Helper to set_last_modified directly

Author: Mike

documentation/main.md

@@ -922,6 +922,16 @@ resource, it is possible to supply the parameter `data_updated` eg.
 
     resource.update_in_hdx(data_updated=True)
 
+If you need to set a specific date for date of update (`last_modified`), you
+can call the following:
+
+    resource.set_date_data_updated(date)
+
+`date` can be a datetime object or string. You can retrieve the date of update
+(`last_modified`) using the getter:
+
+    date = resource.get_date_data_updated()
+
 If the method **set_file_to_upload** is used to supply a file, the resource 
 `last_modified` field is set to now automatically regardless of the value of 
 `data_updated` or whether **mark_data_updated** has been called.

src/hdx/data/dataset.py

@@ -290,7 +290,7 @@ def add_update_resource(
                 updated_resource.set_file_to_upload(
                     resource.get_file_to_upload()
                 )
-            if resource.is_data_updated():
+            if resource.is_marked_data_updated():
                 updated_resource.mark_data_updated()
 
     def add_update_resources(
@@ -334,7 +334,7 @@ def add_update_resources(
                 updated_resource.set_file_to_upload(
                     resource.get_file_to_upload()
                 )
-            if resource.is_data_updated():
+            if resource.is_marked_data_updated():
                 updated_resource.mark_data_updated()
         for resource_index in updated_resource_no_matches:
             resource = resource_objects[resource_index]
@@ -1256,7 +1256,7 @@ def get_reference_period(
         Returns:
             Dict: Dictionary of date information
         """
-        return DateHelper.get_date_info(
+        return DateHelper.get_reference_period_info(
             self.data.get("dataset_date"), date_format, today
         )
 
@@ -1282,7 +1282,7 @@ def set_reference_period(
         Returns:
             None
         """
-        self.data["dataset_date"] = DateHelper.get_hdx_date(
+        self.data["dataset_date"] = DateHelper.get_hdx_reference_period(
             startdate, enddate, ongoing, ignore_timeinfo
         )
 
@@ -1300,7 +1300,10 @@ def set_reference_period_year_range(
         Returns:
             List[int]: The start and end year if supplied or sorted list of years
         """
-        self.data["dataset_date"], retval = DateHelper.get_hdx_date_from_years(
+        (
+            self.data["dataset_date"],
+            retval,
+        ) = DateHelper.get_hdx_reference_period_from_years(
             dataset_year, dataset_end_year
         )
         return retval

src/hdx/data/date_helper.py

@@ -9,30 +9,74 @@
 
 class DateHelper:
     @staticmethod
-    def get_date_info(
-        hdx_date: Dict,
+    def get_hdx_date(
+        date: Union[datetime, str], ignore_timeinfo: bool, max=False
+    ):
+        """Get an HDX date as a string from a datetime.datetime object.
+
+        Args:
+            date (Union[datetime, str]): Date as datetime or string
+            ignore_timeinfo (bool): Ignore time and time zone of date. Defaults to True.
+
+        Returns:
+            str: HDX date as a string
+        """
+        if ignore_timeinfo:
+            timezone_handling = 0
+        else:
+            timezone_handling = 3
+
+        if isinstance(date, str):
+            date = parse_date(date, timezone_handling=timezone_handling)
+
+        if ignore_timeinfo:
+            if max:
+                date = date.replace(
+                    hour=23,
+                    minute=59,
+                    second=59,
+                    microsecond=0,
+                    tzinfo=None,
+                )
+            else:
+                date = date.replace(
+                    hour=0, minute=0, second=0, microsecond=0, tzinfo=None
+                )
+        else:
+            date = date.astimezone(timezone.utc).replace(
+                microsecond=0, tzinfo=None
+            )
+        return date.isoformat()
+
+    @staticmethod
+    def get_reference_period_info(
+        hdx_reference_period: Dict,
         date_format: Optional[str] = None,
         today: datetime = now_utc(),
     ) -> Dict:
-        """Get date as datetimes and strings in specified format. If no format is
-        supplied, the ISO 8601 format is used. Returns a dictionary containing keys
-        startdate (start date as datetime), enddate (end date as datetime),
-        startdate_str (start date as string), enddate_str (end date as string) and
-        ongoing (whether the end date is a rolls forward every day).
+        """Get reference period as datetimes and strings in specified format.
+        If no format is supplied, the ISO 8601 format is used. Returns a
+        dictionary containing keys startdate (start date as datetime), enddate
+        (end date as datetime), startdate_str (start date as string),
+        enddate_str (enddate as string) and ongoing (whether the end date rolls
+        forward every day).
 
         Args:
-            hdx_date (str): Input date
+            hdx_reference_period (str): Input reference period
             date_format (Optional[str]): Date format. None is taken to be ISO 8601. Defaults to None.
             today (datetime): Date to use for today. Defaults to now_utc().
 
