The idea is to convert BPT into some other format, for example:
An important purpose of BPT is to make texts uniform. This is especially useful when working with multiple authors and editors, and/or in multiple formats.
This also means that texts from word processors or other text editors needs to be converted into BPT. Given the variety input-side this will often required manual correction, preferably by the authors themselves.
Conversion is done automatically, i.e with a Python script. Of course the result is checked and can be changed or corrected manually.
In the heart of the pipeline sits the BPT converter. The pipeline produces artifacts (in this case, CTS compliant TEI XML files) which can be examined and downloaded from the jobs page in the data repo. The repo’s pipeline badge indicates the status of the pipeline and links to the pipelines page of the data repo. This feedback gives the data team a tool for quality control.
The BPT converter is a script. Likewise, other moments in the workflow may be automated, e.g. checking the validity of the BPT or other formats, or checking the well-functioning of an online application.
Plain text can be version-managed with git. Environments like GitLab or GitHub also make it possible to automate the workflow as a whole, in a so-called pipeline. An edit in a document could, via a commit, lead to a triggering of the workflow and so, for example, to an immediate change online.
However, it makes sense to check things manually, certainly before publishing entire texts online.
A set-up like this allows a division of labour.