quebras semanticas

Author: ramalho

ferramentas/sembreak.py

@@ -0,0 +1,22 @@
+#!/usr/bin/env python3
+
+import fileinput
+import re
+
+
+RE_PERIOD = re.compile(r'(\w\w)\. +([A-Z])')
+
+
+def sembreak(text):
+    return RE_PERIOD.sub(r'\1.\n\2', text)
+
+
+def main():
+    for line in fileinput.input(encoding="utf-8"):
+        if line[0] != '.':
+            line = sembreak(line)
+        print(line, end='')
+
+
+if __name__ == '__main__':
+    main()

ferramentas/test_sembreak.py

@@ -0,0 +1,13 @@
+import pytest
+
+from sembreak import sembreak
+
+@pytest.mark.parametrize("given,expected", [
+    ('aaa bbb.', 'aaa bbb.'),
+    ('aaa bbb. Ccc', 'aaa bbb.\nCcc'),
+    ('aaa bbb.  Ccc', 'aaa bbb.\nCcc'),
+    ('aaa bbb. ccc', 'aaa bbb. ccc'),
+])
+def test_sembreak(given:str, expected:str):
+    res = sembreak(given)
+    assert res == expected