[compat] Add `compat_datetime_from_timestamp` (#11902)

Authored by: pzhlkj6612, seproDev

Co-authored-by: sepro <sepro@sepr0.com>

test/test_compat.py

@@ -1,6 +1,7 @@
 #!/usr/bin/env python3
 
 # Allow direct execution
+import datetime as dt
 import os
 import sys
 import unittest
@@ -12,7 +13,7 @@ import struct
 
 from yt_dlp import compat
 from yt_dlp.compat import urllib  # isort: split
-from yt_dlp.compat import compat_etree_fromstring, compat_expanduser
+from yt_dlp.compat import compat_etree_fromstring, compat_expanduser, compat_datetime_from_timestamp
 from yt_dlp.compat.urllib.request import getproxies
 
 
@@ -59,6 +60,45 @@ class TestCompat(unittest.TestCase):
     def test_struct_unpack(self):
         self.assertEqual(struct.unpack('!B', b'\x00'), (0,))
 
+    def test_compat_datetime_from_timestamp(self):
+        self.assertEqual(
+            compat_datetime_from_timestamp(0),
+            dt.datetime(1970, 1, 1, 0, 0, 0, tzinfo=dt.timezone.utc))
+        self.assertEqual(
+            compat_datetime_from_timestamp(1),
+            dt.datetime(1970, 1, 1, 0, 0, 1, tzinfo=dt.timezone.utc))
+        self.assertEqual(
+            compat_datetime_from_timestamp(3600),
+            dt.datetime(1970, 1, 1, 1, 0, 0, tzinfo=dt.timezone.utc))
+
+        self.assertEqual(
+            compat_datetime_from_timestamp(-1),
+            dt.datetime(1969, 12, 31, 23, 59, 59, tzinfo=dt.timezone.utc))
+        self.assertEqual(
+            compat_datetime_from_timestamp(-86400),
+            dt.datetime(1969, 12, 31, 0, 0, 0, tzinfo=dt.timezone.utc))
+
+        self.assertEqual(
+            compat_datetime_from_timestamp(0.5),
+            dt.datetime(1970, 1, 1, 0, 0, 0, 500000, tzinfo=dt.timezone.utc))
+        self.assertEqual(
+            compat_datetime_from_timestamp(1.000001),
+            dt.datetime(1970, 1, 1, 0, 0, 1, 1, tzinfo=dt.timezone.utc))
+        self.assertEqual(
+            compat_datetime_from_timestamp(-1.25),
+            dt.datetime(1969, 12, 31, 23, 59, 58, 750000, tzinfo=dt.timezone.utc))
+
+        self.assertEqual(
+            compat_datetime_from_timestamp(-1577923200),
+            dt.datetime(1920, 1, 1, 0, 0, 0, tzinfo=dt.timezone.utc))
+        self.assertEqual(
+            compat_datetime_from_timestamp(4102444800),
+            dt.datetime(2100, 1, 1, 0, 0, 0, tzinfo=dt.timezone.utc))
+
+        self.assertEqual(
+            compat_datetime_from_timestamp(173568960000),
+            dt.datetime(7470, 3, 8, 0, 0, 0, tzinfo=dt.timezone.utc))
+
 
 if __name__ == '__main__':
     unittest.main()

test/test_utils.py

@@ -101,11 +101,13 @@ from yt_dlp.utils import (
     remove_start,
     render_table,
     replace_extension,
+    datetime_round,
     rot47,
     sanitize_filename,
     sanitize_path,
     sanitize_url,
     shell_quote,
+    strftime_or_none,
     smuggle_url,
     str_to_int,
     strip_jsonp,
@@ -409,6 +411,25 @@ class TestUtil(unittest.TestCase):
         self.assertEqual(datetime_from_str('now+1day', precision='hour'), datetime_from_str('now+24hours', precision='auto'))
         self.assertEqual(datetime_from_str('now+23hours', precision='hour'), datetime_from_str('now+23hours', precision='auto'))
 
+    def test_datetime_round(self):
+        self.assertEqual(datetime_round(dt.datetime.strptime('1820-05-12T01:23:45Z', '%Y-%m-%dT%H:%M:%SZ')),
+                         dt.datetime(1820, 5, 12, tzinfo=dt.timezone.utc))
+        self.assertEqual(datetime_round(dt.datetime.strptime('1969-12-31T23:34:45Z', '%Y-%m-%dT%H:%M:%SZ'), 'hour'),
+                         dt.datetime(1970, 1, 1, 0, tzinfo=dt.timezone.utc))
+        self.assertEqual(datetime_round(dt.datetime.strptime('2024-12-25T01:23:45Z', '%Y-%m-%dT%H:%M:%SZ'), 'minute'),
+                         dt.datetime(2024, 12, 25, 1, 24, tzinfo=dt.timezone.utc))
+        self.assertEqual(datetime_round(dt.datetime.strptime('2024-12-25T01:23:45.123Z', '%Y-%m-%dT%H:%M:%S.%fZ'), 'second'),
+                         dt.datetime(2024, 12, 25, 1, 23, 45, tzinfo=dt.timezone.utc))
+        self.assertEqual(datetime_round(dt.datetime.strptime('2024-12-25T01:23:45.678Z', '%Y-%m-%dT%H:%M:%S.%fZ'), 'second'),
+                         dt.datetime(2024, 12, 25, 1, 23, 46, tzinfo=dt.timezone.utc))
+
+    def test_strftime_or_none(self):
+        self.assertEqual(strftime_or_none(-4722192000), '18200512')
+        self.assertEqual(strftime_or_none(0), '19700101')
+        self.assertEqual(strftime_or_none(1735084800), '20241225')
+        # Throws OverflowError
+        self.assertEqual(strftime_or_none(1735084800000), None)
+
     def test_daterange(self):
         _20century = DateRange('19000101', '20000101')
         self.assertFalse('17890714' in _20century)

