FHIRflat Base Class

FHIRflat Base Class#

class fhirflat.resources.base.FHIRFlatBase(*, resource_type: str = 'DomainResource', fhir_comments: str | List[str] = None, id: Id = None, implicitRules: Uri = None, _implicitRules: FHIRPrimitiveExtensionType = None, language: Code = None, _language: FHIRPrimitiveExtensionType = None, meta: MetaType = None, contained: List[ResourceType] = None, extension: List[ExtensionType] = None, modifierExtension: List[ExtensionType] = None, text: NarrativeType = None)#

Base class for FHIR resources to add FHIRflat functionality.

classmethod attr_lists() → list[str]#

Attributes which take a list of FHIR types.

classmethod cleanup(data: dict) → dict#

Apply resource-specific changes to references and default values

classmethod create_fhir_resource(data: str | dict) → FHIRFlatBase | ValidationError#

Load data into a dictionary-like structure, then apply resource-specific changes and unpack flattened data like codeableConcepts back into structured data. Creates a FHIR resource from the data.

classmethod fhir_bulk_import(file: str) → FHIRFlatBase | list[FHIRFlatBase]#

Takes a ndjson file containing FHIR resources as json strings and returns a list of populated FHIR resources.

Parameters:

file – Path to the .ndjson file containing FHIR data

classmethod fhir_file_to_flat(source_file: str, output_name: str | None = None)#

Converts a .ndjson file of exported FHIR resources to a FHIRflat parquet file.

Parameters:

  • source_file – Path to the FHIR resource file.

  • output_name – Name of the parquet file to be generated, optional, defaults to {resource}.parquet

classmethod flat_fields() → list[str]#

All fields that are present in the FHIRflat representation

classmethod from_flat(file: str) → FHIRFlatBase | list[FHIRFlatBase]#

Takes a FHIRflat parquet file and populates the resource with the data.

Parameters:

file – Path to the parquet FHIRflat file containing clinical data

Return type:

FHIRFlatBase or list[FHIRFlatBase]

classmethod ingest_backbone_elements(mapped_data: Series) → Series#

Unflattens ordered lists of data and forms the correct FHIR format which won’t be flattened after ingestion (*_dense columns).

Extends the flat2fhir.expand_concepts function specifically for data ingestion.

Parameters:

mapped_data – Pandas series of FHIRflat-like dictionaries ready to be converted to FHIR format.

classmethod ingest_to_flat(data: DataFrame) → DataFrame | None#

Takes a pandas dataframe containing the populated mapping file and a dictionary representing the FHIRflat resource and creates the FHIRflat parquet file. Performs data formatting on the date and coding columns to account for simplifications parquet makes when saving.

Parameters:

data – Pandas dataframe containing the raw data

Return type:

A dataframe containing the FHIRflat data.

to_flat(filename: str | None = None) → None | Series#

Generates a FHIRflat parquet file from the resource, or returns a Series

Parameters:

filename – Name of the parquet file to be generated.

classmethod validate_fhirflat(df: DataFrame, return_frames: bool = False) → tuple[FHIRFlatBase | DataFrame, None | DataFrame]#

Takes a FHIRflat dataframe and validates the data against the FHIR schema. Returns a dataframe of valid resources and a dataframe of the FHIRflat data that produced validation errors, with a validation_error column describing the error.

Parameters:

  • df – Pandas dataframe containing the FHIRflat data

  • return_frames – If True, returns the valid FHIR resources & errors as dataframes, even if only one row is present in the source.

Returns:

  • valid_resources – A dataframe containing the valid FHIR resources

  • errors – A dataframe containing the flat_dict and validation errors.

Raises:

ValidationError – If a single FHIR resource is present and is invalid.

Contents