Separate metadata for different texts in a parallel corpus document

[Amir Zeldes]

Currently (V2.2.1), metadata is annotatable per document. Since parallel corpus documents are each one document containing many texts in the relANNIS data model, there is no way to give them separate meta-annotations. However this would make sense for e.g. language (text1 is English, text2 is German), name of the translator, etc. This is already possible in PAULA and SaltNPepper but relANNIS and the ANNIS interface need to be adapted to fit this possibility. Example query:

cat="NP" & meta::is_translated="true"

Outputs all cat="NP" annotations in texts of a parallel document that are translated (but not the original text, which is meta-annotated as meta::is_translated="false"