XSLT to the Rescue

XSLT TO THE RESCUE

The immense amount of work that I have had in the recent months kept me away from this blog for much longer than I would wish, but also forced me to brush up some skills that I started acquiring long ago.

I had so much work that I was forced to subcontract a large part of it. One of the projects that I needed to subcontract required that the translator followed the provided terminology strictly.

I had a multiterm termbase, but how do I provide that to a translator who only uses Cafetran?  The termbase had entries in many languages and also entries flagged as blacklisted.

Fortunately, you can export the whole termbase from Multiterm in tbx (termbase exchange) format. Cafetran can read TBX, but the file still contained all the redundant languages and blacklisted terms, and as much as I like Cafetran, it is too dumb for that. But there is this thing called XSLT, which is used to transform XML files into other XML files or other formats. And TBX is basically XML inside.

After a brief look at the TBX file helped me determine what I wanted to keep and what to skip. The following code did the job:

xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
 <xsl:output omit-xml-declaration="yes" indent="yes"/>
 <xsl:strip-space elements="*"/>
 <xsl:template match="node()|@*">
  <xsl:copy>
   <xsl:apply-templates select="node()|@*"/>
  </xsl:copy>
 </xsl:template>
 <xsl:template match="langSet[not(@xml:lang='EN-GB' or @xml:lang='PL-PL')]"/>
 <xsl:template match="tig[descripGrp/descrip[@type='Forbidden']='True']"/>
</xsl:stylesheet>
I'm not much of an XSLT guru, and I found what I needed at Stackoverflow. The only real work I did was to figure out what to put in the last two elements before the closing tag. What this stylesheet does is to make a copy of the TBX file, unless the conditions stated in the last two elements are matched. And these conditions are: Skip langSet elements unless they contain English or Polish, and then skip tig elements, if they contain forbidden vocabulary.

To make any practical use of this stylesheet you need a tool that understands XML and can perform an XSLT transformation. If you are a translator, then Okapi Rainbow is a good choice. It is a free toolkit that offers many other functionalities useful for translators. The file to be transformed needs to be loaded on the first tab, and load the XSLT file under Utilities > XML Utilities > XSL Transformation... Shortly after clicking Execute you should get your transformed file, with a file name similar to the original, provided that your stylesheet is correct.

All in all, XSLT is a powerful tool, and if you have a knack for these things (I mean writing a bit of code), it can help you a lot.

Comments

Post a Comment

Popular posts from this blog

Python 3 script for Deepl Translator Pro Among the MT (machine translation) providers Deepl is the new kid on the block, but it has been making waves ever since it appeared. Frankly speaking, as soon as I found out how good the Deepl MT is, I actually abandoned Google Translate. If I am to pay for a service, I'll pay for the one which delivers better results. Before Deepl announced the paid Pro service, there was (and still is) the free Python module called pydeepl. It used to work until recently, but the ‘backdoor’ making it possible to programmatically get free translations in the era of the paid service has recently been plugged. (For the record, I subscribed to the paid service as soon as it appeared). There also is a Deepl Translator plugin for SDL Trados Studio. It is quite simple and almost no settings, except for the API key, can be entered. My Python script for pydeepl was more sophisticated, for example I added a replacement table for cases where I did not like Deep
Read more
Image
 Skrypty w Pythonie Ułatwiające wypełnianie JPK_VAT7M od jpklibre (Oj dawno nic nie pisałem na blogu!) Do niedawna robiłem JPK za pomocą arkusza w Excelu, ale na nowym komputerze Excel 64-bitowy, więc arkusz nie działa. Dlaczego – nie będę się rozpisywać. Pora na rozwiązanie uniwersalne, które będzie działać niezależnie od systemu operacyjnego – skoroszyt  jpklibre , bo komputerów z Windowsem mam 1, a z systemem Linux – kilka. Można oczywiście wypełnić arkusz ręcznie, ale to jest mozolne. Można też poddać w wątpliwość sens arkuszy, gdy pliki JPK powinno generować oprogramowanie księgowe. Owszem, moje generuje, ale uparło się przy tym robić korekty, które nic nie mają wspólnego z rzeczywistością, więc robię sam - półautomatycznie. Żeby nie wypełniać jpklibre całkowicie ręcznie, można skorzystać z rozwiązania Scriptforge, które dostępne jest "fabrycznie" dla LibreOffice Calc od wersji 7.3 LibreOffice. Trzeba tylko doinstalować rozszerzenie APSO , które ułatwi uruchamianie makr
Read more
Macros, scripts, coding In the early years of my freelancing I worked mostly in Microsoft Word. I either did not use any CAT tool or I used the free version of Wordfast (now called Wordfast Classic) which is a set of macros in itself. I was looking for ways to make my tedious work of translating in Microsoft Word a bit easier. Soon I found out about macros. Initially I created them by recording, then I learned more about coding them. Now, though most of my translation work is done outside of Microsoft Word, in CAT tools, I still have most of the macros I created more than 10 years ago. Many of them fell into disuse, but there is one that I use regularly when I need to do some work in Word. It is the macro I called ‘TextFieldFit’. When I draw a text box, I run the macro to reduce the inner margins, remove the border around the box and reduce the font size. I use text boxes to translate drawings that did not get imported into my preferred CAT tool, the SDL Trados Studio. I
Read more