Add support for negative timestamps (#1060)

.gitignore

@@ -1,4 +1,5 @@
 *.py[cod]
+.python-version
 
 # C extensions
 *.so
@@ -47,6 +48,7 @@ docs/_build
 # Editors
 *.swp
 .idea
+.vscode/
 
 # Other
 raw_data

dateparser/conf.py

@@ -120,7 +120,7 @@ def _check_require_part(setting_name, setting_value):
 def _check_parsers(setting_name, setting_value):
     """Returns `True` if the provided list of parsers contains valid values"""
     existing_parsers = [
-        'timestamp', 'relative-time', 'custom-formats', 'absolute-time', 'no-spaces-time'
+        'timestamp', 'relative-time', 'custom-formats', 'absolute-time', 'no-spaces-time', 'negative-timestamp'
     ]  # FIXME: Extract the list of existing parsers from another place (#798)
     unknown_parsers = set(setting_value) - set(existing_parsers)
     if unknown_parsers:

dateparser/date.py

@@ -40,6 +40,7 @@
 RE_SANITIZE_APOSTROPHE = re.compile('|'.join(APOSTROPHE_LOOK_ALIKE_CHARS))
 
 RE_SEARCH_TIMESTAMP = re.compile(r'^(\d{10})(\d{3})?(\d{3})?(?![^.])')
+RE_SEARCH_NEGATIVE_TIMESTAMP = re.compile(r'^([-]\d{10})(\d{3})?(\d{3})?(?![^.])')
 
 
 def sanitize_spaces(date_string):
@@ -112,8 +113,12 @@ def sanitize_date(date_string):
     return date_string
 
 
-def get_date_from_timestamp(date_string, settings):
-    match = RE_SEARCH_TIMESTAMP.search(date_string)
+def get_date_from_timestamp(date_string, settings, negative=False):
+    if negative:
+        match = RE_SEARCH_NEGATIVE_TIMESTAMP.search(date_string)
+    else:
+         match = RE_SEARCH_TIMESTAMP.search(date_string)
+
     if match:
         seconds = int(match.group(1))
         millis = int(match.group(2) or 0)
@@ -166,6 +171,7 @@ def __init__(self, locale, date_string, date_formats, settings=None):
         self._translated_date_with_formatting = None
         self._parsers = {
             'timestamp': self._try_timestamp,
+            'negative-timestamp': self._try_negative_timestamp,
             'relative-time': self._try_freshness_parser,
             'custom-formats': self._try_given_formats,
             'absolute-time': self._try_absolute_parser,
@@ -185,12 +191,18 @@ def _parse(self):
         else:
             return None
 
-    def _try_timestamp(self):
+    def _try_timestamp_parser(self, negative=False):
         return DateData(
-            date_obj=get_date_from_timestamp(self.date_string, self._settings),
+            date_obj=get_date_from_timestamp(self.date_string, self._settings, negative=negative),
             period='time' if self._settings.RETURN_TIME_AS_PERIOD else 'day',
         )
 
+    def _try_timestamp(self):
+        return self._try_timestamp_parser()
+
+    def _try_negative_timestamp(self):
+        return self._try_timestamp_parser(negative=True)
+
     def _try_freshness_parser(self):
         try:
             return freshness_date_parser.get_date_data(self._get_translated_date(), self._settings)

docs/settings.rst

@@ -144,7 +144,7 @@ Language Detection
 Default Languages
 +++++++++++++++++
 
-``DEFAULT_LANGUAGES``: It is a ``list`` of language codes in ISO 639 that will be used as default 
+``DEFAULT_LANGUAGES``: It is a ``list`` of language codes in ISO 639 that will be used as default
 languages for parsing when language detection fails. eg. ["en", "fr"]:
 
     >>> from dateparser import parse
@@ -181,6 +181,9 @@ The following parsers exist:
     followed by additional digits or a period (``.``), those first 10 digits
     are interpreted as `Unix time <https://en.wikipedia.org/wiki/Unix_time>`_.
 
+-    ``'negative-timestamp'``: ``'timestamp'`` for negative timestamps. For
+    example, parses ``-186454800000`` as ``1964-02-03T23:00:00``.
+
 -   ``'relative-time'``: Parses dates and times expressed in relation to the
     current date and time (e.g. “1 day ago”, “in 2 weeks”).
 

tests/test_date.py

@@ -726,6 +726,34 @@ def test_timestamp_in_microseconds(self):
             datetime.fromtimestamp(1570308760).replace(microsecond=263111)
         )
 
+    @parameterized.expand([
+        param(
+            input_timestamp='-1570308760',
+            negative=True,
+            result=datetime.fromtimestamp(-1570308760)
+        ),
+        param(
+            input_timestamp='-1570308760',
+            negative=False,
+            result=None
+        ),
+        param(
+            input_timestamp='1570308760',
+            negative=True,
+            result=None
+        ),
+        param(
+            input_timestamp='1570308760',
+            negative=False,
+            result=datetime.fromtimestamp(1570308760)
+        )
+    ])
+    def test_timestamp_with_negative(self, input_timestamp, negative, result):
+        self.assertEqual(
+            date.get_date_from_timestamp(input_timestamp, None, negative=negative),
+            result
+        )
+
     @parameterized.expand([
         param(date_string='15703087602631'),
         param(date_string='157030876026xx'),

tests/test_date_parser.py

@@ -645,6 +645,18 @@ def test_parse_timestamp(self, date_string, expected):
         self.then_date_obj_exactly_is(expected)
         self.then_period_is('day')
 
+    @parameterized.expand([
+        param('-1484823450', expected=datetime(1922, 12, 13, 13, 2, 30)),
+        param('-1436745600000', expected=datetime(1924, 6, 22, 0, 0)),
+        param('-1015673450000001', expected=datetime(1937, 10, 25, 12, 29, 10, 1))
+    ])
+    def test_parse_negative_timestamp(self, date_string, expected):
+        self.given_local_tz_offset(0)
+        self.given_parser(settings={'TO_TIMEZONE': 'UTC', 'PARSERS': ['negative-timestamp']})
+        self.when_date_is_parsed(date_string)
+        self.then_date_obj_exactly_is(expected)
+        self.then_period_is('day')
+
     @parameterized.expand([
         # Epoch timestamps.
         param('1484823450', expected=datetime(2017, 1, 19, 10, 57, 30)),