From 120954252de0e81e98843eac41ac988c91e46fff Mon Sep 17 00:00:00 2001
From: Nick Jackson <nick@dxw.com>
Date: Tue, 25 Apr 2023 11:47:36 +0100
Subject: [PATCH] Improve robustness of SearchResult date parsing

This is now tested, and will log a warning if a non-empty unparseable string is detected.
---
 judgments/models.py            |  7 ++++++-
 judgments/tests/test_search.py | 34 ++++++++++++++++++++++++++++++++++
 2 files changed, 40 insertions(+), 1 deletion(-)

diff --git a/judgments/models.py b/judgments/models.py
index 12e899866..6df78509e 100644
--- a/judgments/models.py
+++ b/judgments/models.py
@@ -1,3 +1,4 @@
+import logging
 from os.path import dirname, join
 
 from caselawclient.Client import api_client
@@ -29,7 +30,11 @@ def __init__(
 
         try:
             self.date = dateparser.parse(date)
-        except ParserError:
+        except ParserError as e:
+            if date != "":
+                logging.warning(
+                    f'Unable to parse document date "{date}". Full error: {e}'
+                )
             self.date = None
         try:
             self.court = courts.get_by_code(court)
diff --git a/judgments/tests/test_search.py b/judgments/tests/test_search.py
index 46650adf1..a6b22b8c8 100644
--- a/judgments/tests/test_search.py
+++ b/judgments/tests/test_search.py
@@ -1,5 +1,9 @@
+from datetime import datetime
+from logging import WARNING
+from typing import Any
 from unittest.mock import patch
 
+import pytest
 from dateutil import parser as dateparser
 from django.test import TestCase
 from lxml import etree
@@ -159,3 +163,33 @@ def test_create_from_node_with_missing_elements(self, fake_client):
         self.assertEqual(None, search_result.neutral_citation)
         self.assertEqual(None, search_result.court)
         self.assertEqual(None, search_result.content_hash)
+
+
+class TestSearchResultInit:
+    @pytest.mark.parametrize(
+        "date_string, expected",
+        [
+            ["", None],
+            ["2023-05-09", datetime(2023, 5, 9, 0, 0)],
+            ["2023-04-05T06:54:00", datetime(2023, 4, 5, 6, 54)],
+            ["ffffff", None],
+        ],
+    )
+    def test_searchresult_date_parsing(self, date_string: str, expected: Any):
+        search_result = fake_search_result(date=date_string)
+
+        assert search_result.date == expected
+
+    def test_unparseable_non_empty_string_logs_warning(self, caplog):
+        caplog.set_level(WARNING)
+
+        fake_search_result(date="ffffff")
+
+        assert "Unable to parse document date" in caplog.text
+
+    def test_unparseable_empty_string_doesnt_log_warning(self, caplog):
+        caplog.set_level(WARNING)
+
+        fake_search_result(date="")
+
+        assert "Unable to parse document date" not in caplog.text