Rehome the conll-2012 scorer. Not sure if the 2011 scorer can be reco…

…vered easily yet Update some paths in some properties as well to point to the new /u/nlp/data/coref home
stanfordnlp · May 17, 2024 · 53a8f1b · 53a8f1b
1 parent 54aaf2c
commit 53a8f1b
Show file tree

Hide file tree

Showing 14 changed files with 38 additions and 37 deletions.
diff --git a/data/edu/stanford/nlp/dcoref/coref.properties b/data/edu/stanford/nlp/dcoref/coref.properties
@@ -10,8 +10,9 @@ dcoref.postprocessing = true
 dcoref.maxdist = -1
 dcoref.use.big.gender.number = false
 dcoref.replicate.conll = true
-dcoref.conll.scorer = /u/scr/nlp/data/conll-2011/scorer/v4/scorer.pl
+# TODO maybe this can be 2012?
+dcoref.conll.scorer = /u/nlp/data/coref/conll-2012/scorer/v4/scorer.pl
 
-dcoref.logFile = /u/scr/nlp/coref/error_log/temp/result_conlldev.txt
-dcoref.conll2011 = /u/scr/nlp/data/conll-2011/v2/data/dev/data/english/annotations
+dcoref.logFile = /u/nlp/data/coref/error_log/temp/result_conlldev.txt
+dcoref.conll2011 = /u/nlp/data/coref/conll-2011/v2/data/dev/data/english/annotations
 
diff --git a/itest/src/edu/stanford/nlp/dcoref/DcorefSlowITest.java b/itest/src/edu/stanford/nlp/dcoref/DcorefSlowITest.java
@@ -35,14 +35,14 @@ protected void makePropsFile(String path, String workDir, String scorer) throws
   }
 
   public void testDcorefCoNLLResultV4() throws Exception {
-    double finalScore = runDcoref(String.format("%s/conll-2011/scorer/v4/scorer.pl", TestPaths.testHome()));
-    System.out.printf("Final Score (CoNLL 2011, scorer v4): (MUC+B^3+ceafe)/3 = %.2f%n", finalScore);
+    double finalScore = runDcoref(String.format("%s/conll-2012/scorer/v4/scorer.pl", TestPaths.testHome()));
+    System.out.printf("Final Score (CoNLL 2012, scorer v4): (MUC+B^3+ceafe)/3 = %.2f%n", finalScore);
     assertEquals(59.3, finalScore, 0.3); // 2016-07: 59.45
   }
 
   public void testDcorefCoNLLResultV801() throws Exception {
     double finalScore = runDcoref(String.format("%s/conll-2012/scorer/v8.01/scorer.pl", TestPaths.testHome()));
-    System.out.printf("Final Score (CoNLL 2011, scorer v8): (MUC+B^3+ceafe)/3 = %.2f%n", finalScore);
+    System.out.printf("Final Score (CoNLL 2012, scorer v8): (MUC+B^3+ceafe)/3 = %.2f%n", finalScore);
     assertEquals(54.0, finalScore, 0.3); // 2016-07: 54.13
   }
 

diff --git a/src/edu/stanford/nlp/coref/CorefProperties.java b/src/edu/stanford/nlp/coref/CorefProperties.java
@@ -129,7 +129,7 @@ public static boolean useGoldMentions(Properties props) {
 
   public static final String OUTPUT_PATH_PROP = "coref.conllOutputPath";
   public static String conllOutputPath(Properties props) {
-    String returnPath = props.getProperty("coref.conllOutputPath", "/u/scr/nlp/coref/logs/");
+    String returnPath = props.getProperty("coref.conllOutputPath", "/u/nlp/data/coref/logs/");
     if ( ! returnPath.endsWith("/")) {
       returnPath += "/";
     }
@@ -144,7 +144,7 @@ public static void setInput(Properties props, Dataset d) {
   }
 
   private static String getDataPath(Properties props) {
-    String returnPath = props.getProperty("coref.data", "/u/scr/nlp/data/conll-2012/");
+    String returnPath = props.getProperty("coref.data", "/u/nlp/data/coref/conll-2012/");
     if ( ! returnPath.endsWith("/")) {
       returnPath += "/";
     }
@@ -172,7 +172,7 @@ public static String getInputPath(Properties props) {
   }
 
