Removed redundant code.

data_processing.py

-import pandas as pd # add to requirements.txt
-import numpy as np  # add to requirements.txt
-
-
-def format_data(dataframe):
-    """
-        Format the data in the dataframe, converting the columns
-        to the appropriate types and returning the processed dataframe
-
-        inputs:
-            - dataframe: the source dataframe
-        output:
-            - dataframe: the formatted DataFrame with the following columns:
-                - age
-                - sex
-                - n 'creatinine_result_i' columns
-                - n-1 'elapsed_exam_i-(i+1)' columns
-                - min_creatinine
-                - avg_creatinine
-                - max_creatinine
-            - aki_values: the binary aki values of the original dataframe
-    """
-    # Convert
-
-    aki_values = dataframe['aki']
-    dataframe.drop(columns=['aki'], inplace=True)
-    return (dataframe, aki_values)
-
-
-def process_data(dataframe):
-    """
-        Process the data in the dataframe, converting the columns
-        to the appropriate types and returning the processed dataframe
-
-        inputs:
-            - dataframe: the source dataframe
-        output:
-            - dataframe: the formatted dataframe with the following columns:
-                - age
-                - sex
-                - n 'creatinine_result_i' columns
-                - n-1 'elapsed_exam_i-(i+1)' columns
-                - min_creatinine
-                - avg_creatinine
-                - max_creatinine
-            - aki_values: the binary aki values of the original dataframe
-    """
-    # Preprocess the data
-    (dataframe, aki_values) = format_data(dataframe)
-    return (dataframe, aki_values)
-
-
-def calculate_elapsed_time(dataframe, total_tests):
-    """
-        Calculate the elapsed days between each blood test
-
-        inputs: 
-            - dataframe: the source dataframe
-        output:
-            - elapsed_days: dataframe with n-1 columns corresponding to the
-                            elapsed days between each n blood test
-    """
-    columns = get_creatinine_dataframe_columns(total_tests, 'date')
-    exams_time_difference = {}
-    prev = 0
-    for next in range(1, total_tests):
-        exams_time_difference[f'elapsed_exam_{prev}-{next}'] = (dataframe.loc[:][columns[next]]-dataframe.loc[:][columns[prev]])
-        prev = next
-
-    elapsed_days = pd.DataFrame(exams_time_difference).astype('timedelta64[ns]').apply(lambda x: x.dt.days)
-
-    return elapsed_days
-
-
-def get_creatinine_stats(dataframe, total_tests):
-    columns = get_creatinine_dataframe_columns(total_tests, 'result')
-    creatinine_stats = {}
-    creatinine_stats['min_creatinine'] = dataframe.loc[:][columns].min(axis=1)
-    creatinine_stats['avg_creatinine'] = dataframe.loc[:][columns].mean(axis=1)
-    creatinine_stats['max_creatinine'] = dataframe.loc[:][columns].max(axis=1)
-
-    creatine_ds = pd.DataFrame(creatinine_stats)
-    return creatine_ds
-
-
-def get_creatinine_dataframe_columns(total_tests, col_type):
-    """
-        Return a list of creatinine columns in the dataset in the
-        form 'creatinine_x_5', with x equal to:
-            - 'result', if col_type is the result of the test
-            - 'date', if col_type is the date of the test
-
-        inputs:
-            - total_tests: the maximum number of creatinine tests taken by any patient
-            - col_type: the type of the column, either 'result' or 'date'
-        output:
-            - creatinine_columns: list of column names
-    """
-    creatinine_columns = []
-    for i in range(total_tests):
-        creatinine_columns.append(f'creatinine_{col_type}_{i}')
-    return creatinine_columns
-
-
-def convert_to_datetime(dataframe, columns):
-    """
-        Convert the given dataframe columns in the list 'columns' 
-        to datetime type
-        
-        inputs:
-            - dataframe: the source dataframe
-            - columns: list of column names to convert 
-        output:
-            - dataframe: the processed dataframe
-    """
-    for column in columns:
-        dataframe[column] = pd.to_datetime(dataframe[column])
-    return dataframe
-
-
-def to_binary(x):
-    """ 
-        Map the aki and sex columns of the dataframe
-        to binary values
-
-       output:
-            - 0: if the aki value is 'n' or if the sex value is 'm'
-            - 1: if the aki value is 'y' or if the sex value is 'f'
-    """
-    if x == 'y': return 1
-    elif x == 'f': return 1
-    else: return 0
\ No newline at end of file

nhs_aki_algo.py

-import pandas as pd
-from utils import *
-
-def nhs_aki_algo(patient, data_type = 'train'):
-    """
-    """
-    [sex, age, c1, rv1, rv2, rv_ratio, D] = process_patient_data(patient, data_type)
-
-    (low_ri, high_ri) = get_reference_interval(sex, age)
-
-    elapsed_days = 0
-    if elapsed_days <= 0:
-        print("Error: invalid elapsed time between most recent creatinine test results.")
-        return 0
-    elif (elapsed_days > 0) and (elapsed_days <= 7):
-        rv_ratio = c1/rv1
-    elif (elapsed_days > 7) and (elapsed_days <= 365):
-        rv_ratio = c1/rv2
-    elif elapsed_days > 365:
-        return 0
-    
-    if rv_ratio >= 1.5:
-        return 1
-    elif D > 26:
-        return 1
-    else:
-        return 0
-        
-