[pre-commit.ci] pre-commit autoupdate

.pre-commit-config.yaml

@@ -1,19 +1,19 @@
 repos:
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: "v0.6.2"
+    rev: "v0.16.6"
     hooks:
       - id: ruff
         args: [--fix, --show-fixes]
         types_or: [python]
 
   - repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v4.6.0
+    rev: v6.0.0
     hooks:
       - id: check-yaml
       - id: end-of-file-fixer
       - id: trailing-whitespace
 
-  - repo: https://github.com/psf/black
-    rev: 24.8.0
+  - repo: https://github.com/psf/black-pre-commit-mirror
+    rev: 26.5.1
     hooks:
     - id: black

fhirflat/__main__.py

@@ -6,14 +6,12 @@
 
 def main():
     if len(sys.argv) < 2:
-        print(
-            """fhirflat: specify subcommand to run
+        print("""fhirflat: specify subcommand to run
 
                 Available subcommands:
                 transform - Convert raw data into FHIRflat files
                 validate - Validate FHIRflat files against FHIR schemas
-            """
-        )
+            """)
         sys.exit(1)
     subcommand = sys.argv[1]
     if subcommand not in ["transform", "validate"]:

fhirflat/flat2fhir.py

@@ -151,10 +151,10 @@
 
     quant = {}
 
-    for attribute in df.keys():
+    for attribute in df:
         attr = attribute.split(".")[-1]
         if attr == "code":
-            if group + ".system" in df.keys():
+            if group + ".system" in df:
                 # reading in from ingestion pipeline
                 quant["code"] = df[group + ".code"]
                 quant["system"] = df[group + ".system"]
@@ -213,7 +213,7 @@
     klass = get_local_extension_type(k)
 
     prop = klass.schema()["properties"]
     value_type = [key for key in prop.keys() if key.startswith("value")]
 
     if not value_type:  # pragma: no cover
         raise RuntimeError("Inappropriate entry into create_single_extension")
@@ -249,7 +249,7 @@
                 except ValidationError:
                     continue
             else:
                 raise e  # pragma: no cover
 
     raise RuntimeError(f"extension not created from {k, v}")  # pragma: no cover
 
@@ -309,7 +309,7 @@
 
     if klass.nested_extension:
         classes = find_data_class_options(klass, "extension")
-        short_extensions = [s for s in v_dict.keys() if s.count(".") == 0]
+        short_extensions = [s for s in v_dict if s.count(".") == 0]
         expanded_short_extensions = []
         if short_extensions:
             # these get skipped over in expand_concepts because they don't get grouped
@@ -393,7 +393,7 @@
     groups = group_keys(data.keys())
     group_classes = {}
 
     for k in groups.keys():
         group_classes[k] = find_data_class_options(data_class, k)
 
     expanded = {}
@@ -448,9 +448,7 @@
             else:
                 expanded[k] = [expanded[k]]
 
-    dense_cols = {
-        k: k.removesuffix("_dense") for k in data.keys() if k.endswith("_dense")
-    }
+    dense_cols = {k: k.removesuffix("_dense") for k in data if k.endswith("_dense")}
     if dense_cols:
         for old_k, new_k in dense_cols.items():
             data[new_k] = data[old_k]

fhirflat/ingest.py

@@ -106,7 +106,7 @@
     new_tz = ZoneInfo(timezone)
 
     try:
         date_time = datetime.strptime(date_str, date_format)
         date_time_aware = date_time.replace(tzinfo=new_tz)
         if "%H" not in date_format:
             date_time_aware = date_time_aware.date()
@@ -114,7 +114,7 @@
         try:
             # Unconverted data remains in the string (i.e. time is present)
             date, time = date_str.split(" ")
             date = datetime.strptime(date, date_format)
             time = dateutil.parser.parse(time).time()
             date_time = datetime.combine(date, time)
             date_time_aware = date_time.replace(tzinfo=new_tz)
@@ -220,9 +220,7 @@
                             target_length = max(map(len, relevant_result.values()))
                             for k, v in relevant_result.items():
                                 if len(v) < target_length:
-                                    result[k] = relevant_result[k] + [None] * (
-                                        target_length - len(v)
-                                    )
+                                    result[k] = v + [None] * (target_length - len(v))
     return result
 
 
@@ -441,7 +439,7 @@
     sheet_id: str | None = None,
     subject_id="subjid",
     validate: bool = True,
     compress_format: None | str = None,
     parallel: bool = False,
 ):
     """
@@ -500,7 +498,7 @@
 
         df_types = pd.read_csv(sheet_link, header=0, index_col="Resources")
         types = dict(zip(df_types.index, df_types["Resource Type"], strict=True))
         sheet_keys = {r: df_types.loc[r, "Sheet ID"] for r in types.keys()}
         mappings = {
             get_local_resource(r): sheet_link + f"&gid={i}"
             for r, i in sheet_keys.items()
@@ -523,7 +521,7 @@
                 timezone=timezone,
             )
             if df is None:
-                return None
+                return
         else:
             raise ValueError(f"Unknown mapping type {t}")
 
