I was hoping to finish a bunch of modifications before I leave Edinburgh for Brasil on Thursday, but I am not going to make it. I'll be working for NASA again this summer, so I might not have too much time to do Grok stuff. Nonetheless, I hope to finish these changes and get my current CVS checked in as soon as I can (no promises though!). So, basically, I don't know when I can make the next release since at the moment my current view of the code is not particularly stable.
Anyway, I thought I should mention that Grok development is well and alive despite the lack of recent releases (I'm really bad about doing the releases -- any volunteers to be a releaser?) In particular, anyone interested in part-of-speech tagging for English should use the model which is currently checked into CVS, not the model which is in the last release. It's accuracy is *much* higher. Unfortunately, you cannot just use the model with the old release since the classes in the quipu.grok.preprocess.postag package have changed to improve the model.
I'll be flitting from Edinburgh to Rio de Janeiro to Silicon Valley and back to Rio over the next month, but will be able to check the web and email throughout, so feel free to let me know if you have any questions/comments.
Jason
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi all,
I was hoping to finish a bunch of modifications before I leave Edinburgh for Brasil on Thursday, but I am not going to make it. I'll be working for NASA again this summer, so I might not have too much time to do Grok stuff. Nonetheless, I hope to finish these changes and get my current CVS checked in as soon as I can (no promises though!). So, basically, I don't know when I can make the next release since at the moment my current view of the code is not particularly stable.
Anyway, I thought I should mention that Grok development is well and alive despite the lack of recent releases (I'm really bad about doing the releases -- any volunteers to be a releaser?) In particular, anyone interested in part-of-speech tagging for English should use the model which is currently checked into CVS, not the model which is in the last release. It's accuracy is *much* higher. Unfortunately, you cannot just use the model with the old release since the classes in the quipu.grok.preprocess.postag package have changed to improve the model.
I'll be flitting from Edinburgh to Rio de Janeiro to Silicon Valley and back to Rio over the next month, but will be able to check the web and email throughout, so feel free to let me know if you have any questions/comments.
Jason