[MCC-1476025] - Add Publish Function

dataconnect/__init__.py

@@ -11,7 +11,15 @@
     ServerError,
     ValidationError,
 )
-from dataconnect.models import DatasetVersion, DryPublishResult, PaginatedResponse, Pagination, Study, StudyEnvironment
+from dataconnect.models import (
+    DatasetVersion,
+    DryPublishResult,
+    PaginatedResponse,
+    Pagination,
+    PublishResult,
+    Study,
+    StudyEnvironment,
+)
 
 __all__ = [
     # Client
@@ -23,6 +31,7 @@
     "DryPublishResult",
     "PaginatedResponse",
     "Pagination",
+    "PublishResult",
     # Exceptions — catch these in user application code
     "DataConnectError",
     "AuthenticationError",

dataconnect/client.py

@@ -12,7 +12,14 @@
 
 import pandas as pd
 
-from dataconnect.models import Dataset, DatasetVersion, DryPublishResult, PaginatedResponse, StudiesResult
+from dataconnect.models import (
+    Dataset,
+    DatasetVersion,
+    DryPublishResult,
+    PaginatedResponse,
+    PublishResult,
+    StudiesResult,
+)
 from dataconnect.service import DataConnectService, DefaultDataConnectService
 
 _DEFAULT_HOST = "enodia-gateway.platform.imedidata.com"
@@ -96,7 +103,26 @@ def dry_publish(
         data: pd.DataFrame,
         datetime_formats: dict[str, str] | None = None,
     ) -> DryPublishResult:
+        """Validate a dataset against the server without persisting any changes.
 
+        Delegates directly to :meth:`DataConnectService.dry_publish`.
+
+        Args:
+            project_token: Base64-encoded project token identifying the target
+                study, study environment, and project.
+            dataset_name: Name of the dataset to validate.
+            key_columns: Column names that form the unique key for the dataset.
+            source_datasets: UUIDs of the source datasets the published dataset
+                is derived from.
+            data: The dataset to validate as a ``pd.DataFrame``.
+            datetime_formats: Optional mapping of column name → datetime format
+                string (e.g. ``{"visit_date": "yyyy-MM-dd"}``).
+
+        Returns:
+            A :class:`DryPublishResult` containing the server's validation
+            outcome, including per-field validity flags, error messages, and
+            an optional ``invalid_records`` DataFrame.
+        """
         return self._service.dry_publish(
             project_token=project_token,
             dataset_name=dataset_name,
@@ -106,6 +132,43 @@ def dry_publish(
             datetime_formats=datetime_formats,
         )
 
+    def publish(
+        self,
+        project_token: str,
+        dataset_name: str,
+        key_columns: list[str],
+        source_datasets: list[UUID],
+        data: pd.DataFrame,
+        datetime_formats: dict[str, str] | None = None,
+    ) -> PublishResult:
+        """Publish a dataset to the server.
+
+        Delegates directly to :meth:`DataConnectService.publish`.
+
+        Args:
+            project_token: Base64-encoded project token identifying the target
+                study, study environment, and project.
+            dataset_name: Name of the dataset to publish.
+            key_columns: Column names that form the unique key for the dataset.
+            source_datasets: UUIDs of the source datasets the published dataset
+                is derived from.
+            data: The dataset to publish as a ``pd.DataFrame``.
+            datetime_formats: Optional mapping of column name → datetime format
+                string (e.g. ``{"visit_date": "yyyy-MM-dd"}``).
+
+        Returns:
+            A :class:`PublishResult` containing the server's publish outcome,
+            including dataset UUID, version, and record counts.
+        """
+        return self._service.publish(
+            project_token=project_token,
+            dataset_name=dataset_name,
+            key_columns=key_columns,
+            source_datasets=source_datasets,
+            data=data,
+            datetime_formats=datetime_formats,
+        )
+
     # Lifecycle
 
     def close(self) -> None:

dataconnect/models.py

@@ -82,3 +82,18 @@ class DryPublishResult:
     duplicate_record_count: int | None = None
     invalid_record_count: int | None = None
     invalid_records: pd.DataFrame | None = None
+
+
+@dataclass
+class PublishResult:
+    """Result of a publish operation, including status and details."""
+
+    status: bool
+    dataset_name: str | None = None
+    dataset_uuid: str | None = None
+    dataset_version: int | None = None
+    dataset_batch_number: int | None = None
+    valid_record_count: int | None = None
+    duplicate_record_count: int | None = None
+    invalid_record_count: int | None = None
+    invalid_records: pd.DataFrame | None = None

dataconnect/service/base.py

@@ -7,7 +7,14 @@
 
 import pandas as pd
 
-from dataconnect.models import Dataset, DatasetVersion, DryPublishResult, PaginatedResponse, StudiesResult
+from dataconnect.models import (
+    Dataset,
+    DatasetVersion,
+    DryPublishResult,
+    PaginatedResponse,
+    PublishResult,
+    StudiesResult,
+)
 
 
 class DataConnectService(ABC):
@@ -44,7 +51,22 @@ def dry_publish(
         source_datasets: list[UUID],
         data: pd.DataFrame,
         datetime_formats: dict[str, str] | None = None,
-    ) -> DryPublishResult: ...
+    ) -> DryPublishResult:
+        """Validate a dataset against the server without persisting any changes."""
+        ...
+
+    @abstractmethod
+    def publish(
+        self,
+        project_token: str,
+        dataset_name: str,
+        key_columns: list[str],
+        source_datasets: list[UUID],
+        data: pd.DataFrame,
+        datetime_formats: dict[str, str] | None = None,
+    ) -> PublishResult:
+        """Publish a dataset to the server and return the result."""
+        ...
 