   public static String getScorerPath(Properties props) {
-    return props.getProperty("coref.scorer", "/u/scr/nlp/data/conll-2012/scorer/v8.01/scorer.pl");
+    return props.getProperty("coref.scorer", "/u/nlp/data/coref/conll-2012/scorer/v8.01/scorer.pl");
   }
 
   public static Locale getLanguage(Properties props) {

diff --git a/src/edu/stanford/nlp/coref/hybrid/properties/coref-conll-dep.properties b/src/edu/stanford/nlp/coref/hybrid/properties/coref-conll-dep.properties
@@ -19,20 +19,20 @@ coref.conll = true
 
 ## conll 2012
 ## train
-coref.path.trainData = /u/scr/nlp/data/conll-2012/v4/data/train/data/english/annotations/
+coref.path.trainData = /u/nlp/data/coref/conll-2012/v4/data/train/data/english/annotations/
 #coref.path.traindata = /scr/nlp/data/conll-2012/v4/data/train+dev/
 
 ## eval
 #coref.path.evaldata = /scr/nlp/data/conll-2012/v4/data/development/data/english/annotations/
-coref.path.testData = /u/scr/nlp/data/conll-2012/v9/data/test/data/english/annotations
+coref.path.testData = /u/nlp/data/coref/conll-2012/v9/data/test/data/english/annotations
 
 
 ## models
-coref.path.serialized = /u/scr/nlp/data/coref/models/hybrid-conll-dep
+coref.path.serialized = /u/nlp/data/coref/models/hybrid-conll-dep
 
 ## other data
 coref.big.gender.number = edu/stanford/nlp/models/dcoref/gender.data.gz
-coref.path.word2vec = /u/scr/nlp/data/coref/wordvectors/en/vectors.txt.gz
+coref.path.word2vec = /u/nlp/data/coref/wordvectors/en/vectors.txt.gz
 
 #############################################################################
 # mention detection

diff --git a/src/edu/stanford/nlp/coref/hybrid/properties/coref-conll.properties b/src/edu/stanford/nlp/coref/hybrid/properties/coref-conll.properties
@@ -19,20 +19,20 @@ coref.conll = true
 
 ## conll 2012
 ## train
-coref.path.trainData = /u/scr/nlp/data/conll-2012/v4/data/train/data/english/annotations/
+coref.path.trainData = /u/nlp/data/coref/conll-2012/v4/data/train/data/english/annotations/
 #coref.path.traindata = /scr/nlp/data/conll-2012/v4/data/train+dev/
 
 ## eval
 #coref.path.evaldata = /scr/nlp/data/conll-2012/v4/data/development/data/english/annotations/
-coref.path.testData = /u/scr/nlp/data/conll-2012/v9/data/test/data/english/annotations
+coref.path.testData = /u/nlp/data/coref/conll-2012/v9/data/test/data/english/annotations
 
 
 ## models
-coref.path.serialized = /u/scr/nlp/data/coref/models/hybrid-conll/
+coref.path.serialized = /u/nlp/data/coref/models/hybrid-conll/
 
 ## other data
 coref.big.gender.number = edu/stanford/nlp/models/dcoref/gender.data.gz
-coref.path.word2vec = /u/scr/nlp/data/coref/wordvectors/en/vectors.txt.gz
+coref.path.word2vec = /u/nlp/data/coref/wordvectors/en/vectors.txt.gz
 
 #############################################################################
 # mention detection

diff --git a/src/edu/stanford/nlp/coref/hybrid/properties/coref-default-dep.properties b/src/edu/stanford/nlp/coref/hybrid/properties/coref-default-dep.properties
@@ -7,7 +7,7 @@ coref.sieves = SpeakerMatch, PreciseConstructs, pp-rf, cc-rf, pc-rf, ll-rf, pr-r
 
 coref.useConstituencyParse = false
 coref.conll = false
-coref.path.trainData = /u/scr/nlp/data/conll-2012/v4/data/train/data/english/annotations/
+coref.path.trainData = /u/nlp/data/coref/conll-2012/v4/data/train/data/english/annotations/
 
