datajunction-server 0.0.1a37__tar.gz → 0.0.1a37.dev0__tar.gz

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: datajunction-server
-Version: 0.0.1a37
+Version: 0.0.1a37.dev0
 Summary: DataJunction server library for running to a DataJunction server
 Project-URL: repository, https://github.com/DataJunction/dj
 Author-email: DataJunction Authors <yian.shang@gmail.com>

datajunction_server-0.0.1a37.dev0/curl-format.txt

@@ -0,0 +1,9 @@
+     time_namelookup:  %{time_namelookup}s\n
+        time_connect:  %{time_connect}s\n
+     time_appconnect:  %{time_appconnect}s\n
+    time_pretransfer:  %{time_pretransfer}s\n
+       time_redirect:  %{time_redirect}s\n
+  time_starttransfer:  %{time_starttransfer}s\n
+                     ----------\n
+          time_total:  %{time_total}s\n
+

datajunction_server-0.0.1a37.dev0/datajunction_server/__about__.py

@@ -0,0 +1,4 @@
+"""
+Version for Hatch
+"""
+__version__ = "0.0.1a37.dev0"

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/api/health.py

@@ -2,13 +2,13 @@
 Application healthchecks.
 """
 
-import enum
 from typing import List
 
 from fastapi import APIRouter, Depends
 from sqlalchemy import select
 from sqlmodel import Session, SQLModel
 
+from datajunction_server.enum import StrEnum
 from datajunction_server.utils import get_session, get_settings
 
 settings = get_settings()
@@ -16,7 +16,7 @@ settings = get_settings()
 router = APIRouter(tags=["health"])
 
 
-class HealthcheckStatus(str, enum.Enum):
+class HealthcheckStatus(StrEnum):
     """
     Possible health statuses.
     """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/api/helpers.py

@@ -953,9 +953,11 @@ async def query_event_stream(  # pylint: disable=too-many-arguments
             break
 
         # Check the current state of the query
-        query_next = query_service_client.get_query(query_id=query_id)  # type: ignore
-        if query_next.state in END_JOB_STATES:
-            _logger.info(
+        query_next = query_service_client.get_query(  # type: ignore # pragma: no cover
+            query_id=query_id,
+        )
+        if query_next.state in END_JOB_STATES:  # pragma: no cover
+            _logger.info(  # pragma: no cover
                 "query end state detected (%s), sending final event to the client",
                 query_next.state,
             )

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/api/main.py

@@ -79,6 +79,7 @@ app = FastAPI(
     },
     dependencies=dependencies,
 )
+
 app.add_middleware(
     CORSMiddleware,
     allow_origins=settings.cors_origin_whitelist,

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/api/metrics.py

@@ -6,7 +6,6 @@ from http import HTTPStatus
 from typing import List, Optional
 
 from fastapi import Depends, HTTPException, Query
-from sqlalchemy.exc import NoResultFound
 from sqlalchemy.sql.operators import is_
 from sqlmodel import Session, select
 
@@ -103,31 +102,28 @@ async def get_common_dimensions(
     """
     Return common dimensions for a set of metrics.
     """
-    metric_nodes = []
     errors = []
-    for node_name in metric:
-        statement = (
-            select(Node)
-            .where(Node.name == node_name)
-            .where(is_(Node.deactivated_at, None))
-        )
-        try:
-            node = session.exec(statement).one()
-            if node.type != NodeType.METRIC:
-                errors.append(
-                    DJError(
-                        message=f"Not a metric node: {node_name}",
-                        code=ErrorCode.NODE_TYPE_ERROR,
-                    ),
-                )
-            metric_nodes.append(node)
-        except NoResultFound:
+    statement = (
+        select(Node)
+        .where(Node.name.in_(metric))  # type: ignore  # pylint: disable=no-member
+        .where(is_(Node.deactivated_at, None))
+    )
+    metric_nodes = session.exec(statement).all()
+    for node in metric_nodes:
+        if node.type != NodeType.METRIC:
             errors.append(
                 DJError(
-                    message=f"Metric node not found: {node_name}",
-                    code=ErrorCode.UNKNOWN_NODE,
+                    message=f"Not a metric node: {node.name}",
+                    code=ErrorCode.NODE_TYPE_ERROR,
                 ),
             )
+    if not metric_nodes:
+        errors.append(
+            DJError(
+                message=f"Metric nodes not found: {','.join(metric)}",
+                code=ErrorCode.UNKNOWN_NODE,
+            ),
+        )
 
