From 3b2f8d607ecfac9641d63be754fd5241b99d6433 Mon Sep 17 00:00:00 2001
From: Andre Martins <andre.t.martins@gmail.com>
Date: Mon, 13 Oct 2014 20:19:35 +0100
Subject: [PATCH] ENH Added script to create a single gazetteer file.

---
 scripts_ner/create_gazetteer_file.py | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)
 create mode 100644 scripts_ner/create_gazetteer_file.py

diff --git a/scripts_ner/create_gazetteer_file.py b/scripts_ner/create_gazetteer_file.py
new file mode 100644
index 0000000..1608d25
--- /dev/null
+++ b/scripts_ner/create_gazetteer_file.py
@@ -0,0 +1,18 @@
+import os
+import sys
+
+if __name__ == "__main__":
+    path = sys.argv[1] # Path to the folder where the gazetteers are.
+    destination_filepath = sys.argv[2] # Path to the destination file.
+    f_out = open(destination_filepath, 'w')
+    gazetteers_list = [name for name in os.listdir(path) if
+                       not os.path.isdir(path + os.sep + name)]
+    gazetteers_list.sort()
+    for name in gazetteers_list:
+        print 'Adding ', name, '...'
+        filepath = path + os.sep + name
+        f = open(filepath)
+        for line in f:
+            line = line.rstrip('\r\n')
+            f_out.write('%s\t%s\n' % (name, line))
+    f_out.close()