 #############################################################################
 # data & model path
@@ -18,7 +18,7 @@ coref.path.serialized = edu/stanford/nlp/models/hcoref/
 
 ## other data
 coref.big.gender.number = edu/stanford/nlp/models/dcoref/gender.data.gz
-coref.path.word2vec = /u/scr/nlp/data/coref/wordvectors/en/vectors.txt.gz
+coref.path.word2vec = /u/nlp/data/coref/wordvectors/en/vectors.txt.gz
 coref.path.word2vecSerialized = edu/stanford/nlp/models/word2vec/vectors.ser
 
 #############################################################################

diff --git a/src/edu/stanford/nlp/coref/hybrid/properties/dcoref-conll-dep.properties b/src/edu/stanford/nlp/coref/hybrid/properties/dcoref-conll-dep.properties
@@ -23,12 +23,12 @@ coref.path.trainData = /220/log-hcoref/forthesis/cleanup/small/train
 
 ## eval
 #coref.path.evaldata = /scr/nlp/data/conll-2012/v4/data/development/data/english/annotations/
-coref.path.evalData = /u/scr/nlp/data/conll-2012/v9/data/test/data/english/annotations
+coref.path.evalData = /u/nlp/data/coref/conll-2012/v9/data/test/data/english/annotations
 #coref.path.evaldata = /220/log-hcoref/forthesis/cleanup/small/eval
 
 
 ## scorer
-coref.path.scorer.conll = /u/scr/nlp/data/conll-2012/scorer/v8.01/scorer.pl
+coref.path.scorer.conll = /u/nlp/data/coref/conll-2012/scorer/v8.01/scorer.pl
 
 ## other data
 coref.big.gender.number = edu/stanford/nlp/models/dcoref/gender.data.gz

diff --git a/src/edu/stanford/nlp/coref/hybrid/properties/dcoref-conll.properties b/src/edu/stanford/nlp/coref/hybrid/properties/dcoref-conll.properties
@@ -23,7 +23,7 @@ coref.path.trainData = /220/log-hcoref/forthesis/cleanup/small/train
 
 ## eval
 #coref.path.evaldata = /scr/nlp/data/conll-2012/v4/data/development/data/english/annotations/
-coref.path.testData = /u/scr/nlp/data/conll-2012/v9/data/test/data/english/annotations
+coref.path.testData = /u/nlp/data/coref/conll-2012/v9/data/test/data/english/annotations
 #coref.path.evaldata = /220/log-hcoref/forthesis/cleanup/small/eval
 
 
@@ -32,7 +32,7 @@ coref.path.serialized = /home/heeyoung/log-hcoref/ser/DIR/
 coref.path.DIR=../forthesis/cleanup/ser/
 
 ## scorer
-coref.path.scorer.conll = /u/scr/nlp/data/conll-2012/scorer/v8.01/scorer.pl
+coref.path.scorer.conll = /u/nlp/data/coref/conll-2012/scorer/v8.01/scorer.pl
 
 ## other data
 coref.big.gender.number = edu/stanford/nlp/models/dcoref/gender.data.gz

diff --git a/src/edu/stanford/nlp/coref/hybrid/properties/zh-coref-default.properties b/src/edu/stanford/nlp/coref/hybrid/properties/zh-coref-default.properties
@@ -3,9 +3,9 @@ annotators = segment, ssplit, pos, lemma, ner, parse, mention, coref
 coref.algorithm = hybrid
 
 # data paths
-coref.data = /u/scr/nlp/data/conll-2012/
-coref.conllOutputPath = /u/scr/nlp/coref/logs/
-coref.scorer = /u/scr/nlp/data/conll-2012/scorer/v8.01/scorer.pl
+coref.data = /u/nlp/data/coref/conll-2012/
+coref.conllOutputPath = /u/nlp/data/coref/logs/
+coref.scorer = /u/nlp/data/coref/conll-2012/scorer/v8.01/scorer.pl
 
