Fix skipped day when UTC and tz are different days (#1183)

scrapinghub · Dec 18, 2023 · 5bddb9f · 5bddb9f
1 parent 3955530
commit 5bddb9f
Show file tree

Hide file tree

Showing 2 changed files with 24 additions and 7 deletions.
diff --git a/dateparser/parser.py b/dateparser/parser.py
@@ -563,18 +563,15 @@ def _correct_for_time_frame(self, dateobj, tz):
             # Convert dateobj to utc time to compare with self.now
             try:
                 tz = tz or get_timezone_from_tz_string(self.settings.TIMEZONE)
+                tz_offset = tz.utcoffset(dateobj)
             except pytz.UnknownTimeZoneError:
-                tz = None
+                tz_offset = timedelta(hours=0)
 
-            if tz:
-                dateobj_time = (dateobj - tz.utcoffset(dateobj)).time()
-            else:
-                dateobj_time = dateobj.time()
             if "past" in self.settings.PREFER_DATES_FROM:
-                if self.now.time() < dateobj_time:
+                if self.now < dateobj - tz_offset:
                     dateobj = dateobj + timedelta(days=-1)
             if "future" in self.settings.PREFER_DATES_FROM:
-                if self.now.time() > dateobj_time:
+                if self.now > dateobj - tz_offset:
                     dateobj = dateobj + timedelta(days=1)
 
         # Reset dateobj to the original value, thus removing any offset awareness that may

diff --git a/tests/test_date_parser.py b/tests/test_date_parser.py
@@ -1213,6 +1213,16 @@ def test_parsing_strings_containing_only_separator_tokens(
                 datetime(2021, 10, 19, 20, 0),
                 {"PREFER_DATES_FROM": "future"},
             ),
+            param(
+                "10pm EDT",
+                datetime(2021, 10, 20, 2, 0),
+                {"PREFER_DATES_FROM": "future"},
+            ),
+            param(
+                "8am AEDT",
+                datetime(2021, 10, 18, 21, 0),
+                {"PREFER_DATES_FROM": "past"},
+            ),
             param(
                 "11pm AEDT",
                 datetime(2021, 10, 19, 12, 0),
@@ -1223,6 +1233,16 @@ def test_parsing_strings_containing_only_separator_tokens(
                 datetime(2021, 10, 19, 20, 0),
                 {"PREFER_DATES_FROM": "future", "TIMEZONE": "EDT"},
             ),
+            param(
+                "10pm",
+                datetime(2021, 10, 20, 2, 0),
+                {"PREFER_DATES_FROM": "future", "TIMEZONE": "EDT"},
+            ),
+            param(
+                "8am",
+                datetime(2021, 10, 18, 21, 0),
+                {"PREFER_DATES_FROM": "past", "TIMEZONE": "AEDT"},
+            ),
             param(
                 "11pm",
                 datetime(2021, 10, 19, 12, 0),