| Subject: | faulty escaping of output when apostrophes are found (includes patch) |
Parsing apostrophes results in broken output. For example:
parse: I can't parse this.
result:
[
{ relation => 'nominal subject', from => 'parse-4', to => 'I-1' },
{ relation => 'auxiliary', from => 'parse-4', to => 'ca-2' },
{ relation => 'negation modifier', from => 'parse-4', to => 'n't-3' },
{ relation => 'direct object', from => 'parse-4', to => 'this-5' },
]
The string "to => 'n't-3'" isn't valid Perl. Trivial fix follows:
--- StanfordParser.pm 2012-02-17 21:45:16.000000000 -0800
+++ StanfordParser.pm.fixed 2012-02-18 20:30:17.000000000 -0800
@@ -63,8 +63,8 @@
for (TypedDependency td : collxn) {
buf.append("{ relation => '").
append(td.reln().getLongName()).append("', from => '").
- append(td.gov()).append("', to => '").
- append(td.dep()).append("' },\n");
+
append(td.gov().toString().replace("'","\\'")).append("', to => '").
+
append(td.dep().toString().replace("'","\\'")).append("' },\n");
}
buf.append("]\n");
return buf.toString();