Re:Go-Freeling: Natural Language Processing in Go
by andrewstuart2 on Saturday June 06, 2015 @11:21PM
(#9673277)
I'd suggest making sure it's go-gettable and will build following the instructions in the README when placed in the canonical go path ($GOPATH/github.com/advancedlogic/go-freeling). After fixing the paths, I'm still getting a cgo linking error with the mitie.h file.
Also, probably avoid using . imports, as it's considered a Go antipattern.
Seems really promising, but if you want it to become a project that the Go community can get behind and maintain, you'll probably want to make sure it's familiar and easy to get running. :-)
Re:Go-Freeling: Natural Language Processing in Go
by mrkrwtsn on Sunday June 07, 2015 @02:21AM
(#9673616)
I'm unable to get it to run either. Turning text into tokens is something I've wanted to be able to do in golang before, so it would be super nice if it was easy to install with `go get`. I still need to work through some of the other libraries here: http://biosphere.cc/software-engineering/go-machine-learning...
Re:Go-Freeling: Natural Language Processing in Go
by mschoch on Sunday June 07, 2015 @12:18PM
(#9675250)
If you just want to segment larger blocks of text into tokens you can try the segment library (it implements the word boundary portion of unicode annex 29):
If you need more manipulation of tokens after segmentation/tokenization, you could look at the analysis sub-package of bleve. Its intended to be able to be used indepenently of the rest of the library.
Re:Go-Freeling: Natural Language Processing in Go
by Dewie3 on Sunday June 07, 2015 @11:55AM
(#9675122)
Was that last panel supposed to be about you?
Anyway nothing I've written here is really angry, or passionate, or inflammatory (just snarky). But since I don't get your usage of the word proof in this context, what do you mean specifically?
Re:Go-Freeling: Natural Language Processing in Go
by frik on Sunday June 07, 2015 @03:04AM
(#9673701)
Very nice, something I have been looking for.
Please create a wiki with some more information and examples. Is there a training part involved? Does it support languages other than English, if so which data format or how to train the language model? Is there a WordNet integration? Or can I use the FreeLing website as reference http://nlp.lsi.upc.edu/freeling/index.php?option=com_content... ?
Re:Go-Freeling: Natural Language Processing in Go
by nkozyra on Saturday June 06, 2015 @10:47PM
(#9673203)
Excited to try this, although I don't immediately see what model format is used. I've been a bit surprised to see ML and NLP lag a bit on Go compared to rather swift adoption with other languages.
Also, probably avoid using . imports, as it's considered a Go antipattern.
Seems really promising, but if you want it to become a project that the Go community can get behind and maintain, you'll probably want to make sure it's familiar and easy to get running. :-)
I agree though, seems promising.
https://github.com/blevesearch/segment
If you need more manipulation of tokens after segmentation/tokenization, you could look at the analysis sub-package of bleve. Its intended to be able to be used indepenently of the rest of the library.
https://github.com/blevesearch/bleve
Anyway nothing I've written here is really angry, or passionate, or inflammatory (just snarky). But since I don't get your usage of the word proof in this context, what do you mean specifically?
Please create a wiki with some more information and examples. Is there a training part involved? Does it support languages other than English, if so which data format or how to train the language model? Is there a WordNet integration? Or can I use the FreeLing website as reference http://nlp.lsi.upc.edu/freeling/index.php?option=com_content... ?
https://github.com/Hendler/goling