From 901b4ab2ba0af1ee3e8e9d55bf40fc92be986171 Mon Sep 17 00:00:00 2001
From: Franck Dary <franck.dary@lis-lab.fr>
Date: Thu, 27 Feb 2020 17:40:29 +0100
Subject: [PATCH] Made the transitionsSet generation script to output expected
state name
---
UD_any/data/Makefile | 1 +
UD_any/data/getTransitionSets.py | 11 ++++++-----
2 files changed, 7 insertions(+), 5 deletions(-)
diff --git a/UD_any/data/Makefile b/UD_any/data/Makefile
index 9fb308c..0216349 100644
--- a/UD_any/data/Makefile
+++ b/UD_any/data/Makefile
@@ -36,6 +36,7 @@ columns: all_no_test.conllu $(MCD)
cat all_no_test.conllu | sed '/^#/ d' | cut -f$$number | sort --unique > col_$$number.txt ; \
done
./getTransitionSets.py $(MCD) col_*\.txt
+ cat tagger.ts parser.ts > taggerparser.ts
texts:
./getRawText.py $(CONLL2TXT) $(TRAIN_FILES) $(DEV_FILEs) $(TEST_FILES)
diff --git a/UD_any/data/getTransitionSets.py b/UD_any/data/getTransitionSets.py
index 1ee4d74..7a98d7c 100755
--- a/UD_any/data/getTransitionSets.py
+++ b/UD_any/data/getTransitionSets.py
@@ -31,7 +31,7 @@ if __name__ == "__main__" :
striped = line.strip()
if len(striped) == 0 :
continue
- print("WRITE b.0 UPOS " + striped, file=output)
+ print("<tagger> WRITE b.0 UPOS " + striped, file=output)
output.close()
elif nameCol == "XPOS" :
@@ -90,7 +90,7 @@ if __name__ == "__main__" :
print("SHIFT", file=output)
output.close()
output = open("parser.ts", 'w', encoding='utf-8')
- print("REDUCE", file=output)
+ print("<parser> REDUCE", file=output)
labels = set()
labelsList = []
for line in open(colFile, "r", encoding='utf-8') :
@@ -103,8 +103,9 @@ if __name__ == "__main__" :
labelsList.append(striped)
labelsList.sort()
for label in labelsList :
- print("LEFT " + label, file=output)
- print("RIGHT " + label, file=output)
- print("SHIFT", file=output)
+ print("<parser> LEFT " + label, file=output)
+ print("<parser> RIGHT " + label, file=output)
+ print("<parser> EOS", file=output)
+ print("<parser> SHIFT", file=output)
output.close()
--
GitLab