ENH: Make ObjectDataProviders return Pydantic models

src/fmu/dataio/_definitions.py

@@ -69,15 +69,6 @@ class ExportFolder(str, Enum):
     tables = "tables"
 
 
-class Layout(str, Enum):
-    regular = "regular"
-    unset = "unset"
-    cornerpoint = "cornerpoint"
-    table = "table"
-    dictionary = "dictionary"
-    faultroom_triangulated = "faultroom_triangulated"
-
-
 STANDARD_TABLE_INDEX_COLUMNS: Final[dict[str, list[str]]] = {
     "volumes": ["ZONE", "REGION", "FACIES", "LICENCE"],
     "rft": ["measured_depth", "well", "time"],

src/fmu/dataio/_metadata.py

@@ -17,7 +17,6 @@
 from . import types
 from ._definitions import SCHEMA, SOURCE, VERSION, FmuContext
 from ._logging import null_logger
-from ._utils import drop_nones
 from .datastructure._internal import internal
 from .datastructure.meta import meta
 from .exceptions import InvalidMetadataError
@@ -109,8 +108,7 @@ def generate_export_metadata(
     dataio: ExportData,
     fmudata: FmuProvider | None = None,
     compute_md5: bool = True,
-    skip_null: bool = True,
-) -> dict:  # TODO! -> skip_null?
+) -> internal.DataClassMeta:
     """
     Main function to generate the full metadata
 
@@ -141,7 +139,7 @@ def generate_export_metadata(
     objdata = objectdata_provider_factory(obj, dataio)
     masterdata = dataio.config.get("masterdata")
 
-    metadata = internal.DataClassMeta(
+    return internal.DataClassMeta(
         schema_=TypeAdapter(AnyHttpUrl).validate_strings(SCHEMA),  # type: ignore[call-arg]
         version=VERSION,
         source=SOURCE,
@@ -154,6 +152,4 @@ def generate_export_metadata(
         tracklog=generate_meta_tracklog(),
         display=_get_meta_display(dataio, objdata),
         preprocessed=dataio.fmu_context == FmuContext.PREPROCESSED,
-    ).model_dump(mode="json", exclude_none=True, by_alias=True)
-
-    return metadata if not skip_null else drop_nones(metadata)
+    )

src/fmu/dataio/dataio.py

@@ -22,6 +22,7 @@
 from ._metadata import generate_export_metadata
 from ._utils import (
     detect_inside_rms,  # dataio_examples,
+    drop_nones,
     export_file,
     export_metadata_file,
     prettyprint_dict,
@@ -769,8 +770,10 @@ def generate_metadata(
 
         self._update_fmt_flag()
         fmudata = self._get_fmu_provider() if self._fmurun else None
-        self._metadata = generate_export_metadata(
-            obj, self, fmudata, compute_md5=compute_md5
+        self._metadata = drop_nones(
+            generate_export_metadata(
+                obj, self, fmudata, compute_md5=compute_md5
+            ).model_dump(mode="json", exclude_none=True, by_alias=True)
         )
 
         logger.info("The metadata are now ready!")

src/fmu/dataio/datastructure/meta/content.py

@@ -197,9 +197,9 @@ class Content(BaseModel):
     is_prediction: bool = Field(
         title="Is prediction flag",
     )
-    layout: Optional[str] = Field(
+    layout: Optional[enums.Layout] = Field(
         default=None,
-        examples=["regular"],
+        examples=["regular", "cornerpoint"],
     )
     name: str = Field(
         description=(

src/fmu/dataio/datastructure/meta/enums.py

@@ -4,6 +4,17 @@
 from typing import Type
 
 
+class AccessLevel(str, Enum):
+    asset = "asset"
+    internal = "internal"
+    restricted = "restricted"
+
+
+class AxisOrientation(IntEnum):
+    normal = 1
+    flipped = -1
+
+
 class ContentEnum(str, Enum):
     depth = "depth"
     facies_thickness = "facies_thickness"
@@ -52,12 +63,10 @@ class FMUClassEnum(str, Enum):
     dictionary = "dictionary"
 
