- Top Results
I would like to share with you my small project for translation of Event messages from OSD files. It uses DeepL translator, which translates texts using artificial neural networks. The results are more accurate then from other translators like Google translate.
DeepL provides API and Python library for access to its services.
How it works
I created small Python application which works like this:
- It opens *.osd file and parses XML to extract Text elements from EventText groups.
- For every item it searches for source language text, e.g. German 1031. Source language can be defined in configuration.
- The text is send to DeepL via API and the translation for destination language is received. Destination language can be defined in configuration.
- New XML element containing translated text is created., e.g. Czech language 1029.
- At the end the result is saved back to new *.osd file.
Results
...are awesome 😍 😀.
My idea was to use it only as a support for manual translation. I wanted to run it for pre-processsing and then go trough the new file and manualy correct all mistakes and wrong technical terms. I was really surprised to find that most of the automatically translated files don't need any corrections at all.
Are you interested?
If you are interested in the source code and you are already member of Bosch Social Coding platform (Bitbucket) let me know and I will give you access to the repository.
DeepL API
With Free plan access to API you can translated 500 000 characters in a month which is enough for this task. Unfortunatelly you have to fill your credit card information during the registration, even though you won't be charged any fees.
