From 4c71637569c96a5652569d9d59afc9d361ba615c Mon Sep 17 00:00:00 2001
From: Franck Dary <franck.dary@lis-lab.fr>
Date: Wed, 8 Sep 2021 15:52:57 +0200
Subject: [PATCH] Fixed script concatW2V
---
scripts/concatW2V.py | 22 ++++++++++++++++------
1 file changed, 16 insertions(+), 6 deletions(-)
diff --git a/scripts/concatW2V.py b/scripts/concatW2V.py
index 024459e..8100d98 100755
--- a/scripts/concatW2V.py
+++ b/scripts/concatW2V.py
@@ -2,17 +2,27 @@
import sys
-hadFirst = False
+nbLines = 0
+embSize = None
for filename in sys.argv[1:] :
- prefix = filename.split('/')[-1].split('.')[0]
for line in open(filename, "r") :
line = line.strip()
splited = line.split()
if len(splited) == 2 :
- if hadFirst :
- continue
- hadFirst = True
- print(line)
+ if embSize is None :
+ embSize = int(splited[1])
+ elif embSize != int(splited[1]) :
+ print("ERROR : incompatibles embedings sizes %d and %d"%(embSize, int(splited[1])), file=sys.stderr)
+ exit(1)
else :
+ nbLines += 1
+
+print(nbLines, embSize)
+for filename in sys.argv[1:] :
+ prefix = filename.split('/')[-1].split('.')[0]
+ for line in open(filename, "r") :
+ line = line.strip()
+ splited = line.split()
+ if len(splited) > 2 :
print(prefix+"_"+line)
--
GitLab