And... there are some helper methods:
private HashMap<Integer, Integer> mapTokensToIndices(JCas jc) {
HashMap<Integer,Integer> map = new HashMap<Integer,Integer>();
int i = 1;
FSIterator iter = jc.getAnnotationIndex(BaseToken.type).iterator();
while(iter.hasNext()){
BaseToken tok = (BaseToken) iter.next();
map.put(tok.getBegin(), i);
map.put(tok.getEnd(), i);
i++;
}
return map;
}
private String getSpanString(Markable mark,
HashMap<Integer, Integer> tok2ind) {
Integer tok1 = tok2ind.get(mark.getBegin());
Integer tok2 = tok2ind.get(mark.getEnd());
return "1:" + tok1 + " 1:" + tok2;
}
Hopefully that is it!
________________________________
From: Miller, Timothy [[email protected]]
Sent: Friday, March 14, 2014 2:19 PM
To: [email protected]
Subject: RE: Parser for output of cTAKES coreference module to i2b2 scoring
code format?
Kev -- I think I found it. It looks like it was never checked in, as it was
part of a separate eval module that used gpl'd code and couldn't be released
with ctakes. It was also commented out for some reason. I'll just paste the
relevant section here, and hopefully you can use it or at least save you some
time. If you write a UIMA consumer or other tool that you think would help
others and you are willing to share we would gladly incorporate it into ctakes.
// if(i2b2){
// // write system chain file:
// try {
// PrintWriter writer = new
PrintWriter(i2b2Path + "/sysChain/" + docName + ".chain");
// HashMap<Integer,Integer> tok2ind =
mapTokensToIndices(jc);
// FSIterator iter =
jc.getJFSIndexRepository().getAllIndexedFS(CoreferenceChain.type);
// while(iter.hasNext()){
// CoreferenceChain chain =
(CoreferenceChain) iter.next();
// FSList members =
chain.getMembers();
// while(members instanceof
NonEmptyFSList){
// NonEmptyFSList node =
(NonEmptyFSList) members;
// Markable mark =
(Markable) node.getHead();
// writer.print("c=\"");
//
writer.print(mark.getCoveredText());
// writer.print("\" ");
//
writer.print(getSpanString(mark, tok2ind));
// writer.print("||");
// members =
node.getTail();
// }
// // write the type information
// writer.println("t=\"coref
problem\"");
// }
// writer.close();
// } catch (FileNotFoundException e) {
// // TODO Auto-generated catch block
// e.printStackTrace();
// }
//
// }
Tim
________________________________
From: Miller, Timothy [[email protected]]
Sent: Friday, March 14, 2014 1:52 PM
To: [email protected]
Subject: RE: Parser for output of cTAKES coreference module to i2b2 scoring
code format?
Kevin, I think I did write something like that at some point. I just spent the
last 10 minutes looking for it and can't find it. I will poke around a bit more
and let you know if I find anything.
Tim
________________________________
From: Kevin B. Cohen [[email protected]]
Sent: Friday, March 14, 2014 12:41 PM
To: [email protected]; Natural language processing for biology
Subject: Parser for output of cTAKES coreference module to i2b2 scoring code
format?
Hi,
I'm sitting here writing a little parser to convert the output of the cTAKES
coreference resolution module into the format of the scoring code from the i2b2
coreference resolution task, and it occurred to me that multiple people have
probably done that already. Would anyone be willing to share code for that?
Kev
--
Kevin Bretonnel Cohen, PhD
Biomedical Text Mining Group Lead, Computational Bioscience Program,
U. Colorado School of Medicine
303-916-2417303-916-2417
http://compbio.ucdenver.edu/Hunter_lab/Cohen
Call
Send SMS
Add to Skype
You'll need Skype CreditFree via Skype