 # coref options
 coref.sieves = ChineseHeadMatch, ExactStringMatch, PreciseConstructs, StrictHeadMatch1, StrictHeadMatch2, StrictHeadMatch3, StrictHeadMatch4, PronounMatch

diff --git a/src/edu/stanford/nlp/coref/hybrid/properties/zh-dcoref-conll-no-output.properties b/src/edu/stanford/nlp/coref/hybrid/properties/zh-dcoref-conll-no-output.properties
@@ -20,9 +20,9 @@ coref.addMissingAnnotations = true
 coref.specialCaseNewswire = true
 
 # Evaluation
-coref.path.scorer.conll = /u/scr/nlp/data/conll-2012/scorer/v8.01/scorer.pl
+coref.path.scorer.conll = /u/nlp/data/coref/conll-2012/scorer/v8.01/scorer.pl
 #coref.path.output = /scr/nlp/coref/output/
-coref.path.testData = /u/scr/nlp/data/conll-2012/v4/data/development/data/chinese/annotations/
+coref.path.testData = /u/nlp/data/coref/conll-2012/v4/data/development/data/chinese/annotations/
 
 # NER
 ner.model = edu/stanford/nlp/models/ner/chinese.misc.distsim.crf.ser.gz

diff --git a/src/edu/stanford/nlp/coref/hybrid/properties/zh-dcoref-conll.properties b/src/edu/stanford/nlp/coref/hybrid/properties/zh-dcoref-conll.properties
@@ -22,10 +22,10 @@ coref.specialCaseNewswire = true
 coref.input.type = conll
 
 # Evaluation
-coref.path.output = /u/scr/nlp/coref/output/
-coref.data = /u/scr/nlp/data/conll-2012/
-coref.inputPath = /u/scr/nlp/data/conll-2012/v4/data/development/data/chinese/annotations
-coref.scorer = /u/scr/nlp/data/conll-2012/scorer/v8.01/scorer.pl
+coref.path.output = /u/nlp/data/coref/output/
+coref.data = /u/nlp/data/coref/conll-2012/
+coref.inputPath = /u/nlp/data/coref/conll-2012/v4/data/development/data/chinese/annotations
+coref.scorer = /u/nlp/data/coref/conll-2012/scorer/v8.01/scorer.pl
 
 # NER
 ner.model = edu/stanford/nlp/models/ner/chinese.misc.distsim.crf.ser.gz

diff --git a/src/edu/stanford/nlp/coref/statistical/properties/english-conll-training.properties b/src/edu/stanford/nlp/coref/statistical/properties/english-conll-training.properties
@@ -1,3 +1,3 @@
 coref.conll = true
-coref.data = /u/scr/nlp/data/conll-2012/
-coref.statistical.trainingPath = /u/scr/nlp/coref/training/
+coref.data = /u/nlp/data/coref/conll-2012/
+coref.statistical.trainingPath = /u/nlp/data/coref/training/
diff --git a/src/edu/stanford/nlp/coref/statistical/properties/english-default-training.properties b/src/edu/stanford/nlp/coref/statistical/properties/english-default-training.properties
@@ -1,3 +1,3 @@
 coref.conll = false
-coref.data = /u/scr/nlp/data/conll-2012/
-coref.statistical.trainingPath = /u/scr/nlp/coref/training/
+coref.data = /u/nlp/data/coref/conll-2012/
+coref.statistical.trainingPath = /u/nlp/data/coref/training/
diff --git a/src/edu/stanford/nlp/dcoref/Constants.java b/src/edu/stanford/nlp/dcoref/Constants.java
@@ -47,7 +47,7 @@ protected Constants() {} // static class but extended by jcoref
   public static final boolean PRINT_CONLL_OUTPUT = false;
 
   /** Default path for conll scorer script */
-  public static final String conllMentionEvalScript = "/u/scr/nlp/data/conll-2011/scorer/v4/scorer.pl";
+  public static final String conllMentionEvalScript = "/u/nlp/data/coref/conll-2012/scorer/v4/scorer.pl";
 
   /** if true, skip coreference resolution. do mention detection only */
   public static final boolean SKIP_COREF = false;