     @abstractmethod
     def close(self) -> None: ...

dataconnect/service/default.py

@@ -7,11 +7,20 @@
 
 import pandas as pd
 
-from dataconnect.models import Dataset, DatasetVersion, DryPublishResult, PaginatedResponse, Pagination, StudiesResult
+from dataconnect.models import (
+    Dataset,
+    DatasetVersion,
+    DryPublishResult,
+    PaginatedResponse,
+    Pagination,
+    PublishResult,
+    StudiesResult,
+)
 from dataconnect.service.base import DataConnectService
 from dataconnect.service.error_handler import translate_error
 from dataconnect.service.mappers import (
     dry_publish_response_to_domain,
+    publish_response_to_domain,
     resource_to_dataset,
     resource_to_dataset_version,
     resource_to_fetched_data,
@@ -205,6 +214,68 @@ def dry_publish(
         except TransportError as ex:
             raise translate_error(ex) from ex
 
+    def publish(
+        self,
+        project_token: str,
+        dataset_name: str,
+        key_columns: list[str],
+        source_datasets: list[UUID],
+        data: pd.DataFrame,
+        datetime_formats: dict[str, str] | None = None,
+    ) -> PublishResult:
+        """Publish a dataset to the server and return the result.
+
+        Encodes the publish configuration as a JSON ``input_config`` string,
+        sets ``is_dry_publish: False`` so the server treats the call as a
+        live publish run, then delegates to the transport layer.
+
+        The server response is mapped to a :class:`PublishResult` via
+        :func:`publish_response_to_domain`.
+
+        Args:
+            project_token: Base64-encoded project token identifying the target
+                study, study environment, and project.
+            dataset_name: Name of the dataset to publish.
+            key_columns: Column names that form the unique key for the dataset.
+            source_datasets: UUIDs of the source datasets the published dataset
+                is derived from.
+            data: The dataset to publish as a ``pd.DataFrame``.
+            datetime_formats: Optional mapping of column name → datetime format
+                string (e.g. ``{"visit_date": "yyyy-MM-dd"}``).  Defaults to
+                an empty dict when omitted.
+
+        Returns:
+            A :class:`PublishResult` containing the server's publish outcome,
+            including dataset UUID, version, and record counts.
+
+        Raises:
+            DataConnectError: Any :class:`TransportError` from the transport
+                layer is translated by :func:`translate_error` into the public
+                API's :class:`DataConnectError` hierarchy before propagating.
+        """
+        request = PublishRequest(
+            input_config=json.dumps(
+                {
+                    "is_dry_publish": False,
+                    "project_token": project_token,
+                    "dataset_name": dataset_name,
+                    "dataset_description": dataset_name,
+                    "key_columns": key_columns,
+                    "source_datasets": [str(uuid) for uuid in source_datasets],
+                    "datetime_formats": datetime_formats or {},
+                }
+            ),
+            data=data,
+        )
+
+        try:
+            publish_result = self._transport.publish_dataset(request)
+
+            return publish_response_to_domain(publish_result)
+
+        except TransportError as ex:
+            raise translate_error(ex) from ex
+
     def close(self) -> None:
         """Close the underlying transport connection."""
 

dataconnect/service/mappers.py

@@ -15,8 +15,8 @@
 import pyarrow as pa
 
 from dataconnect.exceptions import NotFoundError
-from dataconnect.models import Dataset, DatasetVersion, DryPublishResult, Study, StudyEnvironment
-from dataconnect.transport.models import DataTable, DryPublishResponse, ResourceInfo
+from dataconnect.models import Dataset, DatasetVersion, DryPublishResult, PublishResult, Study, StudyEnvironment
+from dataconnect.transport.models import DataTable, DryPublishResponse, PublishResponse, ResourceInfo
 
 
 def resource_to_study(resource: ResourceInfo) -> Study:
@@ -124,3 +124,34 @@ def dry_publish_response_to_domain(result: DryPublishResponse | None) -> DryPubl
         invalid_record_count=result.invalid_record_count,
         invalid_records=result.invalid_records,
     )
+
+
+def publish_response_to_domain(result: PublishResponse | None) -> PublishResult:
+    """Map a transport-layer ``PublishResponse`` to a ``PublishResult`` domain object.
+
+    ``PublishResponse`` carries flat, typed fields returned by the server after a
+    publish call. The mapping is direct for all shared fields.
+
+    Args:
+        result: The transport-layer result returned by
+            :meth:`Transport.publish_dataset`. Pass ``None`` to obtain a
+            default :class:`PublishResult` with ``status=False`` and all
+            other fields left at their default values.
+
+    Returns:
+        A :class:`PublishResult` suitable for returning to the caller.
+    """
+    if result is None:
+        return PublishResult(status=False)
+
+    return PublishResult(
+        status=result.status,
+        dataset_name=result.dataset_name,
+        dataset_uuid=result.dataset_uuid,
+        dataset_version=result.dataset_version,
+        dataset_batch_number=result.dataset_batch_number,
+        valid_record_count=result.valid_record_count,
+        duplicate_record_count=result.duplicate_record_count,
+        invalid_record_count=result.invalid_record_count,
+        invalid_records=result.invalid_records,
+    )

dataconnect/transport/__init__.py

@@ -12,6 +12,7 @@
     DataTable,
     DryPublishResponse,
     PublishRequest,
+    PublishResponse,
     ResourceInfo,
     ResourceQuery,
 )
@@ -24,5 +25,6 @@
     "DatasetTicket",
     "DataTable",
     "PublishRequest",
+    "PublishResponse",
     "DryPublishResponse",
 ]