         Returns:
             Dict: Dictionary of date information
         """
         result = dict()
-        if hdx_date:
-            if hdx_date[0] == "[" and hdx_date[-1] == "]":
-                hdx_date = hdx_date[1:-1]
-            dataset_dates = hdx_date.split(" TO ")
+        if hdx_reference_period:
+            if (
+                hdx_reference_period[0] == "["
+                and hdx_reference_period[-1] == "]"
+            ):
+                hdx_reference_period = hdx_reference_period[1:-1]
+            dataset_dates = hdx_reference_period.split(" TO ")
             startdate = parse_date(dataset_dates[0])
             enddate = dataset_dates[1]
             if enddate == "*":
@@ -59,15 +103,16 @@ def get_date_info(
             result["ongoing"] = ongoing
         return result
 
-    @staticmethod
-    def get_hdx_date(
+    @classmethod
+    def get_hdx_reference_period(
+        cls,
         startdate: Union[datetime, str],
         enddate: Union[datetime, str, None] = None,
         ongoing: bool = False,
         ignore_timeinfo: bool = True,
     ) -> str:
-        """Get an HDX date from either datetime.datetime objects or strings with option
-        to set ongoing.
+        """Get an HDX reference period from either datetime.datetime objects or
+        strings with option to set ongoing.
 
         Args:
             startdate (Union[datetime, str]): Dataset start date
@@ -76,65 +121,33 @@ def get_hdx_date(
             ignore_timeinfo (bool): Ignore time and time zone of date. Defaults to True.
 
         Returns:
-            str: HDX date
+            str: HDX reference period
         """
-        if ignore_timeinfo:
-            timezone_handling = 0
-        else:
-            timezone_handling = 3
-
-        def get_date_str(dt, max=False):
-            if ignore_timeinfo:
-                if max:
-                    dt = dt.replace(
-                        hour=23,
-                        minute=59,
-                        second=59,
-                        microsecond=0,
-                        tzinfo=None,
-                    )
-                else:
-                    dt = dt.replace(
-                        hour=0, minute=0, second=0, microsecond=0, tzinfo=None
-                    )
-            else:
-                dt = dt.astimezone(timezone.utc).replace(
-                    microsecond=0, tzinfo=None
-                )
-            return dt.isoformat()
-
-        if isinstance(startdate, str):
-            startdate = parse_date(
-                startdate, timezone_handling=timezone_handling
-            )
-        startdate = get_date_str(startdate)
+        startdate = cls.get_hdx_date(startdate, ignore_timeinfo)
         if ongoing:
             enddate = "*"
         else:
             if not enddate:
                 enddate = startdate
             else:
-                if isinstance(enddate, str):
-                    enddate = parse_date(
-                        enddate, timezone_handling=timezone_handling
-                    )
-                enddate = get_date_str(enddate, max=True)
+                enddate = cls.get_hdx_date(enddate, ignore_timeinfo, max=True)
         return f"[{startdate} TO {enddate}]"
 
     @classmethod
-    def get_hdx_date_from_years(
+    def get_hdx_reference_period_from_years(
         cls,
         startyear: Union[str, int, Iterable],
         endyear: Union[str, int, None] = None,
     ) -> Tuple[str, List[int]]:
-        """Get an HDX date from an iterable of years or a start and end year.
+        """Get an HDX reference period from an iterable of years or a start and
+        end year.
 
         Args:
             startyear (Union[str, int, Iterable]): Start year given as string or int or range in an iterable
             endyear (Union[str, int, None]): End year given as string or int
 
         Returns:
-            Tuple[str,List[int]]: (HDX date, the start and end year if supplied or sorted list of years)
+            Tuple[str,List[int]]: (HDX reference period, the start and end year if supplied or sorted list of years)
         """
         retval = list()
         if isinstance(startyear, str):
@@ -152,6 +165,8 @@ def get_hdx_date_from_years(
         startdate = datetime(startyear, 1, 1, tzinfo=timezone.utc)
         enddate = datetime(endyear, 12, 31, 23, 59, 59, tzinfo=timezone.utc)
         return (
-            cls.get_hdx_date(startdate, enddate, ignore_timeinfo=False),
+            cls.get_hdx_reference_period(
+                startdate, enddate, ignore_timeinfo=False
+            ),
             retval,
         )

src/hdx/data/filestore_helper.py

@@ -92,7 +92,7 @@ def dataset_merge_filestore_resource(
         if file_to_upload:
             resource.set_file_to_upload(file_to_upload)
             filestore_resources[resource_index] = file_to_upload
-        data_updated = updated_resource.is_data_updated()
+        data_updated = updated_resource.is_marked_data_updated()
         merge_two_dictionaries(resource, updated_resource)
         cls.resource_check_required_fields(
             resource, check_upload=True, **kwargs

src/hdx/data/resource.py

@@ -11,7 +11,7 @@
 from hdx.data.date_helper import DateHelper
 from hdx.data.hdxobject import HDXError, HDXObject
 from hdx.data.resource_view import ResourceView
-from hdx.utilities.dateparse import now_utc
+from hdx.utilities.dateparse import now_utc, parse_date
 from hdx.utilities.downloader import Download
 from hdx.utilities.typehint import ListTuple
 from hdx.utilities.uuid import is_valid_uuid
@@ -122,7 +122,7 @@ def get_date_of_resource(
         Returns:
             Dict: Dictionary of date information
         """
-        return DateHelper.get_date_info(
+        return DateHelper.get_reference_period_info(
             self.data.get("daterange_for_data"), date_format, today
         )
 
