gangeli / ParsingTime

Parsing Time: Learning to Interpret Time Expressions
http://stanford.edu/~angeli/papers/2012-naacl-temporal.pdf
GNU Lesser General Public License v3.0
30 stars 3 forks source link

Fixes some hard coded paths, generation of tempeval2-english-retok-numbers, and Maven/Eclipse support #1

Closed bethard closed 11 years ago

bethard commented 11 years ago

(1) Hard coded paths like "/lib/data/bidirectional-distsim-wsj-0-18.tagger" or "/workspace/..." were preventing me from running the code on my machine. Note that in the fix for the POS tagger, I followed the approach of edu.stanford.nlp.pipeline.StanfordCoreNLP, loding the models from the classpath by default but allowing this to be overridden with a system property pos.model. I used the model specified by DefaultPaths.DEFAULT_POS_MODEL, but that's "edu/stanford/nlp/models/pos-tagger/english-left3words/english-left3words-distsim.tagger", which may or may not be the same as what you used before. If it's not, you can replace it with any of the models distributed with Stanford Core NLP 1.3.5:

edu/stanford/nlp/models/pos-tagger/english-left3words/english-left3words-distsim.tagger edu/stanford/nlp/models/pos-tagger/wsj-bidirectional/wsj-0-18-bidirectional-distsim.tagger edu/stanford/nlp/models/pos-tagger/english-bidirectional/english-bidirectional-distsim.tagger edu/stanford/nlp/models/pos-tagger/wsj-left3words/wsj-0-18-left3words-distsim.tagger

(2) The TempEval2 script was no longer producing the "aux/coremap/tempeval2-english-retok-numbers" file needed for training. I added what I believe are the appropriate calls to retokFromInit and normalizeFromRetok, but you should definitely review these to make sure I'm calling the right methods in the right order.

(3) I added the necessary pom.xml and Eclipse project files to allow this to be build with Maven and/or Eclipse.