Greetings,
I am trying to figure out how to prepare WS indexes for an english corpus.
I understand that I have to use the genws program which is installed on our machine.
I am calling genws in the following way:
genws ukwac lemma "/storage/corpora/wsidx/ukwac.lemma-ws" /storage/corpora/wsdef/ws-testuk
Questions:
a) Are the above assumptions sane? That is, when sb wants to produce WS indexes all that is needed is to use genws, providing a ws grammar file. (I am using the file available here http://trac.sketchengine.co.uk/attachment/wiki/WSDefFiles/ws-tt.txt?format=raw)
b) I get a glibc error from this command:
>>> 1:"NN.?.?" [word=","]{0,1} [word="and"|word="or"|word=","] [tag="DT.?"|tag="PRP\$"]{0,1} "CD"{0,2} [tag="JJ.?"|tag="RB.?"|word=","]{0,3} "NN.?.?"{0,2} 2:"NN.?.?" [tag!="NN.?.?"] within <s>
no results
*** glibc detected *** free(): invalid next size (fast): 0x0926c960 ***
Aborted
Is this a known problem? Is a rebuild of the genws utility necessary? Can you provide some guidance on how to do that?
Best,
Stefanos