@@ -146,7 +146,7 @@ def set_date_of_resource(
         Returns:
             None
         """
-        self.data["daterange_for_data"] = DateHelper.get_hdx_date(
+        self.data["daterange_for_data"] = DateHelper.get_hdx_reference_period(
             startdate, enddate, ignore_timeinfo=ignore_timeinfo
         )
 
@@ -389,6 +389,7 @@ def update_in_hdx(self, **kwargs: Any) -> None:
             **kwargs: See below
             operation (string): Operation to perform eg. patch. Defaults to update.
             data_updated (bool): If True, set last_modified to now. Defaults to False.
+            date_data_updated (datetime): Date to use for last_modified. Default to None.
 
         Returns:
             None
@@ -410,6 +411,7 @@ def create_in_hdx(self, **kwargs: Any) -> None:
         Args:
             **kwargs: See below
             data_updated (bool): If True, set last_modified to now. Defaults to False.
+            date_data_updated (datetime): Date to use for last_modified. Default to None.
 
         Returns:
             None
@@ -745,11 +747,11 @@ def mark_broken(self) -> None:
         else:
             logger.debug(result)
 
-    def is_data_updated(self) -> bool:
-        """Return if the resource's data is updated
+    def is_marked_data_updated(self) -> bool:
+        """Return if the resource's data is marked to be updated
 
         Returns:
-            bool: Whether resource's data is updated
+            bool: Whether resource's data is marked to be updated
         """
         return self.data_updated
 
@@ -760,3 +762,27 @@ def mark_data_updated(self) -> None:
             None
         """
         self.data_updated = True
+
+    def get_date_data_updated(self) -> datetime:
+        """Get date resource data was updated
+
+        Returns:
+            datetime: Date resource data was updated
+        """
+        return parse_date(self.data["last_modified"])
+
+    def set_date_data_updated(
+        self, date: Union[datetime, str], ignore_timeinfo: bool = True
+    ) -> None:
+        """Set date resource data was updated
+
+        Args:
+            date (Union[datetime, str]): Date resource data was updated
+            ignore_timeinfo (bool): Ignore time and time zone of date. Defaults to True.
+
+        Returns:
+            None
+        """
+        self.data["last_modified"] = DateHelper.get_hdx_date(
+            date, ignore_timeinfo
+        )

tests/hdx/data/test_dataset_core.py

@@ -713,7 +713,7 @@ def test_update_in_hdx(self, configuration, post_update, date_pattern):
         assert dataset["state"] == "active"
         assert len(dataset.resources) == 3
         resource = dataset.get_resource()
-        assert resource.is_data_updated() is False
+        assert resource.is_marked_data_updated() is False
         match = date_pattern.search(resource["last_modified"])
         assert match
 

tests/hdx/data/test_resource.py

@@ -13,6 +13,7 @@
 from hdx.api.configuration import Configuration
 from hdx.data.hdxobject import HDXError
 from hdx.data.resource import Resource
+from hdx.utilities.dateparse import parse_date
 from hdx.utilities.dictandlist import merge_two_dictionaries
 from hdx.utilities.downloader import DownloadError
 
@@ -845,9 +846,9 @@ def test_update_in_hdx(self, configuration, date_pattern, post_update):
         resource = Resource(resource_data)
         resource.mark_data_updated()
         assert resource.data_updated is True
-        assert resource.is_data_updated() is True
+        assert resource.is_marked_data_updated() is True
         resource.create_in_hdx()
-        assert resource.is_data_updated() is False
+        assert resource.is_marked_data_updated() is False
         assert resource["id"] == "74b74ae1-df0c-4716-829f-4f939a046811"
         assert resource.get_file_type() == "xlsx"
         assert resource["state"] == "active"
@@ -1010,3 +1011,16 @@ def test_mark_broken(self, configuration, post_broken):
         assert resource["package_id"] == "6f36a41c-f126-4b18-aaaf-6c2ddfbc5d4d"
         resource.mark_broken()
         assert resource.is_broken() is True
+
+    def test_date_data_updated(self, configuration, read):
+        resource = Resource.read_from_hdx(
+            "74b74ae1-df0c-4716-829f-4f939a046811"
+        )
+        assert resource["id"] == "de6549d8-268b-4dfe-adaf-a4ae5c8510d5"
+        assert resource["name"] == "MyResource1"
+        assert resource["package_id"] == "6f36a41c-f126-4b18-aaaf-6c2ddfbc5d4d"
+        resource.set_date_data_updated("2020-03-02")
+        expected = "2020-03-02T00:00:00+00:00"
+        assert resource.get_date_data_updated().isoformat() == expected
+        resource.set_date_data_updated(parse_date("2020-03-02"))
+        assert resource.get_date_data_updated().isoformat() == expected