Merge pull request #131 from eseglem/bugfix/stac-pydantic-130-validate-item-properties

Adjust ItemProperties Validation.

Author: vincentsarago

CHANGELOG.txt

@@ -8,6 +8,8 @@
 - Enforce required `type` key for `Collection` and `Catalog` models
 - Add queryables link relation type (#123, @constantinius)
 - Fix STAC API Query Extension operator names from ne->neq, le->lte, and ge->gte (#120, @philvarner)
+- Better **datetime** parsing/validation by using Pydantic native types and remove `ciso8601` requirement (#131, @eseglem)
+- move datetime validation in `StacCommonMetadata` model definition (#131, @eseglem)
 
 3.0.0 (2024-01-25)
 ------------------

pyproject.toml

@@ -20,12 +20,7 @@ keywords=["stac", "pydantic", "validation"]
 authors=[{ name = "Arturo Engineering", email = "[email protected]"}]
 license= { text = "MIT" }
 requires-python=">=3.8"
-dependencies = [
-        "click>=8.1.7",
-        "pydantic>=2.4.1",
-        "geojson-pydantic>=1.0.0",
-        "ciso8601~=2.3",
-]
+dependencies = ["click>=8.1.7", "pydantic>=2.4.1", "geojson-pydantic>=1.0.0"]
 dynamic = ["version", "readme"]
 
 [project.scripts]
@@ -37,7 +32,6 @@ repository ="https://github.com/stac-utils/stac-pydantic.git"
 
 [project.optional-dependencies]
 dev = [
-        "arrow>=1.2.3",
         "pytest>=7.4.2",
         "pytest-cov>=4.1.0",
         "pytest-icdiff>=0.8",

stac_pydantic/api/search.py

@@ -1,22 +1,21 @@
 from datetime import datetime as dt
 from typing import Any, Dict, List, Optional, Tuple, Union, cast
 
-from ciso8601 import parse_rfc3339
-from geojson_pydantic.geometries import GeometryCollection  # type: ignore
 from geojson_pydantic.geometries import (
+    GeometryCollection,
     LineString,
     MultiLineString,
     MultiPoint,
     MultiPolygon,
     Point,
     Polygon,
 )
-from pydantic import BaseModel, Field, field_validator, model_validator
+from pydantic import BaseModel, Field, TypeAdapter, field_validator, model_validator
 
 from stac_pydantic.api.extensions.fields import FieldsExtension
 from stac_pydantic.api.extensions.query import Operator
 from stac_pydantic.api.extensions.sort import SortExtension
-from stac_pydantic.shared import BBox
+from stac_pydantic.shared import BBox, UtcDatetime
 
 Intersection = Union[
     Point,
@@ -28,6 +27,8 @@
     GeometryCollection,
 ]
 
+SearchDatetime = TypeAdapter(Optional[UtcDatetime])
+
 
 class Search(BaseModel):
     """
@@ -43,23 +44,18 @@ class Search(BaseModel):
     datetime: Optional[str] = None
     limit: int = 10
 
+    # Private properties to store the parsed datetime values. Not part of the model schema.
+    _start_date: Optional[dt] = None
+    _end_date: Optional[dt] = None
+
+    # Properties to return the private values
     @property
     def start_date(self) -> Optional[dt]:
-        values = (self.datetime or "").split("/")
-        if len(values) == 1:
-            return None
-        if values[0] == ".." or values[0] == "":
-            return None
-        return parse_rfc3339(values[0])
+        return self._start_date
 
     @property
     def end_date(self) -> Optional[dt]:
-        values = (self.datetime or "").split("/")
-        if len(values) == 1:
-            return parse_rfc3339(values[0])
-        if values[1] == ".." or values[1] == "":
-            return None
-        return parse_rfc3339(values[1])
+        return self._end_date
 
     # Check https://docs.pydantic.dev/dev-v2/migration/#changes-to-validators for more information.
     @model_validator(mode="before")
@@ -102,32 +98,43 @@ def validate_bbox(cls, v: BBox) -> BBox:
 
     @field_validator("datetime")
     @classmethod
-    def validate_datetime(cls, v: str) -> str:
-        if "/" in v:
-            values = v.split("/")
-        else:
-            # Single date is interpreted as end date
-            values = ["..", v]
-
-        dates: List[dt] = []
-        for value in values:
-            if value == ".." or value == "":
-                continue
-
-            dates.append(parse_rfc3339(value))
+    def validate_datetime(cls, value: str) -> str:
+        # Split on "/" and replace no value or ".." with None
+        values = [v if v and v != ".." else None for v in value.split("/")]
 
+        # If there are more than 2 dates, it's invalid
         if len(values) > 2:
             raise ValueError(
-                "Invalid datetime range, must match format (begin_date, end_date)"
+                "Invalid datetime range. Too many values. Must match format: {begin_date}/{end_date}"
             )
 
