search

jbjorne / TEES

Turku Event Extraction System
147 stars 44 forks source link

Corpora missing/invalid link #13

Closed jtsui closed 8 years ago

jtsui commented 9 years ago

Just ran python configure.py and looks like some of the corpora links are no longer valid:

http://sourceforge.net/projects/tees/files/data/BioNLP11-corpora-XML-140216.zip http://sourceforge.net/projects/tees/files/data/BioNLP13-corpora-XML-140216.zip

jbjorne commented 9 years ago

Please let me know which TEES version are you using? The latest repository version should point to the correct files (without the XML part), but the odd thing here is that all earlier data files for older releases should be online too, so I wonder what's going on here...

Another problem that I noticed with those files is that the configure.py couldn't install them. This is probably due to them being repacked with Ubuntu's file roller, which produces zip-files that the Python thinks are somehow broken. If you download and unpack manually, the files work, but of course this should be fixed. I'll try to get the files rebuilt sometime next week (our cluster is down for maintenance currently).

Best Regards, Jari

jtsui commented 9 years ago

Thanks for the help.

I just cloned the master branch of the repository a few hours ago, I assume this is the latest version.

jbjorne commented 9 years ago

Indeed it is, and in the TEES/Utils/DefaultSettings.py file the SourceForge URL:s appear to be incorrect. I'll fix this when I rebuild the models.