Support for CCG Parser #1

Activity

1. desilinguist commented on May 12, 2015

   

   desilinguist

   on May 12, 2015

   Contributor

   More actions

   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
   #1.
2. desilinguist

   closed this as completedon May 12, 2015
3. silentrob commented on May 13, 2015

   

   silentrob

   on May 13, 2015

   Author

   More actions

   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.
4. desilinguist commented on May 13, 2015

   

   desilinguist

   on May 13, 2015

   Contributor

   More actions

   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
   #1 (comment)
   .
5. desilinguist

   reopened this on May 13, 2015
6. silentrob commented on May 13, 2015

   

   silentrob

   on May 13, 2015

   Author

   More actions

   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. :)
7. silentrob commented on May 13, 2015

   

   silentrob

   on May 13, 2015

   Author

   More actions

   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.