hpos-data/scripts/recompute_class_blind.py

import pandas as pd
import os
import numpy as np
import re
import math

curdir = os.getcwd()
path_delim = '/'
df = pd.read_excel("data/_tmp_data_21_02_2024_17_29.xlsx", sheet_name="data")

def device_ratio_classification(ratio):
    try:
        if ratio is not None:
            if 0.16 <= ratio <= 0.23:
                return "Normal"
            if 0.23 <= ratio <= 0.25:
                return "Negative Borderline"
            if 0.25 <= ratio <= 0.31:
                return "Sickle Cell Trait"
            if 0.31 <= ratio <= 0.36:
                return "Positive for Sickle Cell. HPLC for Confirmation"
            if 0.36 <= ratio <= 0.7:
                return "Sickle Cell Disease"
        else:
            return "Invalid"
    except Exception as e:
        return "Error"
    return "Invalid"

def find_result_with_borderline_method_1(device_ratio, device_ratio_class, borderline_metric):
    try:
        if device_ratio is not None and borderline_metric is not None:
            if device_ratio_class == "Negative Borderline":
                return "Borderline. Normal" if borderline_metric >= 2.4 else "Borderline. Sickle Cell Trait"
            elif device_ratio_class == "Positive for Sickle Cell. HPLC for Confirmation":
                return "Borderline. Sickle Cell Trait" if borderline_metric >= 1.34 else "Borderline. Sickle Cell Disease"
    except Exception as e:
        handle_exception(e)
        return "Error"
    return str(device_ratio_class)

def device_ratio_borderline_thresholds_br_method_2(ratio):
    try:
        if ratio is not None:
            if 0.11 <= ratio <= 0.237:
                return "Normal"
            if 0.237 <= ratio <= 0.242:
                return "Negative Borderline"
            if 0.242 <= ratio <= 0.318:
                return "Sickle Cell Trait"
            if 0.318 <= ratio <= 0.356:
                return "Positive for Sickle Cell. HPLC for Confirmation"
            if 0.356 <= ratio <= 0.7:
                return "Sickle Cell Disease"
        else:
            return "Invalid"
    except Exception as e:
        handle_exception(e)
        return "Error"
    return "Invalid"

def reclassify_with_borderline_method_2(device_ratio, device_ratio_class, led2_average):
    try:
        if device_ratio is not None and led2_average is not None:
            if device_ratio_class == "Negative Borderline":
                return "Borderline. Normal" if led2_average >= 0.15 else "Borderline. Sickle Cell Trait"
            elif device_ratio_class == "Positive for Sickle Cell. HPLC for Confirmation":
                return "Borderline. Sickle Cell Trait" if led2_average >= 0.19 else "Borderline. Sickle Cell Disease"
    except Exception as e:
        handle_exception(e)
        return "Error"
    return str(device_ratio_class)

def handle_exception(exception):
    pass

df["tmpDeviceRatioClass1"] = df['deviceRatio'].apply(device_ratio_classification)
df["borderlineMetric1"] = (df['led1Average'] - df['led2Average']) / df['deviceRatio']
df['newDeviceRatioClass1'] = df.apply(lambda row: find_result_with_borderline_method_1(
    row['deviceRatio'],
    row['tmpDeviceRatioClass1'],
    row['borderlineMetric1']
), axis=1)

df["tmpDeviceRatioClass2"] = df['deviceRatio'].apply(device_ratio_borderline_thresholds_br_method_2)
df['newDeviceRatioClass2'] = df.apply(lambda row: reclassify_with_borderline_method_2(
    row['deviceRatio'],
    row['tmpDeviceRatioClass2'],
    row['led2Average']
), axis=1)

# clean columns
df = df[["_id", "name", "classificationResult", "newDeviceRatioClass1", "newDeviceRatioClass2", "deviceRatio", "led1Average", "led2Average"]]
print(df)

writer = pd.ExcelWriter(curdir + path_delim + "data/output.xlsx", engine = 'openpyxl')
df.to_excel(writer, sheet_name = 'op', index=False)
writer.close()