     if errors:
         raise DJException(errors=errors)

datajunction_server-0.0.1a37.dev0/datajunction_server/enum.py

@@ -0,0 +1,21 @@
+"""
+Backwards-compatible StrEnum for both Python >= and < 3.11
+"""
+import enum
+import sys
+
+if sys.version_info >= (3, 11):
+    from enum import (  # noqa  # pylint: disable=unused-import  # pragma: no cover
+        IntEnum,
+        StrEnum,
+    )
+else:
+
+    class StrEnum(str, enum.Enum):  # pragma: no cover
+        """Backwards compatible StrEnum for Python < 3.11"""  # pragma: no cover
+
+        def __str__(self):
+            return str(self.value)
+
+    class IntEnum(int, enum.Enum):  # pragma: no cover
+        """Backwards compatible IntEnum for Python < 3.11"""  # pragma: no cover

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/errors.py

@@ -2,13 +2,14 @@
 Errors and warnings.
 """
 
-from enum import Enum
 from typing import Any, Dict, List, Literal, Optional, TypedDict
 
 from sqlmodel import SQLModel
 
+from datajunction_server.enum import IntEnum
 
-class ErrorCode(int, Enum):
+
+class ErrorCode(IntEnum):
     """
     Error codes.
     """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/models/attribute.py

@@ -1,13 +1,13 @@
 """
 Models for attributes.
 """
-import enum
 from typing import TYPE_CHECKING, List, Optional
 
 from sqlalchemy import JSON, String, UniqueConstraint
 from sqlalchemy.sql.schema import Column as SqlaColumn
 from sqlmodel import Field, Relationship
 
+from datajunction_server.enum import StrEnum
 from datajunction_server.models.base import BaseSQLModel
 from datajunction_server.models.node_type import NodeType
 
@@ -36,7 +36,7 @@ class MutableAttributeTypeFields(AttributeTypeIdentifier):
     allowed_node_types: List[NodeType] = Field(sa_column=SqlaColumn(JSON))
 
 
-class UniquenessScope(str, enum.Enum):
+class UniquenessScope(StrEnum):
     """
     The scope at which this attribute needs to be unique.
     """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/models/column.py

@@ -1,7 +1,6 @@
 """
 Models for columns.
 """
-import enum
 from typing import TYPE_CHECKING, List, Optional, Tuple, TypedDict
 
 from pydantic import root_validator
@@ -10,6 +9,7 @@ from sqlalchemy.sql.schema import Column as SqlaColumn
 from sqlalchemy.types import String, Text
 from sqlmodel import Field, Relationship
 
+from datajunction_server.enum import StrEnum
 from datajunction_server.models.base import (
     BaseSQLModel,
     NodeColumns,
@@ -191,7 +191,7 @@ class ColumnAttributeInput(BaseSQLModel):
     column_name: str
 
 
-class SemanticType(str, enum.Enum):
+class SemanticType(StrEnum):
     """
     Semantic type of a column
     """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/models/engine.py

@@ -1,17 +1,17 @@
 """
 Models for columns.
 """
-import enum
 from typing import Optional
 
 from sqlalchemy.sql.schema import Column as SqlaColumn
 from sqlalchemy.types import Enum
 from sqlmodel import Field, SQLModel
 
+from datajunction_server.enum import StrEnum
 from datajunction_server.models.base import BaseSQLModel
 
 
-class Dialect(str, enum.Enum):
+class Dialect(StrEnum):
     """
     SQL dialect
     """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/models/history.py

@@ -2,7 +2,6 @@
 Model for history.
 """
 from datetime import datetime, timezone
-from enum import Enum
 from functools import partial
 from typing import Any, Dict, Optional
 
@@ -10,12 +9,13 @@ from sqlalchemy import DateTime
 from sqlalchemy.sql.schema import Column as SqlaColumn
 from sqlmodel import JSON, Field, SQLModel
 
