Howto:Change the language of reports
Explanation of how you can produce reports in a language different from the language you installed your operating system in.
- 1 Introduction
- 2 Run GRAMPS in a different language
- 3 What about dates and calendars
- 4 How to install a locale
- 5 Run the report
If you have an international family, chances are you want to produce reports in the language used by your distant family members.
GRAMPS is supported in many languages. You can produce reports in all these languages, by running GRAMPS in this language and making the report.
Run GRAMPS in a different language
First, let us learn how to run gramps from a terminal or console (the so called CLI-Command Line Interface, which you find typically in the System Menu of your task bar).
This can be done by typing
gramps, as GRAMPS typically is installed in the directory
/usr/bin. You can also start GRAMPS with
/usr/bin/gramps, although this might be different on other distributions.
This command calls up GRAMPS with the default locale setting of your PC, that is, the language in which you installed your linux OS. If other locales are installed on your system, calling up GRAMPS in this language can be done. To check if there are other installed locales, use:
For example, if the locale nl_BE is installed with encoding nl_BE.UTF-8, you can start it with:
Once you get the locale working and to make things handy going forward, you can create an alias. Open the file
.bashrc (this file is located in your home directory) in a text editor. At the bottom of the file add the following:
alias gramps_nl=' LANG=nl_BE.UTF-8; /usr/bin/gramps'
You will need to make the alias command unique, so in this instance I have used
gramps_nl. Now save the file. The UNIX terminal will need restarted for the new setting to take effect. Enter the alias command
gramps_nl to run gramps in another locale. Multiple alias commands can be setup for multiple locales.
Alternatively you can add a new item to the menu of GNOME or KDE, so this command needs not to be entered in a terminal, but can be called from the menu.
Add KDE Menu Item
- Right Click on
- Click on
- Click on
'New Item', Enter <code>'GRAMPS Genealogy System (nl)'
In each of the fields enter:
GRAMPS NL Locale
Manage genealogical information, perform genealogical research and analysis
Leave all other options as default. To add the GRAMPS icon, click on the icon beside Name and Description fields
'Other icons'-> Click on
- Navigate to directory
- Select image
In the KDE Menu Editor select
'Save' and close the window.
To Launch the locale version of GRAMPS click on K->APPLICATIONS -> OFFICE -> GRAMPS NL Locale
Add GNOME Menu Item
To add a Gnome Menu item, you need to have the menu editor installed; if not installed, (in Fedora) type
yum install alacarte
in a shell and restart.
The process is similar as for KDE above; right-click on the menu and choose "Edit Menus", go to the desired location in the menu tree and add a "New Menu Item". In the launcher properties window enter:
- Name: [Whatever you like; e.g.
Gramps - SV]
env LANG=sv_SE.utf8 /usr/local/bin/gramps
- Comment: [Whatever you like; e.g.
Starts Gramps in Swedish locale]
Locale parameters are exported using the keyword env, and my installation of Gramps resides in /usr/local/bin as opposed to /usr/bin
Add Windows OS Menu Item
On Windows you need to run the command line (cmd.exe) and type two lines similar to this:
SET LANG=nl_BE.UTF-8 "C:\Python25\pythonw.exe" "C:\Program Files\gramps\gramps.py"
This can be saved to a .bat file for easy execution.
What about dates and calendars
Dates and calendars depend on your locale, so changing the LANG variable will run GRAMPS with the date and calendar appropriate to the locale (if implemented).
How to install a locale
For the above to work, you must have the needed locales installed. This will be distribution specific. Add to this wiki if you know how to do this in a distribution not listed.
Before the overview, it is advised to also install the gnome language package for the locale you want, eg for locale nl you need to install the package
language-pack-gnome-nl. The KDE package (kde-i18n-nl) is not needed as GRAMPS uses gnome, even when run in KDE.
Ubuntu and variants
6.06 and later
/usr/share/i18n/SUPPORTED lists all possible locales. Eg, for nl_BE we can choose from nl_BE.UTF-8 UTF-8, nl_BE ISO-8859-1, nl_BE@euro ISO-8859-15, nl_NL.UTF-8 UTF-8, nl_NL ISO-8859-1 and nl_NL@euro ISO-8859-15. We advise to use UTF-8 encoding, so here, nl_BE.UTF-8 UTF-8 is the one we need to install.
