Add CatalogSlice object and function to export data row IDs by slice

Author: attila-papai

docs/source/index.rst

@@ -173,14 +173,14 @@ DataRowMetadata
 
 .. automodule:: labelbox.schema.data_row_metadata
    :members:
-   :show-inheritance:   
+   :show-inheritance:
 
 AnnotationImport
 ----------------------------
 
 .. automodule:: labelbox.schema.annotation_import
    :members:
-   :show-inheritance:      
+   :show-inheritance:
 
 Batch
 ----------------------------
@@ -194,4 +194,19 @@ ResourceTag
 
 .. automodule:: labelbox.schema.resource_tag
    :members:
-   :show-inheritance:
+   :show-inheritance:
+
+Slice
+-----------------------------------------
+
+.. automodule:: labelbox.schema.slice
+   :members: Slice
+   :exclude-members: CatalogSlice
+   :show-inheritance:
+
+CatalogSlice
+-----------------------------------------
+.. automodule:: labelbox.schema.slice
+   :members: CatalogSlice
+   :exclude-members: Slice
+   :show-inheritance:

labelbox/__init__.py

@@ -27,3 +27,4 @@
 from labelbox.schema.resource_tag import ResourceTag
 from labelbox.schema.project_resource_tag import ProjectResourceTag
 from labelbox.schema.media_type import MediaType
+from labelbox.schema.slice import Slice, CatalogSlice

labelbox/client.py

@@ -33,6 +33,7 @@
 from labelbox.schema.user import User
 from labelbox.schema.project import Project
 from labelbox.schema.role import Role
+from labelbox.schema.slice import CatalogSlice
 
 from labelbox.schema.media_type import MediaType
 
@@ -963,7 +964,7 @@ def assign_global_keys_to_data_rows(
             timeout_seconds=60) -> Dict[str, Union[str, List[Any]]]:
         """
         Assigns global keys to data rows.
-        
+
         Args:
             A list of dicts containing data_row_id and global_key.
         Returns:
@@ -1211,3 +1212,27 @@ def _format_failed_rows(rows: List[str],
                     "Timed out waiting for get_data_rows_for_global_keys job to complete."
                 )
             time.sleep(sleep_time)
+
+    def get_catalog_slice(self, slice_id) -> CatalogSlice:
+        """
+        Fetches a Catalog Slice by ID.
+
+        Args:
+             slice_id (str): The ID of the Slice
+        Returns:
+            CatalogSlice
+        """
+        query_str = """
+            query getSavedQueryPyApi($id: ID!) {
+                getSavedQuery(id: $id) {
+                    id
+                    name
+                    description
+                    filter
+                    createdAt
+                    updatedAt
+                }
+            }
+        """
+        res = self.execute(query_str, {'id': slice_id})
+        return Entity.CatalogSlice(self, res['getSavedQuery'])

labelbox/orm/model.py

@@ -377,6 +377,7 @@ class Entity(metaclass=EntityMeta):
         ProjectRole: Type[labelbox.ProjectRole]
         Project: Type[labelbox.Project]
         Batch: Type[labelbox.Batch]
+        CatalogSlice: Type[labelbox.CatalogSlice]
 
     @classmethod
     def _attributes_of_type(cls, attr_type):

labelbox/schema/slice.py

@@ -0,0 +1,59 @@
+from labelbox.orm.db_object import DbObject
+from labelbox.orm.model import Field
+from labelbox.pagination import PaginatedCollection
+
+
+class Slice(DbObject):
+    """
+    A Slice is a saved set of filters (saved query).
+    This is an abstract class and should not be instantiated.
+
+    Attributes:
+        name (datetime)
+        description (datetime)
+        created_at (datetime)
+        updated_at (datetime)
+        filter (json)
+    """
+    name = Field.String("name")
+    description = Field.String("description")
+    created_at = Field.DateTime("created_at")
+    updated_at = Field.DateTime("updated_at")
+    filter = Field.Json("filter")
+
+
+class CatalogSlice(Slice):
+    """
+    Represents a Slice used for filtering data rows in Catalog.
+    """
+
+    def get_data_row_ids(self) -> PaginatedCollection:
+        """
+        Fetches all data row ids that match this Slice
+
+        Returns:
+            A PaginatedCollection of data row ids
+        """
+        query_str = """
+            query getDataRowIdsBySavedQueryPyApi($id: ID!, $from: String, $first: Int!) {
+                getDataRowIdsBySavedQuery(input: {
+                    savedQueryId: $id,
+                    after: $from
+                    first: $first
+                }) {
+                    totalCount
+                    nodes
+                    pageInfo {
+                        endCursor
+                        hasNextPage
+                    }
+                }
+            }
+        """
+        return PaginatedCollection(
+            client=self.client,
+            query=query_str,
+            params={'id': self.uid},
+            dereferencing=['getDataRowIdsBySavedQuery', 'nodes'],
+            obj_class=lambda _, data_row_id: data_row_id,
+            cursor_path=['getDataRowIdsBySavedQuery', 'pageInfo', 'endCursor'])