R packages use a text file format with a .po extension to store translations of messages, warnings, and errors. poio provides functionality to read in these files, fix the metadata, and write the objects back to file.
To install the development version, you first need the devtools package.
install.packages("devtools")Then you can install the poio package using
devtools::install_bitbucket("RL10N/poio")read_po reads PO and POT files into R, and stores them as an object of class "po" (see below for details).
fix_metadata fixes the metadata in a po object.
generate_po_from_pot generates a PO object from a POT object.
write_po writes po objects back to a PO file.
get_n_plural_forms is a convenience function for retriving the number of plural forms for a language from the po object's metadata.
language_codes is a list of all the language codes and the country codes that gettext understands.
plural_forms is a data.frame of the plural forms header string for over 140 common languages.
A typical workflow begins by generating a POT master translation file for a package using tools::xgettext2pot. In this case, we'll use a sample file stored in the poio package. The contents look like this:
library(poio)
pot_file <- system.file("extdata/R-summerof69.pot", package = "poio")
cat(readLines(pot_file), sep = "\n")## # This is a translator comment before the metadata.
## # Other comment types aren't useful here, and should be ignored.
## # Like the "fuzzy" flags comment below.
## #, fuzzy
## msgid ""
## msgstr ""
## "Project-Id-Version: R 3.3.1\n"
## "Report-Msgid-Bugs-To: bugs.r-project.org\n"
## "POT-Creation-Date: 2016-10-05 20:19\n"
## "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
## "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
## "Language-Team: LANGUAGE <LL@li.org>\n"
## "MIME-Version: 1.0\n"
## "Content-Type: text/plain; charset=CHARSET\n"
## "Content-Transfer-Encoding: 8bit\n"
##
## # Now commencing Bryan Adams lyrics
## # Because the song has numbers in it
## #: some_source_file.R:123
## msgid "I got my first real six-string"
## msgstr ""
##
##
##
##
## # This one gets a "c-format" flags comment
## #because it uses c-style sprintf formatting
## #, c-format
## msgid "Bought it at the %f-and-dime"
## msgstr ""
##
## # I don't think that the tools package supports generating
## # source reference comments, but we should preserve them
## # in case someone manually inserts them into their file
## #: some_source_file.R:123
## msgid "Played it till my fingers bled"
## msgstr ""
##
## # Also uses xgettextf
## #, c-format
## msgctxt "Summer as in seasons, not a function that calculates sums"
## msgid "It was the summer of '%d."
## msgstr ""
##
## # Technically the lyric says 'some' guys
## # but I want a countable example
## msgid "Me and %d guy from school"
## msgid_plural "Me and %d guys from school"
## msgstr[0] ""
## msgstr[1] ""
##
## # Testing quote escaping
## msgid "Had a \"band\"" and we tried real hard"
## msgstr ""
##
## # Obsolete messages can also have other comments
## #~ msgid "Jimmy quit and Jody got married"
## #~ msgstr ""
##
## # Countably obsolete. Apologies for bad English.
## # Also note the bad number order.
## #~ msgid "I should've known we'd never get far"
## #~ msgid_plural "I should've known we'd never get fars"
## #~ msgstr[1]""
## #~ msgstr[0] ""
To import the file, use read_po. A description on the object's structure is shown in the "PO Objects" section below.
(pot <- read_po(pot_file))## _
## Source Type: r code
## File Type: pot master translation
##
## Initial comments:
## [1] This is a translator comment before the metadata.
## [2] Other comment types aren't useful here, and should be ignored.
## [3] Like the "fuzzy" flags comment below.
##
## Metadata:
## # A tibble: 9 × 2
## name value
## <chr> <chr>
## 1 Project-Id-Version R 3.3.1
## 2 Report-Msgid-Bugs-To bugs.r-project.org
## 3 POT-Creation-Date 2016-10-05 20:19
## 4 PO-Revision-Date YEAR-MO-DA HO:MI+ZONE
## 5 Last-Translator FULL NAME <EMAIL@ADDRESS>
## 6 Language-Team LANGUAGE <LL@li.org>
## 7 MIME-Version 1.0
## 8 Content-Type text/plain; charset=CHARSET
## 9 Content-Transfer-Encoding 8bit
##
## Direct Translations:
## # A tibble: 6 × 9
## msgid msgstr is_obsolete msgctxt
## <chr> <chr> <lgl> <list>
## 1 I got my first real six-string FALSE <chr [0]>
## 2 Bought it at the %f-and-dime FALSE <chr [0]>
## 3 Played it till my fingers bled FALSE <chr [0]>
## 4 It was the summer of '%d. FALSE <chr [1]>
## 5 Had a \\"band\\"" and we tried real hard FALSE <chr [0]>
## 6 Jimmy quit and Jody got married TRUE <chr [0]>
## # ... with 5 more variables: translator_comments <list>,
## # source_reference_comments <list>, flags_comments <list>,
## # previous_string_comments <list>, msgkey <chr>
##
## Countable Translations:
## # A tibble: 2 × 10
## msgid
## <chr>
## 1 Me and %d guy from school
## 2 I should've known we'd never get far
## # ... with 9 more variables: msgid_plural <chr>, msgstr <list>,
## # is_obsolete <lgl>, msgctxt <list>, translator_comments <list>,
## # source_reference_comments <list>, flags_comments <list>,
## # previous_string_comments <list>, msgkey <chr>
tools::xgettext2pot makes a mess of some of the metadata element that it generates, so they need fixing. fix_metadata auto-guesses sensible options, but you can manually set values if you prefer.
