Skip to content

Commit 617aef8

Browse files
committed
Add a bunch of words not previously included in the say regex. Generally speaking they are not in PTB3 for testing purposes, though
1 parent 1ac93f6 commit 617aef8

File tree

1 file changed

+6
-1
lines changed

1 file changed

+6
-1
lines changed

src/edu/stanford/nlp/trees/EnglishPatterns.java

+6-1
Original file line numberDiff line numberDiff line change
@@ -86,9 +86,14 @@ public class EnglishPatterns {
8686

8787
/** A list of verbs which are verbs of speaking that easily take an S (as a complement or topicalized)
8888
* which is a direct speech ccomp. For example: "He concedes: ``This is a difficult market.''"
89+
* <br>
90+
* TODO: maybe sign, as in ASL? sing, kibitz ... wish?
91+
* quite a few others from the NP_V_S_INF_VERBS_REGEX:
92+
* pledge, protest, bluster, bray, cackle, call, chuckle, chortle, croak, coo, crow, drawl, drone, gasp ...
93+
* signal, bellow, bleat, blubber, boom, chirp, yammer, yap
8994
*/
9095
public static final String sayVerbRegex =
91-
"/^(?i:say|says|said|saying|(?:add|boast|claim|counsel|explain|inform|interject|proclaim|recall|remark|report|respond|shout|whisper|yell)(?:s|ed|ing)?|(?:advis|announc|acknowledg|conced|conclud|decid|declar|inton|not|observ|stat)(?:e|es|ed|ing)|(?:confess)(?:es|ed|ing)?|(?:agree)(?:s|d|ing)?|reply|replied|replies|replying|admit|admits|admitted|admitting|hold|holds|holding|held|write|writes|writing|wrote|tell|tells|telling|told|quipped|quip|quips|quipping|think|thinks|thinking|thought)$/";
96+
"/^(?i:say|says|said|saying|(?:add|boast|chant|claim|complain|counsel|explain|inform|interject|pray|proclaim|purr|recall|remark|report|respond|scream|shout|shriek|sigh|sulk|whisper|whoop|yell|yelp)(?:s|ed|ing)?|(?:advis|announc|acknowledg|conced|conclud|decid|declar|grip|grous|inton|not|observ|propos|stat|whin|whing)(?:e|es|ed|ing)|(?:bitch|confess|screech)(?:es|ed|ing)?|(?:agree)(?:s|d|ing)?|(?:cr|repl)(?:y|ied|ies|ying)|admit|admits|admitted|admitting|hold|holds|holding|held|write|writes|writing|wrote|tell|tells|telling|told|quipped|quip|quips|quipping|think|thinks|thinking|thought)$/";
9297

9398

9499
// TODO: is there some better pattern to look for? We do not have tag information at this point

0 commit comments

Comments
 (0)