-        if not {"..", ""}.intersection(set(values)):
-            if dates[0] > dates[1]:
-                raise ValueError(
-                    "Invalid datetime range, must match format (begin_date, end_date)"
-                )
+        # If there is only one date, insert a None for the start date
+        if len(values) == 1:
+            values.insert(0, None)
+
+        # Cast because pylance gets confused by the type adapter and annotated type
+        dates = cast(
+            List[Optional[dt]],
+            [
+                # Use the type adapter to validate the datetime strings, strict is necessary
+                # due to pydantic issues #8736 and #8762
+                SearchDatetime.validate_strings(v, strict=True) if v else None
+                for v in values
+            ],
+        )
+
+        # If there is a start and end date, check that the start date is before the end date
+        if dates[0] and dates[1] and dates[0] > dates[1]:
+            raise ValueError(
+                "Invalid datetime range. Begin date after end date. "
+                "Must match format: {begin_date}/{end_date}"
+            )
 
-        return v
+        # Store the parsed dates
+        cls._start_date = dates[0]
+        cls._end_date = dates[1]
+        # Return the original string value
+        return value
 
     @property
     def spatial_filter(self) -> Optional[Intersection]:

stac_pydantic/item.py

@@ -1,24 +1,15 @@
-from datetime import datetime as dt
-from typing import Any, Dict, List, Optional, Union
+from typing import Any, Dict, List, Optional
 
-from ciso8601 import parse_rfc3339
 from geojson_pydantic import Feature
-from pydantic import (
-    AnyUrl,
-    ConfigDict,
-    Field,
-    field_serializer,
-    model_serializer,
-    model_validator,
-)
+from pydantic import AnyUrl, ConfigDict, Field, model_serializer, model_validator
 
 from stac_pydantic.links import Links
 from stac_pydantic.shared import (
-    DATETIME_RFC339,
     SEMVER_REGEX,
     Asset,
     StacBaseModel,
     StacCommonMetadata,
+    UtcDatetime,
 )
 from stac_pydantic.version import STAC_VERSION
 
@@ -28,39 +19,12 @@ class ItemProperties(StacCommonMetadata):
     https://github.com/radiantearth/stac-spec/blob/v1.0.0/item-spec/item-spec.md#properties-object
     """
 
-    datetime: Union[dt, str] = Field(..., alias="datetime")
+    # Overide the datetime field to be required
+    datetime: Optional[UtcDatetime]
 
     # Check https://docs.pydantic.dev/dev-v2/migration/#changes-to-config for more information.
     model_config = ConfigDict(extra="allow")
 
-    @model_validator(mode="before")
-    @classmethod
-    def validate_datetime(cls, data: Dict[str, Any]) -> Dict[str, Any]:
-        datetime = data.get("datetime")
-        start_datetime = data.get("start_datetime")
-        end_datetime = data.get("end_datetime")
-
-        if not datetime or datetime == "null":
-            if not start_datetime and not end_datetime:
-                raise ValueError(
-                    "start_datetime and end_datetime must be specified when datetime is null"
-                )
-
-        if isinstance(datetime, str):
-            data["datetime"] = parse_rfc3339(datetime)
-
-        if isinstance(start_datetime, str):
-            data["start_datetime"] = parse_rfc3339(start_datetime)
-
-        if isinstance(end_datetime, str):
-            data["end_datetime"] = parse_rfc3339(end_datetime)
-
-        return data
-
-    @field_serializer("datetime")
-    def serialize_datetime(self, v: dt, _info: Any) -> str:
-        return v.strftime(DATETIME_RFC339)
-
 
 class Item(Feature, StacBaseModel):
     """

stac_pydantic/shared.py

@@ -1,9 +1,17 @@
-from datetime import datetime
+from datetime import timezone
 from enum import Enum, auto
 from typing import Any, Dict, List, Optional, Tuple, Union
 from warnings import warn
 
-from pydantic import BaseModel, ConfigDict, Field
+from pydantic import (
+    AfterValidator,
+    AwareDatetime,
+    BaseModel,
+    ConfigDict,
+    Field,
+    model_validator,
+)
+from typing_extensions import Annotated, Self
 
 from stac_pydantic.utils import AutoValueEnum
 
@@ -15,9 +23,14 @@
 
 SEMVER_REGEX = r"^(0|[1-9]\d*)\.(0|[1-9]\d*)\.(0|[1-9]\d*)(?:-((?:0|[1-9]\d*|\d*[a-zA-Z-][0-9a-zA-Z-]*)(?:\.(?:0|[1-9]\d*|\d*[a-zA-Z-][0-9a-zA-Z-]*))*))?(?:\+([0-9a-zA-Z-]+(?:\.[0-9a-zA-Z-]+)*))?$"
 
