From 3ec2551f99930858f0b8b2aa11c8971c1dce4f7c Mon Sep 17 00:00:00 2001
From: Franck Dary <franck.dary@lis-lab.fr>
Date: Tue, 15 Oct 2019 13:56:54 +0200
Subject: [PATCH] getActionSets now ignore second parts of labels
---
UD_any/data/getActionSets.py | 9 +++++++--
1 file changed, 7 insertions(+), 2 deletions(-)
diff --git a/UD_any/data/getActionSets.py b/UD_any/data/getActionSets.py
index 57666e8..5e7be2a 100755
--- a/UD_any/data/getActionSets.py
+++ b/UD_any/data/getActionSets.py
@@ -57,12 +57,17 @@ if __name__ == "__main__" :
elif nameCol == "LABEL" :
output = open("parser.as", 'w', encoding='utf-8')
print("REDUCE", file=output)
+ labels = set()
for line in open(colFile, "r", encoding='utf-8') :
striped = line.strip()
if len(striped) == 0 or striped == "root" or striped == "_" :
continue
- print("LEFT " + striped, file=output)
- print("RIGHT " + striped, file=output)
+ label = striped.split(':')[0]
+ if label not in labels :
+ labels.add(striped)
+ for label in labels :
+ print("LEFT " + label, file=output)
+ print("RIGHT " + label, file=output)
print("EOS", file=output)
print("Default : SHIFT", file=output)
output.close()
--
GitLab