s434730/RandomSec
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Disable keyers that are super-seeded by others.

Browse Source
This commit is contained in:
Antonin Delpeuch 2018-12-30 21:20:17 +01:00
parent 22cf0517b9
commit 626c17c649
2 changed files with 12 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
main/src/com/google/refine/clustering/binning/KeyerFactory.java
View File

@ -46,12 +46,15 @@ public class KeyerFactory {
static final private List<String> _keyerNames = new LinkedList<>(); static final private List<String> _keyerNames = new LinkedList<>();
static { static {
// Some keyers are disabled as they are super-seeded by others
// See https://github.com/OpenRefine/OpenRefine/pull/1906
put("fingerprint", new FingerprintKeyer()); put("fingerprint", new FingerprintKeyer());
put("ngram-fingerprint", new NGramFingerprintKeyer()); put("ngram-fingerprint", new NGramFingerprintKeyer());
put("metaphone", new MetaphoneKeyer()); // put("metaphone", new MetaphoneKeyer());
put("double-metaphone", new DoubleMetaphoneKeyer()); // put("double-metaphone", new DoubleMetaphoneKeyer());
put("metaphone3", new Metaphone3Keyer()); put("metaphone3", new Metaphone3Keyer());
put("soundex", new SoundexKeyer()); // put("soundex", new SoundexKeyer());
put("cologne-phonetic", new ColognePhoneticKeyer()); put("cologne-phonetic", new ColognePhoneticKeyer());
} }

7
main/src/com/google/refine/clustering/knn/DistanceFactory.java
View File

@ -55,13 +55,18 @@ public class DistanceFactory {
static { static {
put("levenshtein", new VicinoDistance(new LevenshteinDistance())); put("levenshtein", new VicinoDistance(new LevenshteinDistance()));
put("ppm", new VicinoDistance(new PPMDistance()));
// Distances not activated as they are not very useful:
// See https://github.com/OpenRefine/OpenRefine/pull/1906
/*
put("jaccard", new VicinoDistance(new JaccardDistance())); put("jaccard", new VicinoDistance(new JaccardDistance()));
put("jaro", new VicinoDistance(new JaroDistance())); put("jaro", new VicinoDistance(new JaroDistance()));
put("jaro-winkler", new VicinoDistance(new JaroWinklerDistance())); put("jaro-winkler", new VicinoDistance(new JaroWinklerDistance()));
put("jaro-winkler-tfidf", new VicinoDistance(new JaroWinklerTFIDFDistance())); put("jaro-winkler-tfidf", new VicinoDistance(new JaroWinklerTFIDFDistance()));
put("gzip", new VicinoDistance(new GZipDistance())); put("gzip", new VicinoDistance(new GZipDistance()));
put("bzip2", new VicinoDistance(new BZip2Distance())); put("bzip2", new VicinoDistance(new BZip2Distance()));
put("ppm", new VicinoDistance(new PPMDistance())); */
} }
/** /**