Open silentrob opened 9 years ago
Rob,
node-zpar supports the part of speech tagger, the constituency parser, and the dependency parser. On Tue, May 12, 2015 at 7:11 PM Rob Ellis notifications@github.com wrote:
This is great! I was just about to do the same thing today! Does the CCG (conparser) Parser work too or just the DepParser?
— Reply to this email directly or view it on GitHub https://github.com/EducationalTestingService/node-zpar/issues/1.
Thanks for the update. It looks like the conparser is returning one form.
console.log(zpar.parse_sentence('I am going to the market.\n'));
// (S (NP (PRP I)) (VP (VBP am) (VP (VBG going) (PP (TO to) (NP (DT the) (NN market))))) (. .))
What I was looking for is the other format. http://people.sutd.edu.sg/~yue_zhang/doc/doc/ccg.html It may just need a different input param. Ill take a look at the source.
That'll be great! Thanks, Rob.
On Wed, May 13, 2015 at 11:00 AM Rob Ellis notifications@github.com wrote:
Thanks for the update. It looks like the conparser is returning one form.
console.log(zpar.parse_sentence('I am going to the market.\n')); // (S (NP (PRP I)) (VP (VBP am) (VP (VBG going) (PP (TO to) (NP (DT the) (NN market))))) (. .))
What I was looking for is the other format. http://people.sutd.edu.sg/~yue_zhang/doc/doc/ccg.html It may just need a different input param. Ill take a look at the source.
— Reply to this email directly or view it on GitHub https://github.com/EducationalTestingService/node-zpar/issues/1#issuecomment-101704109 .
So the code appears to be fine. I'm not positive the pre-computed english conparse model contains the CCG data. I was having issues training it yesterday, but I will try it again. :)
So the generic.comparser behaves differently from the english.comparser, and neither are able to generate a CCG model.
Mountain:ccg_files robellis$ ../../../dist/generic.conparser/train train.txt -urules.unary -brules.binary -ctrain.input testdb2 1 Training started. Training iteration is started... Loading scores... empty. Segmentation fault: 11 Mountain:ccg_files robellis$ ../../../dist/english.conparser/train train.txt -urules.unary -brules.binary -ctrain.input testdb2 1 Training started. Training iteration is started... Loading scores... empty. Error: unknown constituent: (S\NP)/(S\NP).
I'm curious to know if anyone has gotten this to work, there seems to be very little online.
This is great! I was just about to do the same thing today! Does the CCG (conparser) Parser work too or just the DepParser?