+from datajunction_server.enum import StrEnum
 from datajunction_server.models.node import NodeRevision, NodeStatus
 from datajunction_server.models.user import User
 from datajunction_server.typing import UTCDatetime
 
 
-class ActivityType(str, Enum):
+class ActivityType(StrEnum):
     """
     An activity type
     """
@@ -30,7 +30,7 @@ class ActivityType(str, Enum):
     STATUS_CHANGE = "status_change"
 
 
-class EntityType(str, Enum):
+class EntityType(StrEnum):
     """
     An entity type for which activity can occur
     """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/models/materialization.py

@@ -9,6 +9,7 @@ from sqlalchemy import DateTime, String
 from sqlalchemy.types import Enum
 from sqlmodel import Field, Relationship, SQLModel, UniqueConstraint
 
+from datajunction_server.enum import StrEnum
 from datajunction_server.errors import DJInvalidInputException
 from datajunction_server.models.base import BaseSQLModel
 from datajunction_server.models.node_type import NodeType
@@ -44,7 +45,7 @@ DRUID_AGG_MAPPING = {
 }
 
 
-class MaterializationStrategy(str, enum.Enum):
+class MaterializationStrategy(StrEnum):
     """
     Materialization strategies
     """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/models/measure.py

@@ -1,7 +1,6 @@
 """
 Models for measures.
 """
-import enum
 from typing import TYPE_CHECKING, List, Optional
 
 from pydantic.class_validators import root_validator
@@ -9,13 +8,14 @@ from sqlalchemy.sql.schema import Column as SqlaColumn
 from sqlalchemy.types import Enum, String
 from sqlmodel import Field, Relationship
 
+from datajunction_server.enum import StrEnum
 from datajunction_server.models.base import BaseSQLModel, generate_display_name
 
 if TYPE_CHECKING:
     from datajunction_server.models import Column
 
 
-class AggregationRule(str, enum.Enum):
+class AggregationRule(StrEnum):
     """
     Type of allowed aggregation for a given measure.
     """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/models/node.py

@@ -20,6 +20,7 @@ from sqlalchemy.types import Enum
 from sqlmodel import Field, Relationship, SQLModel
 from typing_extensions import TypedDict
 
+from datajunction_server.enum import StrEnum
 from datajunction_server.errors import DJError, DJInvalidInputException
 from datajunction_server.models.base import (
     BaseSQLModel,
@@ -124,7 +125,7 @@ class BoundDimensionsRelationship(BaseSQLModel, table=True):  # type: ignore
     )
 
 
-class NodeMode(str, enum.Enum):
+class NodeMode(StrEnum):
     """
     Node mode.
 
@@ -138,7 +139,7 @@ class NodeMode(str, enum.Enum):
     DRAFT = "draft"
 
 
-class NodeStatus(str, enum.Enum):
+class NodeStatus(StrEnum):
     """
     Node status.
 
@@ -487,7 +488,7 @@ class AvailabilityState(AvailabilityStateBase, table=True):  # type: ignore
         return True
 
 
-class MetricDirection(str, enum.Enum):
+class MetricDirection(StrEnum):
     """
     The direction of the metric that's considered good, i.e., higher is better
     """
@@ -1111,17 +1112,9 @@ class DimensionAttributeOutput(SQLModel):
     node_name: Optional[str]
     node_display_name: Optional[str]
     is_primary_key: bool
-    type: ColumnType
+    type: str
     path: List[str]
 
-    @root_validator
-    def type_string(cls, values):  # pylint: disable=no-self-argument
-        """
-        Extracts the type as a string
-        """
-        values["type"] = str(values.get("type"))
-        return values
-
 
 class ColumnOutput(BaseSQLModel):
     """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/models/node_type.py

@@ -1,8 +1,8 @@
 """Node type"""
-import enum
+from datajunction_server.enum import StrEnum
 
 
-class NodeType(str, enum.Enum):
+class NodeType(StrEnum):
     """
     Node type.
 

