Improved text formatting

libpurple_to_markdown/__init__.py

@@ -88,9 +88,10 @@ def is_adjacent_messages(first: Message, second: Message) -> bool:
         and second.sent_at - first.sent_at <= MSG_ADJACENTCY_DIST
     )
 
+PUNCTUATION = ('.','?','!',',',':',';')
 
 def merge_texts(text1: str, text2: str) -> str:
-    punctuated = text1.endswith('.?!,:')
+    punctuated = text1.endswith(PUNCTUATION)
     # return text1 + (' ' if punctuated else '. ') + text2
     return text1 + ('' if punctuated else '.') + '\n' + text2
 

libpurple_to_markdown/markdown.py

@@ -7,6 +7,8 @@ from .data import Message
 def normalize_line(line: str) -> str:
     line = re.sub(r'(<[\w ]+>)', r'`\1`', line)
     line = re.sub(r'(\$\$\$)', r'`\1`', line)
+    line = re.sub(r'\s\s', ' ', line)
+    line = re.sub(r'\s+([.:;!?])', r'\1', line)
     return line.strip()
 
 def format_message_as_citation(out: list[str], msg: Message) -> None:

test/test_join.py

@@ -0,0 +1,9 @@
+import datetime
+
+from libpurple_to_markdown import merge_texts
+
+def test_merge():
+    assert merge_texts('Hello World', 'Hello World.') == 'Hello World.\nHello World.'
+    assert merge_texts('Hello World.', 'Hello World.') == 'Hello World.\nHello World.'
+    assert merge_texts('Hello World?', 'Hello World.') == 'Hello World?\nHello World.'
+