Module matchops.constructor_tools

generic specs and func defs for data matching operations

Functions

def add_column_range(new_column_template: str, source_columns_template: list[str], apply_function: collections.abc.Callable[[str, pandas.core.series.Series], str], rng: range) ‑> list[AddColumn]

add an AddColumn definition for a range of array keys

Args

new_column_template : str

string formattable with array index

source_columns_template : list[str]

list of strings as above

apply_function : Callable[[str, pd.Series], str]

first arg is

array index. 2nd is standard row (aka series) of source columns.

rng : range

range object containing indexes to build columns for

Returns

list[AddColumn]

AddColumn list for all indexes in rng

def anes_staff_w_npi_columns(md_name_column: str, md_npi_column: str, staff_column: str) ‑> list[AddColumn]

generate columns for populating anesthesiaStaff rows from data in csv/tsv e.g. using "Anesthesiologist", "MDA NPI", and "Anesthesia Staff" columns from msa_bmd

Args

md_name_column : str

column containing the anesthesiologist name

md_npi_column : str

column containing the anesthesiologist NPI

staff_column : str

column containing multiline text representation of

all provider times

Returns

list[AddColumn]

list of new columns to fully populate

schedule.anesthesiaStaff. column names will map directly to standard flat output keys (e.g. schedule.anesthesiaStaff[0].provider), so mappings.py entries are not required.

def full_name_column(first_column: str, middle_column: str, last_column: str, new_column_name: str, suffix_column: str = None) ‑> AddColumn

given the names of columns for first, middle, and last name, return properly formatted full name

Args

first_column : str

column name that contains first name

last_column : str

column name that contains last name

middle_column : str

column name that contains middle name

new_column_name : str

name of the column to be added

suffix_column : str | None

optional column containing suffix

Returns

AddColumn

AddColumn object to create full name

def graphium_composite_columns(hank_key_template_map: str, source_column: str, repeat_count: int) ‑> list[AddColumn]

Create a series of add columns for a graphium repeating composite field.

Args

hank_key_template_map : str

a string containing a series of Hank keys

separated by "^" that map to the composite field's subfields. Each key

should contain a pair of curly braces to be formatted with the repeat

index (e.g. "schedule.anesthesiaStaff[{}].provider").

source_column : str

the column containing the composite field data.

repeat_count : int

the number of times the composite field repeats.

Example

When unpacked in a DataFrameConstructor's add_columns parameter, the output from the following implementation: graphium_composite_columns( hank_key_template_map=( "schedule.anesthesiaStaff[{}].provider" "^schedule.anesthesiaStaff[{}].provider" "^schedule.anesthesiaStaff[{}].providerNPI" "^" "^schedule.anesthesiaStaff[{}].role" "^schedule.anesthesiaStaff[{}].startTime" "^schedule.anesthesiaStaff[{}].startTime" "^schedule.anesthesiaStaff[{}].endTime" "^schedule.anesthesiaStaff[{}].endTime" ), source_column="AnesthesiaStaff", repeat_count=2, ) will add the provider, providerNPI, role, startTime, and endTime columns for schedule.anesthesiaStaff[0] and schedule.anesthesiaStaff[1]. Increase repeat_count to add an arbitrary number of such columns.

Returns

list[AddColumn]

as described above.

def hl7_joins_for_schedule_columns(patient_name: str, dos: str, mrn: str, dob: str, accn: str) ‑> list[JoinSpec]

get standard joins for hl7 demographic merge using supplied schedule column names for dos (date of service), mrn (medical record number), dob (date of birth) and accn (hospital account number)

def split_address_columns(source_column: str, prefix='') ‑> list[AddColumn]

returns a list of add columns for all address subfields from a source column containing a full address.

Args

source_column : str

column name containing full address

prefix : str

Optional. prefix to add to new column names. Defaults to "". Set to a valid "address" prefix (e.g. "patient_info.patient.address.") to return columns with valid summary_spec keys.

See transform_utils.address_split for details.

def split_name_columns(source_column: str, jmespath_keys=False) ‑> list[AddColumn]

returns a list of AddColumns representing the first, last, and middle name elements extracted from the value contained in source_column. new columns will be of form f'{source_column} {[element]}' where element is one of ["First", "Middle", "Last"], e.g. "Patient Name First".

def split_phone_columns(source_column: str, class_dict=None, prefix='') ‑> list[AddColumn]

returns a list of add columns for homePhone, workPhone, and mobilePhone based on the value of a consolidated source_column.

Example:
    given consolidated source_column value:
        'H: 222-222-2222

W: 888-888-8888 M: 555-555-5555' returns new columns equivalent to: { "homePhone": "+1 222-222-2222", "workPhone": "+1 888-888-8888", "mobilePhone": "+1 555-555-5555", }

Args:
    source_column (str): column name containing consolidated phone numbers
    class_dict (dict[str, str]): Optional. dictionary mapping phone class
        to format string. Defaults to None.
    prefix (str): Optional. prefix to add to new column names. Defaults to "".
        Set to a valid "person" prefix (e.g. "patient_info.patient.") to
        return columns with valid summary_spec keys.

See transform_utils.phone_split for details.

def standard_demographics_constructor(split_name_column_list: list[str], split_phone_column_list: list[str], add_columns: list[AddColumn], post_process: collections.abc.Callable[[pandas.core.frame.DataFrame], pandas.core.frame.DataFrame]) ‑> DataFrameConstructor

used for source CSVs following the Hank AI CSV field standard

Args

split_name_column_list : list[str]

list of *.fullName columns to split

split_phone_column_list : list[str]

list of *.homePhone columns to split

add_columns : list[AddColumn]

additional implementation specific AddColumns

post_process : Callable[[pd.DataFrame], pd.DataFrame]

standard post_process

Returns

DataFrameConstructor

demographics dataframe constructor