Fix regex logic

florisboard · Feb 23, 2021 · 247aa37 · 247aa37
1 parent db4a468
commit 247aa37
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 6 deletions.
diff --git a/clb.py b/clb.py
@@ -4,13 +4,13 @@
 import gzip
 import msgpack
 import io
-import re
+import regex
 import time
 
-STR_FAIL_REGEX = "[.,]+"
+STR_VALIDATE_REGEX = r"^((\p{L}\p{M}*)|\'|\-)+$"
 
 def __validate_str(str):
-    return re.search(STR_FAIL_REGEX, str) == None and len(str) != 0
+    return regex.search(STR_VALIDATE_REGEX, str) != None and len(str) != 0
 
 def __freq_for_index(index, len_list):
     """

diff --git a/flict.py b/flict.py
@@ -185,9 +185,9 @@ def insertNgram(self, ngram: Ngram):
 
 
 def clb_to_flict(src_path: str, dst_path: str):
-    word_regex = regex.compile(r"^\sword=(\p{L}\p{M}*)+,f=[0-9]+$")
-    bigram_regex = regex.compile(r"^\s\sbigram=(\p{L}\p{M}*)+,f=[0-9]+$")
-    trigram_rgeex = regex.compile(r"^\s\s\strigram=(\p{L}\p{M}*)+,f=[0-9]+$")
+    word_regex = regex.compile(r"^\sword=((\p{L}\p{M}*)|\'|\-)+,f=[0-9]+$")
+    bigram_regex = regex.compile(r"^\s\sbigram=((\p{L}\p{M}*)|\'|\-)+,f=[0-9]+$")
+    trigram_rgeex = regex.compile(r"^\s\s\strigram=((\p{L}\p{M}*)|\'|\-)+,f=[0-9]+$")
     try:
         ptree = FlictRootNode()
         with io.open(src_path, "r", encoding="utf-8") as f_src: