summaryrefslogtreecommitdiff
path: root/phrase2_extraction
diff options
context:
space:
mode:
authorPatrick Simianer <p@simianer.de>2016-06-07 17:10:19 +0200
committerPatrick Simianer <p@simianer.de>2016-06-07 17:10:19 +0200
commit30d1bd804d1e9891287289388015bfecc5e25a12 (patch)
treea2f83a5998c5ee80aa9f0247117300159e53b08f /phrase2_extraction
parent913d602044ef8083ba5dcf68b46b075cf345006f (diff)
phrase2_extraction: adjacent nt, MAX_SRC_SZ
Diffstat (limited to 'phrase2_extraction')
-rwxr-xr-xphrase2_extraction/phrase2_extraction.rb5
1 files changed, 4 insertions, 1 deletions
diff --git a/phrase2_extraction/phrase2_extraction.rb b/phrase2_extraction/phrase2_extraction.rb
index 01bdae9..1f268cd 100755
--- a/phrase2_extraction/phrase2_extraction.rb
+++ b/phrase2_extraction/phrase2_extraction.rb
@@ -7,7 +7,7 @@ module PhrasePhraseExtraction
DEBUG = false
MAX_NT = 1 # Chiang: 2
MAX_SEED_NUM_WORDS = 3 # Chiang: 10 words, -> phrases!
-MAX_SRC_SZ = 7 # Chiang: 5 words, -> words!
+MAX_SRC_SZ = 8 # Chiang: 5 words, -> words!
FORBID_SRC_ADJACENT_SRC_NT = true # Chiang:true
class Rule
@@ -649,6 +649,8 @@ def PhrasePhraseExtraction.remove_adjacent_nt rules
prev = false
end
}
+ b
+=begin
c = false
prev = false
r.target.each { |i|
@@ -663,6 +665,7 @@ def PhrasePhraseExtraction.remove_adjacent_nt rules
end
}
b || c
+=end
}
end