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/models/partition.py

@@ -1,5 +1,4 @@
 """Partition-related models."""
-import enum
 from typing import TYPE_CHECKING, List, Optional
 
 from pydantic.class_validators import validator
@@ -8,6 +7,7 @@ from sqlalchemy import JSON
 from sqlalchemy import Column as SqlaColumn
 from sqlmodel import Field, Relationship, SQLModel
 
+from datajunction_server.enum import StrEnum
 from datajunction_server.models.base import BaseSQLModel
 from datajunction_server.models.column import Column
 from datajunction_server.sql.parsing.types import TimestampType
@@ -16,7 +16,7 @@ if TYPE_CHECKING:
     from datajunction_server.models.materialization import Materialization
 
 
-class PartitionType(str, enum.Enum):
+class PartitionType(StrEnum):
     """
     Partition type.
 
@@ -27,7 +27,7 @@ class PartitionType(str, enum.Enum):
     CATEGORICAL = "categorical"
 
 
-class Granularity(str, enum.Enum):
+class Granularity(StrEnum):
     """
     Time dimension granularity.
     """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/models/query.py

@@ -3,13 +3,13 @@ Models for queries.
 """
 
 from datetime import datetime
-from enum import Enum
 from typing import Any, List, Optional
 
 import msgpack
 from pydantic import AnyHttpUrl, validator
 from sqlmodel import Field, SQLModel
 
+from datajunction_server.enum import IntEnum
 from datajunction_server.models.base import BaseSQLModel
 from datajunction_server.typing import QueryState, Row
 
@@ -135,7 +135,7 @@ class QueryWithResults(BaseSQLModel):
         return datetime.fromisoformat(value) if isinstance(value, str) else value
 
 
-class QueryExtType(int, Enum):
+class QueryExtType(IntEnum):
     """
     Custom ext type for msgpack.
     """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/sql/dag.py

@@ -47,7 +47,9 @@ def get_dimensions(
             continue
         processed.add(current_node)
 
-        for column in current_node.current.columns:
+        current_revision = current_node.current
+        current_primary_key = current_node.current.primary_key()
+        for column in current_revision.columns:
             # Include the dimension if it's a column belonging to a dimension node
             # or if it's tagged with the dimension column attribute (but not
             # additionally linked to a dimension)
@@ -69,11 +71,11 @@ def get_dimensions(
                 dimensions.append(
                     DimensionAttributeOutput(
                         name=f"{current_node.name}.{column.name}",
-                        node_name=current_node.current.name,
-                        node_display_name=current_node.current.display_name,
+                        node_name=current_revision.name,
+                        node_display_name=current_revision.display_name,
                         is_primary_key=column.name
-                        in {pk.name for pk in current_node.current.primary_key()},
-                        type=column.type,
+                        in {pk.name for pk in current_primary_key},
+                        type=str(column.type),
                         path=join_path_str,
                     ),
                 )
@@ -90,7 +92,7 @@ def check_convergence(path1: List[str], path2: List[str]) -> bool:
     final element, the dimension attribute.
     """
     if path1 == path2:
-        return True
+        return True  # pragma: no cover
     len1 = len(path1)
     len2 = len(path2)
     min_len = min(len1, len2)
@@ -131,8 +133,14 @@ def get_shared_dimensions(
     """
     Return a list of dimensions that are common between the nodes.
     """
-    common = group_dimensions_by_name(metric_nodes[0])
-    for node in set(metric_nodes[1:]):
+    common_parents = set()
+    for metric_node in metric_nodes:
+        immediate_parent = metric_node.current.parents[0]
+        common_parents.add(immediate_parent)
+
+    parents = list(common_parents)
+    common = group_dimensions_by_name(parents[0])
+    for node in parents[1:]:
         node_dimensions = group_dimensions_by_name(node)
 
         # Merge each set of dimensions based on the name and path

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/sql/parsing/types.py

@@ -13,12 +13,13 @@ field_type=IntegerType(), is_optional=True, doc='an optional field'))
 """
 
 import re
-from enum import Enum
 from typing import TYPE_CHECKING, Any, ClassVar, Dict, Generator, Optional, Tuple, cast
 
 from pydantic import BaseModel, Extra
 from pydantic.class_validators import AnyCallable
 
+from datajunction_server.enum import StrEnum
+
 if TYPE_CHECKING:
     from datajunction_server.sql.parsing import ast
 
@@ -657,7 +658,7 @@ class DateTimeBase(PrimitiveType, Singleton):
     """
 
