From 1dc2def8545ec7a96dcb4ddf19edd178191c6ac7 Mon Sep 17 00:00:00 2001
From: Pritimay Sarkar <pritimay@sminnovations.in>
Date: Thu, 22 Feb 2024 00:17:48 +0530
Subject: [PATCH] add reclassification script

---
 scripts/recompute_class_blind.py | 98 ++++++++++++++++++++++++++++++++
 1 file changed, 98 insertions(+)
 create mode 100644 scripts/recompute_class_blind.py

diff --git a/scripts/recompute_class_blind.py b/scripts/recompute_class_blind.py
new file mode 100644
index 0000000..66c7d47
--- /dev/null
+++ b/scripts/recompute_class_blind.py
@@ -0,0 +1,98 @@
+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()
\ No newline at end of file