Document CorA XML format - Githubissues

comphist / cora

A web-based, token-level annotation tool for non-standard language data

http://www.linguistics.rub.de/comphist/resources/cora/

MIT License

10 stars 6 forks source link

Document CorA XML format #14

Closed mbollmann closed 7 years ago

mbollmann commented 9 years ago

Originally reported by: Marcel Bollmann (Bitbucket: mbollmann, GitHub: mbollmann)

DTD for CorA XML is probably outdated
Other schema languages might be more appropriate to describe CorA XML?
Technical document should briefly explain the format (and especially the semantics behind the tags, which a schema won't cover)

Bitbucket: https://bitbucket.org/mbollmann/cora/issue/11

mbollmann commented 7 years ago

Original comment by Marcel Bollmann (Bitbucket: mbollmann, GitHub: mbollmann):

We now have:

a basic description of CorA XML in the user docs
a RelaxNG schema
a Python script for validation of CorA XML files with said schema, plus validation of internal logic (such as ID ranges)

This can certainly still be improved, but should be enough to close this issue for now.

mbollmann commented 7 years ago

Original comment by Marcel Bollmann (Bitbucket: mbollmann, GitHub: mbollmann):

Add script for XML validation (re #11)

mbollmann commented 8 years ago

Original comment by Marcel Bollmann (Bitbucket: mbollmann, GitHub: mbollmann):

Schematron file to check various content properties (ID reference validity etc.) is still in the works.

mbollmann commented 8 years ago

Original comment by Marcel Bollmann (Bitbucket: mbollmann, GitHub: mbollmann):

Replace outdated DTD by RelaxNG schema (re issue #11)

mbollmann commented 8 years ago

Original comment by Marcel Bollmann (Bitbucket: mbollmann, GitHub: mbollmann):

Aiko Freyth currently working on a RELAX NG schema

mbollmann commented 8 years ago

Original comment by Marcel Bollmann (Bitbucket: mbollmann, GitHub: mbollmann):

Ways to validate RELAX NG + Schematron:

There doesn't seem to be a simple stand-alone CLI tool, though it would seem to be trivial to make with lxml.

mbollmann commented 8 years ago

Original comment by Marcel Bollmann (Bitbucket: mbollmann, GitHub: mbollmann):

After some research, I would probably try to go with RELAX NG for the main schema language.

I'm not sure if it's capable to express our ID reference ranges ("t3_d1..t7_d2"), but it could probably be combined with Schematron for this purpose. At least Wikipedia states that combining these two systems is a typical solution to get the strengths of both systems.

mbollmann commented 8 years ago

Original comment by Marcel Bollmann (Bitbucket: mbollmann, GitHub: mbollmann):

Commit 7917f2c adds an informal description of the format to the user documentation.