-# https://tools.ietf.org/html/rfc3339#section-5.6
-# Unused, but leaving it here since it's used by dependencies
-DATETIME_RFC339 = "%Y-%m-%dT%H:%M:%SZ"
+# Allows for some additional flexibility in the input datetime format. As long as
+# the input value has timezone information, it will be converted to UTC timezone.
+UtcDatetime = Annotated[
+    # Input value must be in a format which has timezone information
+    AwareDatetime,
+    # Convert the input value to UTC timezone
+    AfterValidator(lambda d: d.astimezone(timezone.utc)),
+]
 
 
 class MimeTypes(str, Enum):
@@ -106,41 +119,76 @@ class Provider(StacBaseModel):
     https://github.com/radiantearth/stac-spec/blob/v1.0.0/collection-spec/collection-spec.md#provider-object
     """
 
-    name: str = Field(..., alias="name", min_length=1)
+    name: str = Field(..., min_length=1)
     description: Optional[str] = None
     roles: Optional[List[str]] = None
     url: Optional[str] = None
 
 
 class StacCommonMetadata(StacBaseModel):
     """
-    https://github.com/radiantearth/stac-spec/blob/v1.0.0/item-spec/common-metadata.md#date-and-time-range
+    https://github.com/radiantearth/stac-spec/blob/v1.0.0/item-spec/common-metadata.md
     """
 
-    title: Optional[str] = Field(None, alias="title")
-    description: Optional[str] = Field(None, alias="description")
-    start_datetime: Optional[datetime] = Field(None, alias="start_datetime")
-    end_datetime: Optional[datetime] = Field(None, alias="end_datetime")
-    created: Optional[datetime] = Field(None, alias="created")
-    updated: Optional[datetime] = Field(None, alias="updated")
-    platform: Optional[str] = Field(None, alias="platform")
-    instruments: Optional[List[str]] = Field(None, alias="instruments")
-    constellation: Optional[str] = Field(None, alias="constellation")
-    mission: Optional[str] = Field(None, alias="mission")
-    providers: Optional[List[Provider]] = Field(None, alias="providers")
-    gsd: Optional[float] = Field(None, alias="gsd", gt=0)
+    # Basic
+    title: Optional[str] = None
+    description: Optional[str] = None
+    # Date and Time
+    datetime: Optional[UtcDatetime] = None
+    created: Optional[UtcDatetime] = None
+    updated: Optional[UtcDatetime] = None
+    # Date and Time Range
+    start_datetime: Optional[UtcDatetime] = None
+    end_datetime: Optional[UtcDatetime] = None
+    # Provider
+    providers: Optional[List[Provider]] = None
+    # Instrument
+    platform: Optional[str] = None
+    instruments: Optional[List[str]] = None
+    constellation: Optional[str] = None
+    mission: Optional[str] = None
+    gsd: Optional[float] = Field(None, gt=0)
+
+    @model_validator(mode="after")
+    def validate_datetime_or_start_end(self) -> Self:
+        # When datetime is null, start_datetime and end_datetime must be specified
+        if not self.datetime and (not self.start_datetime or not self.end_datetime):
+            raise ValueError(
+                "start_datetime and end_datetime must be specified when datetime is null"
+            )
+
+        return self
+
+    @model_validator(mode="after")
+    def validate_start_end(self) -> Self:
+        # Using one of start_datetime or end_datetime requires the use of the other
+        if (self.start_datetime and not self.end_datetime) or (
+            not self.start_datetime and self.end_datetime
+        ):
+            raise ValueError(
+                "use of start_datetime or end_datetime requires the use of the other"
+            )
+        return self
 
 
 class Asset(StacCommonMetadata):
     """
     https://github.com/radiantearth/stac-spec/blob/v1.0.0/item-spec/item-spec.md#asset-object
     """
 
-    href: str = Field(..., alias="href", min_length=1)
+    href: str = Field(..., min_length=1)
     type: Optional[str] = None
     title: Optional[str] = None
     description: Optional[str] = None
     roles: Optional[List[str]] = None
+
     model_config = ConfigDict(
         populate_by_name=True, use_enum_values=True, extra="allow"
     )
+
+    @model_validator(mode="after")
+    def validate_datetime_or_start_end(self) -> Self:
+        # Overriding the parent method to avoid requiring datetime or start/end_datetime
+        # Additional fields MAY be added on the Asset object, but are not required.
+        # https://github.com/radiantearth/stac-spec/blob/v1.0.0/item-spec/item-spec.md#additional-fields-for-assets
+        return self

tests/api/extensions/test_fields.py

@@ -1,4 +1,4 @@
-from datetime import datetime
+from datetime import datetime, timezone
 
 from shapely.geometry import Polygon
 
@@ -15,7 +15,7 @@ def test_fields_filter_item():
     item = Item(
         id="test-fields-filter",
         geometry=Polygon.from_bounds(0, 0, 0, 0),
-        properties={"datetime": datetime.utcnow(), "foo": "foo", "bar": "bar"},
+        properties={"datetime": datetime.now(timezone.utc), "foo": "foo", "bar": "bar"},
         assets={},
         links=[
             {"href": "http://link", "rel": "self"},