This is a subtask of issue #2836.
The class GrobidModelConverter converts the metadata extracted by Grobid into the BibSonomy data model.
Currently, it's just using Grobid's
toBibTeX() method which is quite imperfect:
1. It does not properly extract the year, it is just using
publication_date. Compare this with the effort done in the
toTEI method to extract data data.
2. It's entrytype selection approach is wrong, as items with a "booktitle" can be identified as "article" (though that requires a "journal").
3. See the
GrobidModelConverter.convert() for further suggestions for improvement.
Improve the model conversion and create JUnit tests with various different inputs. Basically, the different entry types should be covered and the required fields for each entry type should be considered.
There are two ways to implement this:
1. By implementing our own conversion method.
2. By improving Grobid's