search

MorphDiv / TeDDi_sample

Text Data Diversity Sample (TeDDi Sample)
Other
5 stars 3 forks source link

concatenated segments in the glosses #232

Closed tsamardzic closed 4 years ago

tsamardzic commented 4 years ago

Fixed Piraha glosses as Chris requested, e.g.

2806c2806 \< \<glossing> * dig_up_VB -DUR -frustrated_completion -old_info -hearsay paca_NN --- > \<glossing> * dig_up_VB-DUR-frustrated_completion-old_info-hearsay paca_NN

It was a matter of adding one simple line in the script piraha_tree_to_text.py.

bambooforest commented 4 years ago

@christianbentz -- i'd be careful merging in new data without running the database generation in test mode first. maybe you did? otherwise, for new or updated corpus data, i suggest that @ximenina always runs the load-database.py script. otherwise, you might get into a case in the future where someone merged badly formatted data, but it's not caught when it's submitted to PR.

also, delete these branches with the Delete branch button after accepting the PR. it'll keep the repo tidy.

bambooforest commented 4 years ago

my bad, i copied piraha under the english folder in the test directory (which means it should have failed). everything works fine at the moment. disregard previous message -- but do test in the future.

tsamardzic commented 4 years ago

@bambooforest totally agree, we thought only @ximenina had the right to merge things. What do you think about restricting the rights?

christianbentz commented 4 years ago

@bambooforest hey, thanks for being attentive. Yes, I wanted to work on the new files by Tanja, so I accepted the pull request without running the test code. Won't happen again:)