     # pylint: disable=invalid-name
-    class Unit(str, Enum):
+    class Unit(StrEnum):
         """
         Units used for date and time functions and intervals
         """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/typing.py

@@ -7,13 +7,14 @@ Custom types for annotations.
 from __future__ import annotations
 
 import datetime
-from enum import Enum
 from types import ModuleType
 from typing import Any, Iterator, List, Literal, Optional, Tuple, TypedDict, Union
 
 from pydantic.datetime_parse import parse_datetime
 from typing_extensions import Protocol
 
+from datajunction_server.enum import StrEnum
+
 
 class SQLADialect(Protocol):  # pylint: disable=too-few-public-methods
     """
@@ -48,7 +49,7 @@ Row = Tuple[Any, ...]
 Stream = Iterator[Row]
 
 
-class TypeEnum(str, Enum):
+class TypeEnum(StrEnum):
     """
     PEP 249 basic types.
 
@@ -64,7 +65,7 @@ class TypeEnum(str, Enum):
     UNKNOWN = "UNKNOWN"
 
 
-class QueryState(str, Enum):
+class QueryState(StrEnum):
     """
     Different states of a query.
     """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/datajunction_server/utils.py

@@ -4,7 +4,6 @@ Utility functions.
 import logging
 import os
 import re
-from enum import Enum
 from functools import lru_cache
 from string import ascii_letters, digits
 
@@ -19,6 +18,7 @@ from starlette.requests import Request
 from yarl import URL
 
 from datajunction_server.config import Settings
+from datajunction_server.enum import StrEnum
 from datajunction_server.errors import DJException
 from datajunction_server.models.user import User
 from datajunction_server.service_clients import QueryServiceClient
@@ -52,6 +52,7 @@ def get_settings() -> Settings:
     return Settings()
 
 
+@lru_cache(maxsize=None)
 def get_engine() -> Engine:
     """
     Create the metadata engine.
@@ -103,7 +104,7 @@ def get_issue_url(
     return baseurl % query_arguments
 
 
-class VersionUpgrade(str, Enum):
+class VersionUpgrade(StrEnum):
     """
     The version upgrade type
     """

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/pyproject.toml

@@ -2,6 +2,9 @@
 requires = ["hatchling"]
 build-backend = "hatchling.build"
 
+[tool.hatch.build.targets.wheel]
+packages = ["datajunction_server"]
+
 [tool.pdm.build]
 includes = ["dj"]
 

datajunction_server-0.0.1a37.dev0/tester.py

@@ -0,0 +1,24 @@
+def profile(func):
+    from functools import wraps
+
+    @wraps(func)
+    def wrapper(*args, **kwargs):
+        from line_profiler import LineProfiler
+        prof = LineProfiler()
+        try:
+            return prof(func)(*args, **kwargs)
+        finally:
+            prof.print_stats()
+
+    return wrapper
+
+@profile
+def tester():
+    retval = []
+    for x in range(100):
+        retval.extend([1234, 1234])
+    return retval
+
+
+tester()
+

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/tests/api/metrics_test.py

@@ -458,21 +458,15 @@ def test_raise_common_dimensions_metric_not_found(
     )
     assert response.status_code == 500
     assert response.json() == {
-        "message": "Metric node not found: default.foo\nMetric node not found: default.bar",
         "errors": [
             {
                 "code": 203,
-                "message": "Metric node not found: default.foo",
-                "debug": None,
                 "context": "",
-            },
-            {
-                "code": 203,
-                "message": "Metric node not found: default.bar",
                 "debug": None,
-                "context": "",
+                "message": "Metric nodes not found: default.foo,default.bar",
             },
         ],
+        "message": "Metric nodes not found: default.foo,default.bar",
         "warnings": [],
     }
 

{datajunction_server-0.0.1a37 → datajunction_server-0.0.1a37.dev0}/tests/api/nodes_test.py

@@ -4957,7 +4957,7 @@ def test_resolving_downstream_status(client_with_service_setup: TestClient) -> N
         (metric3, NodeType.METRIC),
     ]:
         response = client_with_service_setup.post(
-            f"/nodes/{node_type.value}/",
+            f"/nodes/{node_type.value}/",  # pylint: disable=no-member
             json=node,
         )
         assert response.status_code == 201