@@ -557,7 +555,7 @@
                     UserWarning,
                     stacklevel=2,
                 )
-                return None
+                return
 
         valid_time = timeit.default_timer()
         print(f"{resource.__name__} validation in " + str(valid_time - dict_time))

fhirflat/resources/base.py

@@ -77,7 +77,7 @@
         data = expand_concepts(data, cls)
 
         # create lists for properties which are lists of FHIR types
         for field in [x for x in data.keys() if x in cls.attr_lists()]:
             if not isinstance(data[field], list):
                 data[field] = [data[field]]
 
@@ -89,7 +89,7 @@
     @classmethod
     def validate_fhirflat(
         cls, df: pd.DataFrame, return_frames: bool = False
-    ) -> tuple[FHIRFlatBase | pd.DataFrame, None | pd.DataFrame]:
+    ) -> tuple[FHIRFlatBase | pd.DataFrame, pd.DataFrame | None]:
         """
         Takes a FHIRflat dataframe and validates the data against the FHIR
         schema. Returns a dataframe of valid resources and a dataframe of the
@@ -276,7 +276,7 @@
             flat_df.drop(columns=system_columns, inplace=True)
 
             potential_dense_cols = [
                 x for x in cls.backbone_elements.keys() if x in flat_df.columns
             ]
 
             long_list_cols = [
@@ -355,7 +355,7 @@
 
         df.to_parquet(output_name)
 
-    def to_flat(self, filename: str | None = None) -> None | pd.Series:
+    def to_flat(self, filename: str | None = None) -> pd.Series | None:
         """
         Generates a FHIRflat parquet file from the resource, or returns a Series
 

fhirflat/resources/condition.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from typing import ClassVar, TypeAlias, Union
+from typing import ClassVar, TypeAlias
 
 from fhir.resources import fhirtypes
 from fhir.resources.condition import Condition as _Condition
@@ -30,22 +30,18 @@
 
 class Condition(_Condition, FHIRFlatBase):
     extension: list[
-        Union[
-            presenceAbsenceType,
-            prespecifiedQueryType,
-            timingPhaseType,
-            timingPhaseDetailType,
-            fhirtypes.ExtensionType,
-        ]
+        presenceAbsenceType
+        | prespecifiedQueryType
+        | timingPhaseType
+        | timingPhaseDetailType
+        | fhirtypes.ExtensionType
     ] = Field(
         None,
         alias="extension",
         title="Additional content defined by implementations",
-        description=(
-            """
+        description=("""
             Contains the G.H 'age' and 'birthSex' extensions,
-            and allows extensions from other implementations to be included."""
-        ),
+            and allows extensions from other implementations to be included."""),
         # if property is element of this resource.
         element_property=True,
         union_mode="smart",

fhirflat/resources/diagnosticreport.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from typing import ClassVar, Union
+from typing import ClassVar
 
 from fhir.resources import fhirtypes
 from fhir.resources.diagnosticreport import (
@@ -21,17 +21,15 @@
 
 class DiagnosticReport(_DiagnosticReport, FHIRFlatBase):
     extension: list[
-        Union[timingPhaseType, timingPhaseDetailType, fhirtypes.ExtensionType]
+        timingPhaseType | timingPhaseDetailType | fhirtypes.ExtensionType
     ] = Field(
         None,
         alias="extension",
         title="List of `Extension` items (represented as `dict` in JSON)",
-        description=(
-            """
+        description=("""
             Contains the Global.health 'timingPhase' extension,
             and allows extensions from other implementations to be included.
-            """
-        ),
+            """),
         # if property is element of this resource.
         element_property=True,
         # this trys to match the type of the object to each of the union types
@@ -86,7 +84,7 @@ def cleanup(cls, data: dict) -> dict:
                 "study",
                 "composition",
             }
-            | {x for x in data.keys() if x.endswith(".reference")}
+            | {x for x in data if x.endswith(".reference")}
         ).intersection(data.keys()):
             data[field] = {"reference": data[field]}
 

fhirflat/resources/encounter.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from typing import ClassVar, TypeAlias, Union
+from typing import ClassVar, TypeAlias
 
 from fhir.resources import fhirtypes
 from fhir.resources.encounter import Encounter as _Encounter
@@ -22,22 +22,18 @@
 
 class Encounter(_Encounter, FHIRFlatBase):
     extension: list[
-        Union[
-            relativePeriodType,
-            timingPhaseType,
-            timingPhaseDetailType,
-            fhirtypes.ExtensionType,
-        ]
+        relativePeriodType
+        | timingPhaseType
+        | timingPhaseDetailType
+        | fhirtypes.ExtensionType
     ] = Field(
         None,
         alias="extension",
         title="List of `Extension` items (represented as `dict` in JSON)",
-        description=(
-            """
+        description=("""
             Contains the Global.health 'eventTiming' and 'relativePeriod' extensions,
             and allows extensions from other implementations to be included.
-            """
-        ),
+            """),
         # if property is element of this resource.
         element_property=True,
         # this trys to match the type of the object to each of the union types