pot_fixed <- fix_metadata(pot, "Language-Team" = "Team RL10N!")## Updating the Project-Id-Version to 'poio 0.0-3'.
## Updating the Report-Msgid-Bugs-To to 'https://github.com/RL10N/poio/issues'.
## Updating the PO-Revision-Date to '2017-01-09 20:55:22-0500'.
## Updating the Last-Translator to 'Richie Cotton <richierocks@gmail.com>'.
## Updating the Language-Team to 'Team RL10N!'.
## Updating the Content-Type to 'text/plain; charset=UTF-8'.
pot_fixed$metadata## # A tibble: 9 × 2
## name value
## <chr> <chr>
## 1 Project-Id-Version poio 0.0-3
## 2 Report-Msgid-Bugs-To https://github.com/RL10N/poio/issues
## 3 POT-Creation-Date 2016-10-05 20:19
## 4 PO-Revision-Date 2017-01-09 20:55:22-0500
## 5 Last-Translator Richie Cotton <richierocks@gmail.com>
## 6 Language-Team Team RL10N!
## 7 MIME-Version 1.0
## 8 Content-Type text/plain; charset=UTF-8
## 9 Content-Transfer-Encoding 8bit
Now you need to choose some languages to translate your messages into. Suitable language codes can be found in the language_codes dataset included in the package.
data(language_codes)
str(language_codes, vec.len = 8)## List of 2
## $ language: Named chr [1:245] "aa" "ab" "ace" "ae" "af" "ak" "am" "an" ...
## ..- attr(*, "names")= chr [1:245] "Afar" "Abkhazian" "Achinese" "Avestan" "Afrikaans" "Akan" "Amharic" "Aragonese" ...
## $ country : Named chr [1:249] "AF" "AX" "AL" "DZ" "AS" "AD" "AO" "AI" ...
## ..- attr(*, "names")= chr [1:249] "Afghanistan" "Åland Islands" "Albania" "Algeria" "American Samoa" "Andorra" "Angola" "Anguilla" ...
Then, for each language that you want to create a translation for, generate a po object and write it to file. If your current working directory is the root of your package, the correct file name is automatically generated.
for(lang in c("de", "fr_BE"))
{
po <- generate_po_from_pot(pot, lang)
write_po(po)
}po objects are lists with class "po" (to allow S3 methods), containing the following elements:
- source_type: A string. Either
"r"or"c", depending upon whether the messages originated from R-level code, or C-level code. - file_type: Either
"po"or"pot", depending upon whether the messages originated from a PO (language-specific) or POT (master translation) file. Determined from the file name. - initial_comments: A
charactervector of comments added by the translator. - metadata: A
data_frameof file metadata with columns "name" and "value". - direct: A
data_frameof messages with a direct translation, as created bystop,warning,messageorgettext; its columns are described below. - countable: A
data_frameof messages where the translation depends upon a countable value, as created byngettext; its columns are described below.
The direct element of the po object has the following columns.
- msgid: Character. The untranslated (should be American English) message.
- msgstr: Character. The translated message, or empty strings in the case of POT files.
- is_obsolete: Logical. Is the message obsolete?
- msgctxt: List of character. Disambiguating context information to allow multiple messages with the same ID.
- translator_comments: List of character. Comments added by the translator, typically to explain unclear messages, or why translation choices were made.
- source_reference_comments: List of character. Links to where the message occured in the source, in the form "filename:line".
- flags_comments: List of character. Typically used to describe formatting directives. R uses C-style formatting, which would imply a
"c-format"flag. For example%ddenotes an integer, and%sdenotes a string."fuzzy"flags can appear when PO files are merged. - previous_string_comment: List of character. When PO files are merged with an updated POT file ,and a fuzzy flag is generated, the old msgid is stored in a previous string comment.
The countable element of the po object takes the same form as the direct element, with two differences.
- msgid_plural: Character. The plural form of the untranslated message.
- msgstr: This is now a list of character (rather than character.)
The msgtools package, which has higher level tools for working with messages and translations.
The Pology python library has some useful documentation on the PO file format.
The GNU gettext utility.
This package was developed as part of the RL10N project, funded by the R Consortium.