 
-class AccessLevel(str, Enum):
-    asset = "asset"
-    internal = "internal"
-    restricted = "restricted"
-
-
-class AxisOrientation(IntEnum):
-    normal = 1
-    flipped = -1
+class Layout(str, Enum):
+    regular = "regular"
+    unset = "unset"
+    cornerpoint = "cornerpoint"
+    table = "table"
+    dictionary = "dictionary"
+    faultroom_triangulated = "faultroom_triangulated"

src/fmu/dataio/providers/objectdata/_base.py

@@ -4,7 +4,7 @@
 from copy import deepcopy
 from dataclasses import dataclass, field
 from datetime import datetime
-from typing import TYPE_CHECKING, Final
+from typing import TYPE_CHECKING, Any, Final
 from warnings import warn
 
 from fmu.dataio._definitions import ConfigurationError, ValidFormats
@@ -20,10 +20,9 @@
 from fmu.dataio.providers._base import Provider
 
 if TYPE_CHECKING:
-    from fmu.dataio._definitions import Layout
     from fmu.dataio.dataio import ExportData
     from fmu.dataio.datastructure.meta.content import BoundingBox2D, BoundingBox3D
-    from fmu.dataio.datastructure.meta.enums import FMUClassEnum
+    from fmu.dataio.datastructure.meta.enums import FMUClassEnum, Layout
     from fmu.dataio.datastructure.meta.specification import AnySpecification
     from fmu.dataio.types import Inferrable
 
@@ -62,7 +61,7 @@ class ObjectDataProvider(Provider):
     # result properties; the most important is metadata which IS the 'data' part in
     # the resulting metadata. But other variables needed later are also given
     # as instance properties in addition (for simplicity in other classes/functions)
-    _metadata: dict = field(default_factory=dict)
+    _metadata: AnyContent | UnsetAnyContent | None = field(default=None)
     name: str = field(default="")
     time0: datetime | None = field(default=None)
     time1: datetime | None = field(default=None)
@@ -80,53 +79,44 @@ def __post_init__(self) -> None:
 
         content_model = self._get_validated_content(self.dataio.content)
         named_stratigraphy = self._get_named_stratigraphy()
-
         self.name = named_stratigraphy.name
-        self._metadata["name"] = self.name
-        self._metadata["stratigraphic"] = named_stratigraphy.stratigraphic
-        self._metadata["offset"] = named_stratigraphy.offset
-        self._metadata["alias"] = named_stratigraphy.alias
-        self._metadata["top"] = named_stratigraphy.top
-        self._metadata["base"] = named_stratigraphy.base
-
-        self._metadata["content"] = (usecontent := content_model.content)
+
+        metadata: dict[str, Any] = {}
+        metadata["name"] = self.name
+        metadata["stratigraphic"] = named_stratigraphy.stratigraphic
+        metadata["offset"] = named_stratigraphy.offset
+        metadata["alias"] = named_stratigraphy.alias
+        metadata["top"] = named_stratigraphy.top
+        metadata["base"] = named_stratigraphy.base
+
+        metadata["content"] = (usecontent := content_model.content)
         if content_model.content_incl_specific:
-            self._metadata[usecontent] = getattr(
+            metadata[usecontent] = getattr(
                 content_model.content_incl_specific, usecontent, None
             )
 
-        self._metadata["tagname"] = self.dataio.tagname
-        self._metadata["format"] = self.fmt
-        self._metadata["layout"] = self.layout
-        self._metadata["unit"] = self.dataio.unit or ""
-        self._metadata["vertical_domain"] = list(self.dataio.vertical_domain.keys())[0]
-        self._metadata["depth_reference"] = list(self.dataio.vertical_domain.values())[
-            0
-        ]
-
-        self._metadata["spec"] = (
-            spec.model_dump(mode="json", exclude_none=True)
-            if (spec := self.get_spec())
-            else None
-        )
-        self._metadata["bbox"] = (
-            bbox.model_dump(mode="json", exclude_none=True)
-            if (bbox := self.get_bbox())
-            else None
+        metadata["tagname"] = self.dataio.tagname
+        metadata["format"] = self.fmt
+        metadata["layout"] = self.layout
+        metadata["unit"] = self.dataio.unit or ""
+        metadata["vertical_domain"] = list(self.dataio.vertical_domain.keys())[0]
+        metadata["depth_reference"] = list(self.dataio.vertical_domain.values())[0]
+
+        metadata["spec"] = self.get_spec()
+        metadata["bbox"] = self.get_bbox()
+        metadata["time"] = self._get_timedata()
+        metadata["table_index"] = self.table_index
+        metadata["undef_is_zero"] = self.dataio.undef_is_zero
+
+        metadata["is_prediction"] = self.dataio.is_prediction
+        metadata["is_observation"] = self.dataio.is_observation
+        metadata["description"] = generate_description(self.dataio.description)
+
+        self._metadata = (
+            UnsetAnyContent.model_validate(metadata)
+            if metadata["content"] == "unset"
+            else AnyContent.model_validate(metadata)
         )
-        self._metadata["time"] = (
-            timedata.model_dump(mode="json", exclude_none=True)
-            if (timedata := self._get_timedata())
-            else None
-        )
-
-        self._metadata["table_index"] = self.table_index
-        self._metadata["undef_is_zero"] = self.dataio.undef_is_zero
-
-        # timedata:
-        self._metadata["is_prediction"] = self.dataio.is_prediction
-        self._metadata["is_observation"] = self.dataio.is_observation
-        self._metadata["description"] = generate_description(self.dataio.description)
         logger.info("Derive all metadata for data object... DONE")
 