fhirflat/resources/extension_types.py

@@ -10,7 +10,7 @@
 
 class AbstractType(_AbstractType):
     @classmethod
-    def __get_validators__(cls) -> "CallableGenerator":
+    def __get_validators__(cls) -> CallableGenerator:
         from . import extension_validators as validators
 
         yield getattr(validators, cls.__resource_type__.lower() + "_validator")

fhirflat/resources/extension_validators.py

@@ -37,7 +37,7 @@
 import importlib
 import typing
 from pathlib import Path
-from typing import TYPE_CHECKING, Type, Union
+from typing import TYPE_CHECKING
 
 from fhir.resources.core.fhirabstractmodel import FHIRAbstractModel
 from pydantic.v1.class_validators import make_generic_validator
@@ -72,7 +72,7 @@ def __init__(self):
             "dateTimeExtension": (None, ".extensions"),
         }
 
-    def get_fhir_model_class(self, model_name: str) -> Type[FHIRAbstractModel]:
+    def get_fhir_model_class(self, model_name: str) -> type[FHIRAbstractModel]:
         """
         Returns the extension class by finding the 'datetimeextension' file and
         importing the type class.
@@ -95,11 +95,11 @@ def run_validator_for_fhir_type(self, model_type_cls, v, values, config, field):
         return v
 
     def fhir_model_validator(
-        self, model_name: str, v: Union[StrBytes, dict, Path, FHIRAbstractModel]
+        self, model_name: str, v: StrBytes | dict | Path | FHIRAbstractModel
     ):
         """ """
-        model_class: Type[BaseModel] | Type[FHIRAbstractModel] = (
-            self.get_fhir_model_class(model_name)
+        model_class: type[BaseModel | FHIRAbstractModel] = self.get_fhir_model_class(
+            model_name
         )
 
         if isinstance(v, (str, bytes)):
@@ -108,7 +108,7 @@ def fhir_model_validator(
             except ValidationError as exc:
                 if TYPE_CHECKING:
                     model_class = typing.cast(
-                        Type[BaseModel], model_class
+                        type[BaseModel], model_class
                     )  # pragma: no cover
                 errors = exc.errors()
                 if (
@@ -200,61 +200,61 @@ def fhir_model_validator(
         return v
 
 
-def timingphase_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def timingphase_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("timingPhase", v)
 
 
-def timingdetail_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def timingdetail_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("timingDetail", v)
 
 
-def timingphasedetail_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def timingphasedetail_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("timingPhaseDetail", v)
 
 
-def relativeday_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def relativeday_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("relativeDay", v)
 
 
-def relativestart_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def relativestart_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("relativeStart", v)
 
 
-def relativeend_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def relativeend_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("relativeEnd", v)
 
 
-def relativeperiod_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def relativeperiod_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("relativePeriod", v)
 
 
-def approximatedate_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def approximatedate_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("approximateDate", v)
 
 
-def duration_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def duration_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("Duration", v)
 
 
-def age_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def age_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("Age", v)
 
 
-def birthsex_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def birthsex_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("birthSex", v)
 
 
-def race_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def race_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("Race", v)
 
 
-def presenceabsence_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def presenceabsence_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("presenceAbsence", v)
 
 
-def prespecifiedquery_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def prespecifiedquery_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("prespecifiedQuery", v)
 
 
-def datetimeextension_validator(v: Union[StrBytes, dict, Path, FHIRAbstractModel]):
+def datetimeextension_validator(v: StrBytes | dict | Path | FHIRAbstractModel):
     return Validators().fhir_model_validator("dateTimeExtension", v)

fhirflat/resources/extensions.py

@@ -7,7 +7,7 @@
 
 from __future__ import annotations
 
-from typing import Any, ClassVar, Union
+from typing import Any, ClassVar
 
 from fhir.resources import fhirtypes
 from fhir.resources.datatype import DataType as _DataType
@@ -216,7 +216,7 @@ class timingPhaseDetail(_ISARICExtension):
 
     nested_extension: ClassVar[bool] = True
 
-    extension: list[Union[et.timingPhaseType, et.timingDetailType]] = Field(
+    extension: list[et.timingPhaseType | et.timingDetailType] = Field(
         None,
         alias="extension",
         title="List of `Extension` items (represented as `dict` in JSON)",
@@ -367,7 +367,7 @@ class relativePeriod(_ISARICExtension):
 
     nested_extension: ClassVar[bool] = True
 
-    extension: list[Union[et.relativeStartType, et.relativeEndType]] = Field(
+    extension: list[et.relativeStartType | et.relativeEndType] = Field(
         None,
         alias="extension",
         title="List of `Extension` items (represented as `dict` in JSON)",
@@ -712,7 +712,7 @@ class dateTimeExtension(_FHIRPrimitiveExtension):
     resource_type: str = Field(default="dateTimeExtension", const=True)
 
     extension: list[
-        Union[et.approximateDateType, et.relativeDayType, fhirtypes.ExtensionType]
+        et.approximateDateType | et.relativeDayType | fhirtypes.ExtensionType
     ] = Field(
         None,
         alias="extension",