Automate lang links #6618

rikivillalba · 2024-11-16T19:46:20Z

In response to #6613

This is my proposal to automate link generation:

The logic is in .translation_links.R in the vignettes directory, and it is sourced in each vignette
The message "Translations of this document are available in the following languages" is translated in the same source (there is a switch which defaults to english)
Currently the output format is en | fr | es | ..., that is, not a list but all inline.

Note that currently, in documents with a TOC (datatable-faq.Rmd) the link list is below the toc in the packaged vignettes. That is unchanged as the TOC is built automatically by commonmark.
Also, #6394 should be considered.

codecov · 2024-11-16T19:53:05Z

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.60%. Comparing base (546259d) to head (7c661b4).
Report is 3 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #6618   +/-   ##
=======================================
  Coverage   98.60%   98.60%           
=======================================
  Files          79       79           
  Lines       14518    14518           
=======================================
  Hits        14316    14316           
  Misses        202      202

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

tdhock · 2024-11-19T14:09:20Z

vignettes/.translation_links.R

+block = sprintf(
+  "%s: %s\n", 
+  switch(lang,
+    fr = "Des traductions de ce document sont disponibles dans les langues suivantes",


This method for defining translations is non-standard. I think this would be better as gettext(...)
or to simplify, just use English. Only list the other languages at the top of the English vignette. and then the other vignettes can just link back to English.

Yes, I think in gettext however I don't figure out how to get the string for the appropriate language to compile the vignettes in all languages, other than set Sys.setLanguage() in each call.
Another option is to hardcode the header and let the code generate only the links.
But I think your second option is ok.

rikivillalba · 2024-11-23T22:43:04Z

I made the following changes:

The script is now is a function to which the "translations available" message is an arg, so it is not harcoded but passed from the vignette text.
The message only prints if there are any translations other than current language (as now french version of datatable-joins.Rmd is missing)
renamed .translation_links.R to _translation_links.R so ls can see it (it is part of the vignettes after all)
NOTE: I manually merge datatable-keys-fast-subset.Rmd, datatable-reference-semantics.Rmd and datatable-secondary-indices-and-auto-indexing.Rmd because Using hyperlinks instead of vignette() calls for readability; doing the same for vignette titles without links for consistency #6617 modified lines adjacent to this branch.

_translation_links.R must be bundled in the package so r CMD check can build the vignettes. However that source is 'knitted' and present in each .R file generated from .Rmd and installed in doc.

French vignettes are no modified at all. An equivalent to "from English original vignette" message linking back to english might be added to translated vignettes, or code similar to that of the english ones, it is up to the owner to decide. In the script case it will add all available links excluding that of the current vignette language determined from the folder name.

As part of spanish translators there is work in progres in https://github.com/cienciadedatos/traduccion-vignettes-datatable
tks

MichaelChirico · 2024-12-06T07:31:57Z

vignettes/_translation_links.R

+# build a link list of alternative languages (may be character(0))
+.write.translation.links <- function(fmt) {
+    url = "https://rdatatable.gitlab.io/data.table/articles"
+    path = dirname(knitr::current_input(TRUE))


is there a litedown equivalent (#6583)?

MichaelChirico · 2024-12-06T07:33:06Z

vignettes/_translation_links.R

+      pattern = glob2rx(knitr::current_input(FALSE))
+    )
+    transl_lang = ifelse(dirname(translation) == ".", "en", dirname(translation))
+    block = if (length(which(transl_lang != lang))) {


Suggested change

block = if (length(which(transl_lang != lang))) {

block = if (!all(transl_lang == lang)) {

MichaelChirico · 2024-12-06T07:34:55Z

vignettes/_translation_links.R

@@ -0,0 +1,25 @@
+# build a link list of alternative languages (may be character(0))


Suggested change

# build a link list of alternative languages (may be character(0))

# build a link list of alternative languages (may be character(0))

# idea is to look like 'Other languages: en | fr | de'

Just commenting on expected output to facilitate reading/future refactoring

MichaelChirico · 2024-12-06T07:36:45Z

vignettes/_translation_links.R

+          transl_lang[transl_lang != lang],
+          file.path(url, sub("(?i)\\.Rmd$", ".html", translation[transl_lang != lang]))
+      )))
+    } else ""


WDYT about leaving unlinked text in for the current language instead?

Available languages: fr | es

vs

Available languages: en | fr | es

MichaelChirico · 2024-12-06T07:37:55Z

vignettes/datatable-benchmarking.Rmd

-* [French](https://rdatatable.gitlab.io/data.table/articles/fr/datatable-benchmarking.html)
+```{r echo=FALSE, file='_translation_links.R'}
+```
+`r .write.translation.links("Translations of this document are available in: %s")`


IIUC the idea is the translation team will just translate this string themselves directly in the .Rmd file, is that right? (seems fine, just want to confirm my understanding)

MichaelChirico · 2024-12-06T07:39:42Z

vignettes/_translation_links.R

we also have this helper script, let's keep a consistent naming style (personally, I like your snake case style here):

https://github.com/Rdatatable/data.table/blob/master/vignettes/.check.translations.R

Also note that this should probably be added to .Rbuildignore? Maybe just ignore all vignettes/_.* files?

rikivillalba mentioned this pull request Nov 16, 2024

translated vignette proposal : all pages link back to EN root vignette #6613

Open

tdhock reviewed Nov 19, 2024

View reviewed changes

rebase from master

7c661b4

rikivillalba force-pushed the automate-lang-links branch from fa2522b to 7c661b4 Compare November 23, 2024 21:57

MichaelChirico added the translation issues/PRs related to message translation projects label Dec 3, 2024

MichaelChirico reviewed Dec 6, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Automate lang links #6618

Automate lang links #6618

rikivillalba commented Nov 16, 2024

codecov bot commented Nov 16, 2024 •

edited

Loading

tdhock Nov 19, 2024

rikivillalba Nov 19, 2024

rikivillalba commented Nov 23, 2024

MichaelChirico Dec 6, 2024

MichaelChirico Dec 6, 2024

MichaelChirico Dec 6, 2024

MichaelChirico Dec 6, 2024

MichaelChirico Dec 6, 2024

MichaelChirico Dec 6, 2024

	block = if (length(which(transl_lang != lang))) {
	block = if (!all(transl_lang == lang)) {

		@@ -0,0 +1,25 @@
		# build a link list of alternative languages (may be character(0))

	# build a link list of alternative languages (may be character(0))
	# build a link list of alternative languages (may be character(0))
	# idea is to look like 'Other languages: en \| fr \| de'

Automate lang links #6618

Are you sure you want to change the base?

Automate lang links #6618

Conversation

rikivillalba commented Nov 16, 2024

codecov bot commented Nov 16, 2024 • edited Loading

Codecov Report

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rikivillalba commented Nov 23, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codecov bot commented Nov 16, 2024 •

edited

Loading