with the url option is perfect for this.

The package is currently setup to publish each report into the top-level of the S3 bucket. It may be desirable to allow users to specify a subdirectory of the bucket, for instance organizing reports by project or time-period.

Relatedly, the Addin could also automatically update the index.html each time it publishes a new report. This would be done by grabbing the existing index.html (if any), and adding the name of the new report as a new entry.

In some cases, rendering a report is computationally expensive. Consistent use of the cache chunk option avoids this, but that may not always be the case. In case the user has already used the "knit" button to compile the report, it will be good for the addin to detect the existence of a compiled report, and offer to just upload that, rather than rendering again.

Perhaps easier than making multiple dialog boxes, perhaps this can just be a separate function, so that the package installs 2 Addins: "Publish" & "Render and Publish". In this case, "Publish" would return an error if there is no already-rendered version.

Ideally, there would be some way to detect if the current source code of the active .rmd is consistent with the previously-rendered version, but that sounds dicey.

The way I use the file.path function currently to form the S3 object name means that on windows, the object name will be invalid, containing backslashes instead of forward slashes.

Currently, it is hard-coded into the Addin that the user is trying to publish an rmarkdown report with an HTML output format. It is easy to imagine PDF being desired instead, and even .docx format could be accommodated as well.

Ideally, the prompt would offer users a radio button or dropdown box to select the bucket they want. It seems this is not possible, with rstudioapi, so I would have to turn the Addin into a shiny gadget to make this happen, which is more complexity than I want.

I think it may be best to just change the message provided in the current prompt to something like "If you are not sure what bucket to use, please hit cancel and run aws.s3::bucketlist() to view your currently accessible buckets". Changing buckets may not be that common anyway, so storing the last used bucket as a preference is probably good enough.