Improved normalization

libpurple_to_markdown/__main__.py

@@ -46,7 +46,7 @@ MAX_AVERAGE_MESSAGES_PER_PERIOD = 120
 
 def group_messages_by_period(messages: Iterable[Message]) -> dict[str, list[Message]]:
     possible_period_keys = [
-        (lambda msg: 'Full History'),
+        (lambda msg: 'full'),
         year_period_key,
         year_quarter_period_key,
         year_and_month_period_key,
@@ -108,6 +108,11 @@ def main():
 
     for chat_id, messages_in_chat_original in messages_by_chat_id.items():
         messages_in_chat = merge_adjacent_messages(messages_in_chat_original)
+        if len(messages_in_chat) <= 2:
+            logger.info(
+                '  "%s": Skipped due to too few messages', chat_id)
+            continue
+
         messages_by_period = group_messages_by_period(messages_in_chat)
         logger.info(
             '  "%s": %d messages, %d periods (%d msg/period avg)',
@@ -118,7 +123,8 @@ def main():
         )
 
         for period_key, messages in messages_by_period.items():
-            output_file = args.output / chat_id / f'{period_key}.md'
+            file_escaped_chat_id = chat_id.replace(' ','-')
+            output_file = args.output / chat_id / f'{file_escaped_chat_id}-{period_key}.md'
             output_file.parent.mkdir(exist_ok=True)
             logger.info('Writing % 5d messages to %s', len(messages), output_file)
             with open(output_file, 'w') as f:

libpurple_to_markdown/synctech_sms.py

@@ -17,7 +17,17 @@ from .data import MYSELF, Message
 
 logger = logging.getLogger(__name__)
 
+def is_named_number(num: str) -> str:
+    try:
+        int(num.removeprefix('+').replace(' ', ''))
+        return False
+    except ValueError:
+        return True
+
 def normalize_phone_number(num: str) -> str:
+    if is_named_number(num):
+        return num
+
     num = num.replace(' ', '')
     if num.startswith('00'):
         num = '+' + num.removeprefix('00')
@@ -33,14 +43,24 @@ def sms_soup_to_message(soup: bs4.BeautifulSoup) -> Message:
     sent_at = datetime.datetime.fromtimestamp(int(soup['date']) / 1000)
 
     phone_num = normalize_phone_number(soup['address'])
-    contact_name = soup.get('contact_name') or phone_num
+
+    if is_named_number(phone_num):
+        contact_name = phone_num
+        phone_num = None
+    else:
+        contact_name = soup.get('contact_name') or phone_num
+    if contact_name == '(Unknown)':
+        contact_name = None
+
     if soup['type'] == '2':
         sender = MYSELF
     else:
-        sender = contact_name
+        sender = contact_name or phone_num
 
     text = soup['body']
-    chat_id = f'SMS {contact_name} {phone_num}'
+
+    chat_id_parts = ['SMS', contact_name or phone_num]
+    chat_id = ' '.join(p for p in chat_id_parts if p)
     return Message(sent_at, sender, text, chat_id=chat_id)
 
 

test/test_synctech_sms.py

@@ -1,5 +1,6 @@
 from libpurple_to_markdown import synctech_sms
 
+
 def test_normalize_phone_number():
     assert synctech_sms.normalize_phone_number('+45 12 34 56 78') == '+4512345678'
     assert synctech_sms.normalize_phone_number('+4512345678') == '+4512345678'
@@ -8,3 +9,6 @@ def test_normalize_phone_number():
     assert synctech_sms.normalize_phone_number('12 34 56 78') == '+4512345678'
     assert synctech_sms.normalize_phone_number('441234567890') == '+441234567890'
     assert synctech_sms.normalize_phone_number('004712345678') == '+4712345678'
+
+def test_dont_normalize_weird():
+    assert synctech_sms.normalize_phone_number('Midttrafik') == 'Midttrafik'