yt_dlp/YoutubeDL.py

@@ -2717,11 +2717,7 @@ class YoutubeDL:
                 ('modified_timestamp', 'modified_date'),
         ):
             if info_dict.get(date_key) is None and info_dict.get(ts_key) is not None:
-                # Working around out-of-range timestamp values (e.g. negative ones on Windows,
+                info_dict[date_key] = strftime_or_none(info_dict[ts_key])
-                # see http://bugs.python.org/issue1646728)
-                with contextlib.suppress(ValueError, OverflowError, OSError):
-                    upload_date = dt.datetime.fromtimestamp(info_dict[ts_key], dt.timezone.utc)
-                    info_dict[date_key] = upload_date.strftime('%Y%m%d')
 
         if not info_dict.get('release_year'):
             info_dict['release_year'] = traverse_obj(info_dict, ('release_date', {lambda x: int(x[:4])}))

yt_dlp/compat/__init__.py

@@ -1,3 +1,4 @@
+import datetime as dt
 import os
 import xml.etree.ElementTree as etree
 
@@ -27,6 +28,13 @@ def compat_ord(c):
     return c if isinstance(c, int) else ord(c)
 
 
+def compat_datetime_from_timestamp(timestamp):
+    # Calling dt.datetime.fromtimestamp with negative timestamps throws error in Windows
+    # Ref: https://github.com/yt-dlp/yt-dlp/issues/5185, https://github.com/python/cpython/issues/81708,
+    # https://github.com/yt-dlp/yt-dlp/issues/6706#issuecomment-1496842642
+    return (dt.datetime.fromtimestamp(0, dt.timezone.utc) + dt.timedelta(seconds=timestamp))
+
+
 # Python 3.8+ does not honor %HOME% on windows, but this breaks compatibility with youtube-dl
 # See https://github.com/yt-dlp/yt-dlp/issues/792
 # https://docs.python.org/3/library/os.path.html#os.path.expanduser

yt_dlp/utils/_utils.py

@@ -47,6 +47,7 @@ import xml.etree.ElementTree
 from . import traversal
 
 from ..compat import (
+    compat_datetime_from_timestamp,
     compat_etree_fromstring,
     compat_expanduser,
     compat_HTMLParseError,
@@ -1376,6 +1377,7 @@ def datetime_round(dt_, precision='day'):
     if precision == 'microsecond':
         return dt_
 
+    time_scale = 1_000_000
     unit_seconds = {
         'day': 86400,
         'hour': 3600,
@@ -1383,8 +1385,8 @@ def datetime_round(dt_, precision='day'):
         'second': 1,
     }
     roundto = lambda x, n: ((x + n / 2) // n) * n
-    timestamp = roundto(calendar.timegm(dt_.timetuple()), unit_seconds[precision])
+    timestamp = roundto(calendar.timegm(dt_.timetuple()) + dt_.microsecond / time_scale, unit_seconds[precision])
-    return dt.datetime.fromtimestamp(timestamp, dt.timezone.utc)
+    return compat_datetime_from_timestamp(timestamp)
 
 
 def hyphenate_date(date_str):
@@ -2056,18 +2058,13 @@ def strftime_or_none(timestamp, date_format='%Y%m%d', default=None):
     datetime_object = None
     try:
         if isinstance(timestamp, (int, float)):  # unix timestamp
-            # Using naive datetime here can break timestamp() in Windows
+            datetime_object = compat_datetime_from_timestamp(timestamp)
-            # Ref: https://github.com/yt-dlp/yt-dlp/issues/5185, https://github.com/python/cpython/issues/94414
-            # Also, dt.datetime.fromtimestamp breaks for negative timestamps
-            # Ref: https://github.com/yt-dlp/yt-dlp/issues/6706#issuecomment-1496842642
-            datetime_object = (dt.datetime.fromtimestamp(0, dt.timezone.utc)
-                               + dt.timedelta(seconds=timestamp))
         elif isinstance(timestamp, str):  # assume YYYYMMDD
             datetime_object = dt.datetime.strptime(timestamp, '%Y%m%d')
         date_format = re.sub(  # Support %s on windows
             r'(?<!%)(%%)*%s', rf'\g<1>{int(datetime_object.timestamp())}', date_format)
         return datetime_object.strftime(date_format)
-    except (ValueError, TypeError, AttributeError):
+    except (ValueError, TypeError, AttributeError, OverflowError, OSError):
         return default