Update file types supported by drive_download()

[ateucher]

Since Google Docs now supports exporting to Markdown, it seemed like a good idea to enable that support in drive_download().

The only code I changed manually was to add the default .md extension for text/markdown mimetype in data-raw/extension-mime-type-defaults.csv. Then I just ran data-raw/mime-types.R to pull in the Google MIME types and create the csv files in inst/extdata/data/.

This does work now:

library(googledrive)

drive_auth(email = "andy@andyteucher.ca")

drive_download(
  "https://docs.google.com/document/d/1EK6WZyEUfdy6sj3GjLII20Ok7qTcmRD-vv-osfHnkPU",
  path = "my_document.md",
)
#> File downloaded:
#> • 'Test for google drive' <id: 1EK6WZyEUfdy6sj3GjLII20Ok7qTcmRD-vv-osfHnkPU>
#> Saved locally as:
#> • 'my_document.md'

^{Created on 2025-07-09 with reprex v2.1.1}

[ateucher]

Converted to draft - I'm going to do a bit more work around extensions and defaults

[ateucher]

I don't think there is any more to do around adding default extensions to mimetypes. All of the changes here come from changes in mime::mimemap and the importFormats and exportFormats in Google's about endpoint and it seems sensible to just accept them.

I did however discover that if a mimetype does not have an extension in inst/extdata/data/mime_tbl.csv, .NA is appended to the given filename if that mimetype is specified in type. Eg.,

drive_download(
  "https://docs.google.com/document/d/1EK6WZyEUfdy6sj3GjLII20Ok7qTcmRD-vv-osfHnkPU",
  type = "text/x-markdown",
  path = "my_document.md",
  overwrite = TRUE
)
#> File downloaded:
#> • 'Test for google drive' <id: 1EK6WZyEUfdy6sj3GjLII20Ok7qTcmRD-vv-osfHnkPU>
#> Saved locally as:
#> • 'my_document.md.NA'

I added a fix and a test for this, so now it does not add .NA:

drive_download(
  "https://docs.google.com/document/d/1EK6WZyEUfdy6sj3GjLII20Ok7qTcmRD-vv-osfHnkPU",
  type = "text/x-markdown",
  path = "my_document.md",
  overwrite = TRUE
)
#> File downloaded:
#> • 'Test for google drive' <id: 1EK6WZyEUfdy6sj3GjLII20Ok7qTcmRD-vv-osfHnkPU>
#> Saved locally as:
#> • 'my_document.md'

^{Created on 2025-07-10 with reprex v2.1.1}

[ateucher]

The opposite also works:

tfile <- tempfile(fileext = ".md")
writeLines(
  "# Title

  text
  ",
  tfile
)

drive_auth("andy@andyteucher.ca")

drive_upload(tfile, type = "document")
#> Local file:
#> • '/var/folders/_f/n9fw7ctx3fqf2ty9ylw502g80000gn/T//Rtmpk9yyHy/file82652c912409.md'
#> Uploaded into Drive file:
#> • 'file82652c912409.md' <id: 1APiJm3DzCRdq2WH8_X2SE5xCDYxyOU1lVYjVmoa30ns>
#> With MIME type:
#> • 'application/vnd.google-apps.document'

^{Created on 2025-07-10 with reprex v2.1.1}

[jennybc]

text/x-markdown appears to be a rather weird/legacy MIME type, but if it's one of the official ones (and it is), I agree we need to plan for it. This is a note to myself.

[jennybc]

Thanks! As for my commits, this was mostly an effort to recollect how these MIME type / file extension tables are constructed. The way I had rigged this was not terribly clear. I think it's somewhat improved -- it'll have to do for now. And we've got tests for upload/download.

[ateucher]

Great, thanks so much for finishing it off!

[jrosell]

@ateucher thanks for your PR!.

It would have been awesome if the code were not lost in the process of upload + download, but I guess it's a google issue and not an implementation issue.

pak::pak("tidyverse/googledrive")
#> ℹ Loading metadata database
#> ✔ Loading metadata database ... done
#> 
#> 
#> ✔ All system requirements are already installed.
#> 
#> ℹ No downloads are needed
#> ✔ 1 pkg + 24 deps: kept 24 [4.8s]

library(googledrive)

tfile <- tempfile(fileext = ".md")
code <- r"(

# Title

```r
1+1
```

text
)"

writeLines(code, tfile, sep = "\n")
cat(paste0(readLines(tfile), "\n"))
#> 
#>  
#>  # Title
#>  
#>  ```r
#>  1+1
#>  ```
#>  
#>  text
#> 

options(gargle_oauth_email = Sys.getenv("GOOGLE_EMAIL"))
drive_auth_configure(
  path = Sys.getenv("GOOGLE_API_CREDENTIALS_OAUTH20_CLIENT_ID_DESTOP_FILE")
)
drive_upload(tfile, type = "document")
#> Local file:
#> • '/tmp/Rtmp2Lg9bf/file27bff5042c5a8.md'
#> Uploaded into Drive file:
#> • 'file27bff5042c5a8.md' <id: 1aX6tY9RIyMiq_RB1LSDeai7judGcVhGNKi4JhEdD7f8>
#> With MIME type:
#> • 'application/vnd.google-apps.document'


drive_download(
  basename(tfile),
  path = "markdown-googledrive.md",
  overwrite = TRUE
)
#> File downloaded:
#> • 'file27bff5042c5a8.md' <id: 1aX6tY9RIyMiq_RB1LSDeai7judGcVhGNKi4JhEdD7f8>
#> Saved locally as:
#> • 'markdown-googledrive.md'
cat(paste0(readLines("markdown-googledrive.md"), "\n"))
#> # Title
#>  
#>  1+1
#>  
#>  text

^{Created on 2025-09-10 with reprex v2.1.1.9000}

[jennybc]

If I send this markdown:

# Title

```r
1+1
```

```rust
fn main() { println!("Hello, world!"); }
```

text

then re-download it, I get:

# Title

```
1+1
```

```rust
fn main() { println!("Hello, world!"); }
```

text  

The code fences are retained. But the language identifier for r is lost, where the one for rust is retained. I think this is a Google problem, because R is not available in this dropdown list of languages:

Screen.Recording.2025-09-10.at.3.26.47.PM.mov

It would be great to include R there, since it is more popular than some of these languages. But it's beyond the scope of what googledrive can do.