From 357eceabad7a0c4590ecc1d8bc1fd5794146d30e Mon Sep 17 00:00:00 2001
From: deroshkin <deroshkinsf@gmail.com>
Date: Mon, 22 Nov 2021 21:25:08 +0100
Subject: [PATCH] Added emdash 0x2014 as a word separator

---
 novel_stats/novel_stats.py | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/novel_stats/novel_stats.py b/novel_stats/novel_stats.py
index 7d660ca..189975d 100755
--- a/novel_stats/novel_stats.py
+++ b/novel_stats/novel_stats.py
@@ -4,6 +4,7 @@
 import argparse
 import collections
 import tempfile
+import re
 
 CHAPTER_MARKER = '## '
 STATUS_MARKER = '[status]: # '
@@ -11,12 +12,13 @@ ACT_MARKER = '[act]: # '
 # Standard markdown comment marker, supported by Pandoc and Calibre's ebook-convert.
 COMMENT_MARKER = '[//]: # '
 TITLE_MARKER = '# '
+WORD_SEPS = [' ','—']
 
 
 def count_words(line):
     count = 0
 
-    for word in line.strip().split(' '):
+    for word in re.split('|'.join(WORD_SEPS), line.strip()):
         if not word.strip() or word == '*' or word.startswith('#'):
             continue