GiellaLT provides an infrastructure for rule-based language technology aimed at minority and indigenous languages, and streamlines building anything from keyboards to speech technology. Read more about Why. See also How to get started, and our Privacy document.
The bibles are an important source of text in a language. It is even more imprtant source of parallel text, since the bible translations generally preserve the format of the text irrespectible of the target language. In this project, we have recieved bible texts for North and Lule Sámi, Bokmål, Nynorsk, Swedish and Finnish. Since the bibles have a specific format which requires separate processing than the other files, there is a documentation included.
In general..
The bible files are part of the general conversion process, which is described in the documentscorpus_conversion.html and corpus_conversion_tech.html. This conversion process is aimed to produce texts that are ready for linguistic analysis and research. The texts contain some structure, such as the sections and paragraphs. Similarly, titles, lists and tables are marked separately. The original files recieved from text producers are preserved as such and the conversion process is always started from the original file. The bibles however, have a special structure with books, chapters and verses; this structure is implemented very differently depending mostly on the type of the source file. For example, paratext is a rich description language designed for bibles but in text- or word documents, less information of the structure of the bible text is available.
One goal of the processint of the bible text is to preserve or restore the original structure of the text. The bible texts can be better compared, the text can be searched using verse and chapter number and the structural information can be used in text alignment.
The XML format of the analyzed text is basically the following:
<book title="Fyrste Mosebok">
<chapter number="1">
<section title="Gud skaper verda">
<p>
<verse number="1"> I opphavet skapte Gud himmelen og jorda. </verse>
<verse number="2"> Jorda var aud og tom, og mørker låg over havdjupet.
Men Guds Ande sveiv over vat net. </verse>
...
</p>
...
</section>
...
</chapter>
See the description of the dtd for the exact format. This format is achieved by simplifying the original structure in some cases. For example, in many bible versions, the verse is printed without the number and taken out from the paragraph. Some versions may not have subtitles, so the section layer may be missing. The paragraph-marks are not added when they are not explicitely marked in the document.
The xml-formatted bible text is stored at
corp/orig/lang/nt and ot -directories. The files are
marked with suffix .bible.xml.
Most of the bible texts are under standard license conditions, which are explained in… At the moment, the corpus hierarchy contains the following bible texts:
bible2xml.plUsage:
corpus-analyze.pl [OPTIONS] file.bible.xml
The structure of the bible text is converted to the project’s internal format. The verse-markings are removed and the verses are combined to form longer paragraphs. A sentence many times spreads over more than one verse, so when the verses are combined, there sentece structure is preserved.
(…)