Fix lowercase 'a' token formatting to lowercase meridiem

The 'a' format token was returning the literal string 'a' instead of a
lowercase meridiem indicator (am/pm). Only the uppercase 'A' token was
handled in _format_localizable_token, so 'a' fell through to the default
branch that returns the token unchanged.

This also broke round-tripping: the parser already expects lowercase
am/pm for the 'a' token, but formatting produced 'a' instead.

Author: vidigoat

src/pendulum/formatting/formatter.py

@@ -347,14 +347,16 @@ def _format_localizable_token(
             first_day = cast("int", locale.get("translations.week_data.first_day"))
 
             return locale.ordinalize((dt.day_of_week % 7 - first_day) % 7 + 1)
-        elif token == "A":
+        elif token in ["A", "a"]:
             key = "translations.day_periods"
             if dt.hour >= 12:
                 key += ".pm"
             else:
                 key += ".am"
 
-            return cast("str", locale.get(key))
+            meridiem = cast("str", locale.get(key))
+
+            return meridiem.lower() if token == "a" else meridiem
         else:
             return token
 

tests/formatting/test_formatter.py

@@ -140,6 +140,13 @@ def test_am_pm():
     assert f.format(d.set(hour=11), "A") == "AM"
 
 
+def test_lowercase_am_pm():
+    f = Formatter()
+    d = pendulum.datetime(2016, 8, 28, 23)
+    assert f.format(d, "a") == "pm"
+    assert f.format(d.set(hour=11), "a") == "am"
+
+
 def test_hour():
     f = Formatter()
     d = pendulum.datetime(2016, 8, 28, 7)