Problem of extractDependencyCounts in Jane 2.0

Skip to first unread message


Feb 3, 2012, 1:42:28 AM2/3/12
to jane Users
In Janes User Manual , the command to extract dependency counts is

"bin/x86_64/extractDependencyCounts.x86_64-standard --dependencyTree
e.parse.gz --leftOut e.counts.left.3gram.gz --rightOut e.counts.right.
3gram.gz --headOut e.counts.head.gz"

But in When I tried it on Jane 2.0 , it just show the following error
[error 3/2/2012 15:28:09 extractdependencybinary] no parse file given

Then in the Useage:
extractDependencyCounts.x86_64-standard [OPTIONS]

format format of the dependency structure, out of
(simple, *stanford, berkeley)
headOut head output file
leftOut left output file
rightOut right output file
headMarker head marker
ngramMax n in ngram (default: 3)
startSentence start sentence
endSentence end sentence
sentenceSplit output dependencies sentence-wise and split
with given string
asPerplexityText output dependencies to as text to evaluate
it's Perplexity

It seems the "dependencyTree" option has been removed in the version

so .. how to specify a input parsed tree file in the version 2.0 ?

Jan-Thorsten Peter

Feb 4, 2012, 4:12:17 AM2/4/12
to jane Users
The option was renamed from "dependencyTree" to "parseFile". This
should work:
"bin/x86_64/extractDependencyCounts.x86_64-standard --parseFile
e.parse.gz --leftOut e.counts.left.3gram.gz --rightOut e.counts.right.
3gram.gz --headOut e.counts.head.gz"

Sorry, we will update it in the manual and usage.


Feb 4, 2012, 7:28:03 PM2/4/12
to jane Users
thanks for the answer , i solved the problem by modifying the code
yesterday ♪~(´ε` )

and found it should be *parseFile* XDDD
Reply all
Reply to author
0 new messages