     @property
@@ -168,11 +158,8 @@ def get_spec(self) -> AnySpecification | None:
         raise NotImplementedError
 
     def get_metadata(self) -> AnyContent | UnsetAnyContent:
-        return (
-            UnsetAnyContent.model_validate(self._metadata)
-            if self._metadata["content"] == "unset"
-            else AnyContent.model_validate(self._metadata)
-        )
+        assert self._metadata is not None
+        return self._metadata
 
     def _get_validated_content(self, content: str | dict | None) -> AllowedContent:
         """Check content and return a validated model."""

src/fmu/dataio/providers/objectdata/_faultroom.py

@@ -3,10 +3,10 @@
 from dataclasses import dataclass
 from typing import Final
 
-from fmu.dataio._definitions import ExportFolder, Layout, ValidFormats
+from fmu.dataio._definitions import ExportFolder, ValidFormats
 from fmu.dataio._logging import null_logger
 from fmu.dataio.datastructure.meta.content import BoundingBox3D
-from fmu.dataio.datastructure.meta.enums import FMUClassEnum
+from fmu.dataio.datastructure.meta.enums import FMUClassEnum, Layout
 from fmu.dataio.datastructure.meta.specification import FaultRoomSurfaceSpecification
 from fmu.dataio.readers import FaultRoomSurface
 

src/fmu/dataio/providers/objectdata/_provider.py

@@ -92,9 +92,9 @@
 import pandas as pd
 import xtgeo
 
-from fmu.dataio._definitions import ExportFolder, Layout, ValidFormats
+from fmu.dataio._definitions import ExportFolder, ValidFormats
 from fmu.dataio._logging import null_logger
-from fmu.dataio.datastructure.meta.enums import FMUClassEnum
+from fmu.dataio.datastructure.meta.enums import FMUClassEnum, Layout
 from fmu.dataio.readers import FaultRoomSurface
 
 from ._base import (

src/fmu/dataio/providers/objectdata/_tables.py

@@ -8,11 +8,10 @@
 from fmu.dataio._definitions import (
     STANDARD_TABLE_INDEX_COLUMNS,
     ExportFolder,
-    Layout,
     ValidFormats,
 )
 from fmu.dataio._logging import null_logger
-from fmu.dataio.datastructure.meta.enums import FMUClassEnum
+from fmu.dataio.datastructure.meta.enums import FMUClassEnum, Layout
 from fmu.dataio.datastructure.meta.specification import TableSpecification
 
 from ._base import (

src/fmu/dataio/providers/objectdata/_xtgeo.py

@@ -7,11 +7,11 @@
 import pandas as pd
 import xtgeo
 
-from fmu.dataio._definitions import ExportFolder, Layout, ValidFormats
+from fmu.dataio._definitions import ExportFolder, ValidFormats
 from fmu.dataio._logging import null_logger
 from fmu.dataio._utils import npfloat_to_float
 from fmu.dataio.datastructure.meta.content import BoundingBox2D, BoundingBox3D
-from fmu.dataio.datastructure.meta.enums import FMUClassEnum
+from fmu.dataio.datastructure.meta.enums import FMUClassEnum, Layout
 from fmu.dataio.datastructure.meta.specification import (
     CPGridPropertySpecification,
     CPGridSpecification,
@@ -21,9 +21,7 @@
     SurfaceSpecification,
 )
 
-from ._base import (
-    ObjectDataProvider,
-)
+from ._base import ObjectDataProvider
 
 if TYPE_CHECKING:
     import pandas as pd