search

dkpro / dkpro-core

Collection of software components for natural language processing (NLP) based on the Apache UIMA framework.
https://dkpro.github.io/dkpro-core
Other
196 stars 67 forks source link

BratReader does not support N1 Reference tags #1391

Open alaindesilets opened 5 years ago

alaindesilets commented 5 years ago

If you run ReadBrat.java in the attached Maven project, you get this errror:

Exception in thread "main" java.lang.IllegalStateException: Unknown annotation format: [N1  Reference T1 Wikipedia:Q95  Google]

Yet, the annotation in question is perfectly valid, according to the brat format specification:

https://brat.nlplab.org/standoff.html

I should point out that we are encountering this problem because we are using BratReader to read some files that were previously annotated by MetamapLite, and that tool makes use of the Reference line in the .ann files.

reckart commented 5 years ago

So far, normalizations (N) are not supported by the BratReader/BratWriter in DKPro Core.

alaindesilets commented 5 years ago

Any plans to add them?

Btw we noticed that you guys released the fix we sent you last month thx.

I started the process for getting the paperwork done to allow us to contribute to dkpdo. So if you give me an idea of how to go about adding support for reference brat lines into the Brat reader and writer we can implement it.

Le dim. 7 juil. 2019 à 06:03, Richard Eckart de Castilho < notifications@github.com> a écrit :

So far, normalizations (N) are not supported by the BratReader/BratWriter in DKPro Core.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/dkpro/dkpro-core/issues/1391?email_source=notifications&email_token=AAIMA4BBAIFAHXSKRB3PQDTP6G5PFA5CNFSM4H6UMP3KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZLIM6I#issuecomment-508987001, or mute the thread https://github.com/notifications/unsubscribe-auth/AAIMA4FEGIRQVAGRBXYBBJ3P6G5PFANCNFSM4H6UMP3A .

alaindesilets commented 5 years ago

On Sun, Jul 7, 2019 at 6:42 AM Alain Désilets alaindesilets0@gmail.com wrote:

So if you give me an idea of how to go about adding support for reference brat lines into the Brat reader and writer we can implement it.

Also:

reckart commented 5 years ago

The contribution guidelines include links to coding convention templates.

Yes, there are unit tests for the brat module.

If I understand it correctly, then the normalization annotations in brat are essentially like attributes. The BratReader could probably just treat them as String attributes. The BratWriter would need some way of configuring which attributes should be rendered as normalization attributes instead of regular attributes.

alaindesilets commented 5 years ago

Thx.

Le lun. 8 juil. 2019 à 05:46, Richard Eckart de Castilho < notifications@github.com> a écrit :

The contribution guidelines https://dkpro.github.io/contributing/ include links to coding convention templates.

Yes, there are unit tests for the brat module.

If I understand it correctly, then the normalization annotations in brat are essentially like attributes. The BratReader could probably just treat them as String attributes. The BratWriter would need some way of configuring which attributes should be rendered as normalization attributes instead of regular attributes.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/dkpro/dkpro-core/issues/1391?email_source=notifications&email_token=AAIMA4FDLTBX6QXDXN4UVWTP6MEIDA5CNFSM4H6UMP3KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZMSCUI#issuecomment-509157713, or mute the thread https://github.com/notifications/unsubscribe-auth/AAIMA4BSLBBKDPCWCYVJBZ3P6MEIDANCNFSM4H6UMP3A .