Default string-based "as_of" dates to a time of 11:59:59 PM UTC

Resolves #24

When someone instantiates a CladeTime object using string-based
date formats (YYYY-MM-DD) for sequence_as_of or tree_as_of,
set the corresponding timestamp to 11:59:59 to ensure that the
entire day is included when searching for S3 object versions
that match the date.

src/cladetime/cladetime.py

@@ -32,15 +32,17 @@ class CladeTime:
         sequence metadata files that will be used by CladeTime
         properties and methods. Can be a datetime object or a
         string in YYYY-MM-DD format, both of which will be treated as
-        UTC. The default value is the current time.
+        UTC. The default value is the current UTC time. Dates passed
+        as YYYY-MM-DD strings will be set to 11:59:59 PM UTC.
     tree_as_of : datetime.datetime | str | None
         Sets the version of the Nextstrain reference tree that will be
         used by CladeTime. Can be a datetime object or a string in
         YYYY-MM-DD format, both of which will be treated as UTC.
         The default value is :any:`sequence_as_of<sequence_as_of>`,
         unless sequence_as_of is before reference tree availability
-        (2024-08-01), in which case tree_as_of will default to the
-        current time.
+        (2024-08-01), in which case tree_as_of will default to current
+        time UTC. Dates passed as YYYY-MM-DD strings will be
+        set to 11:59:59 PM UTC.
 
     Attributes
     ----------

src/cladetime/sequence.py

@@ -274,10 +274,10 @@ def filter_metadata(
 
     # Apply filters for min and max sequence collection date, if applicable
     if collection_min_date is not None:
-        collection_min_date = _get_date(collection_min_date)
+        collection_min_date = _get_date(collection_min_date).replace(hour=0, minute=0, second=0)
         filtered_metadata = filtered_metadata.filter(pl.col("date") >= collection_min_date)
     if collection_max_date is not None:
-        collection_max_date = _get_date(collection_max_date)
+        collection_max_date = _get_date(collection_max_date).replace(hour=0, minute=0, second=0)
         filtered_metadata = filtered_metadata.filter(pl.col("date") <= collection_max_date)
 
     # Create state mappings based on state_format parameter, including a DC alias, since

src/cladetime/util/reference.py

@@ -28,7 +28,11 @@ def _get_date(original_date: datetime | str | None) -> datetime:
         new_date = original_date.replace(tzinfo=timezone.utc)
     elif isinstance(original_date, str):
         try:
-            new_date = datetime.strptime(original_date, "%Y-%m-%d").replace(tzinfo=timezone.utc)
+            new_date = (
+                datetime.strptime(original_date, "%Y-%m-%d")
+                .replace(hour=11, minute=59, second=59)
+                .replace(tzinfo=timezone.utc)
+            )
         except ValueError as e:
             raise ValueError(f"Invalid date format: {original_date}") from e
 

tests/integration/test_cladetime_integration.py

@@ -109,9 +109,9 @@ def test_assign_old_tree(test_file_path, tmp_path, test_sequences):
         expected_summary, old_assigned_clades.summary.collect(), check_column_order=False, check_row_order=False
     )
 
-    # metadata should reflect ncov metadata as of 2024-11-01
     assert old_assigned_clades.meta.get("sequence_as_of") == datetime(2024, 11, 1, tzinfo=timezone.utc)
-    assert old_assigned_clades.meta.get("tree_as_of") == datetime(2024, 8, 2, tzinfo=timezone.utc)
+    assert old_assigned_clades.meta.get("tree_as_of") == datetime(2024, 8, 2, 11, 59, 59, tzinfo=timezone.utc)
+    # nextclade metadata should reflect its state on tree_as_of (2024-08-02)
     assert old_assigned_clades.meta.get("nextclade_dataset_version") == "2024-07-17--12-57-03Z"
     assert old_assigned_clades.meta.get("nextclade_version_num") == "3.8.2"
     assert old_assigned_clades.meta.get("assignment_as_of") == "2024-11-01 00:00"

tests/unit/test_cladetime.py

@@ -26,8 +26,8 @@ def test_cladetime_no_args():
             # (metadata for reference trees started publishing in Aug, 2024)
             "2024-09-01",
             "2024-01-01",
-            datetime(2024, 9, 1, tzinfo=timezone.utc),
-            datetime(2024, 9, 1, tzinfo=timezone.utc),
+            datetime(2024, 9, 1, 11, 59, 59, tzinfo=timezone.utc),
+            datetime(2024, 9, 1, 11, 59, 59, tzinfo=timezone.utc),
         ),
         (
             # sequence_as_of set to current date, tree_as_of defaults to sequence_as_of
@@ -41,7 +41,7 @@ def test_cladetime_no_args():
             None,
             "2024-09-01",
             datetime(2025, 7, 13, 16, 21, 34, tzinfo=timezone.utc),
-            datetime(2024, 9, 1, tzinfo=timezone.utc),
+            datetime(2024, 9, 1, 11, 59, 59, tzinfo=timezone.utc),
         ),
         (
             # tree_as_of set to sequence_as_of
@@ -62,7 +62,7 @@ def test_cladetime_no_args():
             # defaults to current date
             "2023-12-21",
             None,
-            datetime(2023, 12, 21, tzinfo=timezone.utc),
+            datetime(2023, 12, 21, 11, 59, 59, tzinfo=timezone.utc),
             datetime(2025, 7, 13, 16, 21, 34, tzinfo=timezone.utc),
         ),
         (
@@ -85,7 +85,7 @@ def test_cladetime_no_args():
             # 2024-08-01, so it should revert to current date
             "2023-07-13",
             "2074-07",
-            datetime(2023, 7, 13, tzinfo=timezone.utc),
+            datetime(2023, 7, 13, 11, 59, 59, tzinfo=timezone.utc),
             datetime(2025, 7, 13, 16, 21, 34, tzinfo=timezone.utc),
         ),
     ],