Now, we backup the old locale file, and add the required locale. In a terminal you can backup with the commands:
mkdir ~/backup/locales sudo cp /var/lib/locales/supported.d/local ~/backup/locales/local.old
With your favorite editor, you now need to change the
local file as root, by adding the locale you want. So for nl_BE this would be adding the line nl_BE.UTF-8 UTF-8 at the bottom. Watch out: end with a newline, so put an ENTER after this line. Next we need to regenerate the locales:
sudo dpkg-reconfigure locales
This command must not generate errors; if it did, you did something wrong.
5.10 and before Execute
sudo dpkg-reconfigure locales
and select from the choices given the locales you need. Using UTF8 is advised, eg for nl_BE: choose nl_BE.UTF-8
Run the report
Having started GRAMPS in another language, all reports you make will also be in this language. Of course, things you typed in yourself are not translated.
Tips for international use
As what you type is not translated, some tips:
- for places: use the actual placename, for example München and not Munich. You could add an alternate location to the place with the translated name. (Or wait until the better place handling feature is fully implemented and available).
- for notes: use the languages the people who are most interested in this section will want it in. This version enables multiple notes, so you are able to make different notes, one per language, with the possibility to use filters to determine what appears
- for descriptions: Try to use built-in description names as much as possible, as those are translated. Eg, try to avoid custom events, and use the builtin events of GRAMPS as much as possible.
Automatically translate names and places
From Gramps version 3.2, GRAMPS doesn't support language tagging of alternative names --- whether for a person or a location. That is, you can add multiple alternative names, each in a different language, but still only one of them will be the primary one --- and will be the one picked up for a report. Nevertheless, when exporting to a relative not speaking your language, or when preparing a report for a part of your family having different language preferences than yourself, you need to change the primary language, but only in a temporary manner --- so that you still continue maintaining your own DB in your preferred language.
Luckily, the simplicity of the GRAMPS XML format allows the use of a small XSL stylesheet, fed by a local translation table of the names you want translated, which you can use for this purpose. Here's how you can easily do it on Linux:
- Ensure your administrator installs the "xsltproc" package.
- Export your GRAMPS DB (without media) in the .gramps format (i.e., GZipped GRAMPS XML). Save it (let's call it Smith.gramps). Ensure you enable the required censorship on the database (such as restrict the data on living people, etc.)
- Store the translation XSL (gramps-translate-names.xsl) near the Smith.gramps. Look at the XSL source for advanced parameters allowing you to customize the translation.
- Create the translation dictionary (t.xml) near it as well:
<?xml version="1.0" encoding="UTF-8"?> <dict xmlns="http://tarunz.org/xml/dict/0.1.0/"> <w> <s lang="ru">Хана</s> <s lang="ru">Ханна</s> <s lang="de">Hane</s> <s lang="en">Hannah</s> </w> </dict>
- Populate it with the translations for all the stems you want translated. The translation breaks down names into stems using a customizable delimiter set. This allows translation of compound names, like Menachem-Mendl, using the same rules, as given for the simple names comprising the compound, i.e., Menachem and Mendl in this case.
- Run the translation, using the xsltproc command. To just use the defaults (Russian to English translation), do this:
gunzip < Smith.gramps | xsltproc gramps-translate-names.xsl - |gzip > en.Smith.gramps
- See the XSL source for more advanced parameters. E.g., if you want to collect the stems missing translation into a file "missing", translate into German, and use "[LIVING]" as the censorship marker, the command line will be something like this:
gunzip < Smith.gramps | xsltproc --stringparam living '[LIVING]' --param log-missing 1 --stringparam lang-to de gramps-translate-names.xsl - 2> missing |gzip > de.Smith.gramps
- Use the "missing" file obtained above to collect stems to add to the translation dictionary.
- You can use the following Makefile stanza to facilitate your processing:
GZIP=gzip GUNZIP=gunzip XSLT=xsltproc TRANSLATEXSL=gramps-translate.xsl en.%.gramps: %.gramps $(GUNZIP) < $< | $(XSLT) --stringparam living '[LIVING]' $(TRANSLATEXSL) - | $(GZIP) > $@
- Create a new database in gramps, and import the resulting en.Smith.gramps into it. Review the database, and now produce, using the language translated into, all the reports, export into GEDCOM, etc.
- GRAMPS XML version is hardcoded in the XSL -- make sure it matches yours. "zcat *.gramps|head" will give you the GRAMPS XML format version used in that GRAMPS file. The script works has been tested either 1.3.0 (Gramps v3.2) or 1.4.0 GRAMPS XML (Gramps v3.3).