feat: enhance error handling and configuration validation

- Added a global exception handler to log unhandled exceptions and return a generic 500 JSON response without exposing details to the client.
- Updated the configuration to validate the `DATABASE_URL` format, ensuring it starts with `sqlite://` or `postgresql://`, and log warnings for invalid formats.
- Introduced safe parsing for numeric environment variables (`HTTP_PORT`, `INIT_DATA_MAX_AGE_SECONDS`) with defaults on invalid values, including logging warnings for out-of-range values.
- Enhanced the duty schedule parser to enforce limits on the number of schedule rows and the length of full names and duty strings, raising appropriate errors when exceeded.
- Updated internationalization messages to include generic error responses for import failures and parsing issues, improving user experience.
- Added unit tests to verify the new error handling and configuration validation behaviors.

tests/test_utils.py

@@ -1,6 +1,6 @@
 """Unit tests for utils (dates, user, handover)."""
 
-from datetime import date
+from datetime import date, timedelta
 
 import pytest
 
@@ -62,6 +62,23 @@ def test_validate_date_range_from_after_to():
     assert exc_info.value.kind == "from_after_to"
 
 
+def test_validate_date_range_too_large():
+    """Range longer than MAX_DATE_RANGE_DAYS raises range_too_large."""
+    from duty_teller.utils.dates import MAX_DATE_RANGE_DAYS
+
+    # from 2023-01-01 to 2025-06-01 is more than 731 days
+    with pytest.raises(DateRangeValidationError) as exc_info:
+        validate_date_range("2023-01-01", "2025-06-01")
+    assert exc_info.value.kind == "range_too_large"
+
+    # Exactly MAX_DATE_RANGE_DAYS + 1 day
+    from_d = date(2024, 1, 1)
+    to_d = from_d + timedelta(days=MAX_DATE_RANGE_DAYS + 1)
+    with pytest.raises(DateRangeValidationError) as exc_info:
+        validate_date_range(from_d.isoformat(), to_d.isoformat())
+    assert exc_info.value.kind == "range_too_large"
+
+
 # --- user ---