stanojevic / beer Goto Github PK

Hi again Milos,

./beer --help indicates that --norm does not tokenize input, while the github documentation states the opposite, i.e. that one should use this option in order to tokenize (which is also what --norm name suggests).

Hi Milos,

When running BEER to evaluate French-to-English or German-to-English, I got an error in some situations. This only happens for some files, not all of them an not for other translation directions. I have not been able to find an explanation nor a way to correct this. I tried with --norm, --NoPunct and/or --NoLower options as well, but nothing changed.

I ran:

./beer --lang en --verbose -r ref -s out

ref and out files are attached.

beer_error_minimal_example.tar.gz

Here is the log (with --verbose)

Loading beer.alignment.MeteorAligner . . .
DONE

LOAD LEARNER FROM /people/pecheux/src/beer/models/en_func_exact_para_stem_syn_dep_weka MADE OF beer.learningToRank.Weka@479d31f3

Loading beer.ingredients.DependencyMatching . . .
Reading POS tagger model from edu/stanford/nlp/models/pos-tagger/english-left3words/english-left3words-distsim.tagger ... done [0.9 sec].
Loading depparse model file: edu/stanford/nlp/models/parser/nndep/english_SD.gz ...
PreComputed 100000, Elapsed Time: 1.969 (s)
Initializing dependency parser done [5.9 sec].
DONE
Loading beer.ingredients.Kendall . . .
DONE
Loading beer.ingredients.KendallScaled . . .
DONE
Loading beer.ingredients.PETcount . . .
DONE
Loading beer.ingredients.PETnodes . . .
DONE
Loading beer.ingredients.CharNGrams . . .
DONE
Loading beer.ingredients.FunctionAndContentWordsCoverage . . .
DONE
Loading beer.ingredients.WordMatchingCoverage . . .
DONE
Exception in thread "main" java.util.NoSuchElementException
at edu.stanford.nlp.process.DocumentPreprocessor$PlainTextIterator.next(DocumentPreprocessor.java:373)
at edu.stanford.nlp.process.DocumentPreprocessor$PlainTextIterator.next(DocumentPreprocessor.java:227)
at beer.ingredients.DependencyMatching$.parse(DependencyMatching.scala:196)
at beer.ingredients.DependencyMatching.eval(DependencyMatching.scala:22)
at beer.BEER$$anonfun$6$$anonfun$7.apply(BEER.scala:64)
at beer.BEER$$anonfun$6$$anonfun$7.apply(BEER.scala:64)
at scala.collection.immutable.List.map(List.scala:276)
at beer.BEER$$anonfun$6.apply(BEER.scala:64)
at beer.BEER$$anonfun$6.apply(BEER.scala:63)
at beer.BEER.directFactorsAndTokens(BEER.scala:240)
at beer.BEER.factorsAndTokens(BEER.scala:223)
at beer.Evaluation$$anonfun$evaluate$2$$anonfun$apply$5.apply(Evaluation.scala:120)
at beer.Evaluation$$anonfun$evaluate$2$$anonfun$apply$5.apply(Evaluation.scala:119)
at scala.collection.immutable.List.foreach(List.scala:381)
at beer.Evaluation$$anonfun$evaluate$2.apply(Evaluation.scala:119)
at beer.Evaluation$$anonfun$evaluate$2.apply(Evaluation.scala:107)
at scala.collection.immutable.List.foreach(List.scala:381)
at beer.Evaluation$.evaluate(Evaluation.scala:107)
at beer.Evaluation$.main(Evaluation.scala:23)
at beer.Evaluation.main(Evaluation.scala)

@stanojevic anyway to force BEER to generate one score per line?