Tongtong rbc

.gitignore

@@ -0,0 +1 @@
+.env

anaemia.py

@@ -0,0 +1,52 @@
+testdict = {
+    "mcv":{
+		"test_found":True,
+		"test_value":60,
+		"test_unit":False,
+		"test_ref_min":85,
+		"test_ref_max":False
+		},
+    "rbc_count":{
+ 		"test_found":True,
+		"test_value":60,
+		"test_unit":False,
+		"test_ref_min":85,
+		"test_ref_max":False       
+    },
+	"hb":{
+		"test_found":True,
+		"test_value":11,
+		"test_unit":False,
+		"test_ref_min":12,
+		"test_ref_max":16
+		},
+}
+#mcv need fl RBC need 10^6/uL
+#returns tuple - bool whether anaemia, and bool whether iron deficient 
+def anaemia_analysis (hbdict):
+	output_phrase = "I didn't catch that"
+	print ("in anaemia analysis")
+	hblevel = hbdict["hb"]["test_value"]
+	if hblevel < 7:
+		return ":large_orange_circle: Your haemoglobin level is dangerously low. Please visit a doctor immediately."
+	if hblevel < hbdict["hb"]["test_ref_min"]:
+		if not hbdict["mcv"]["test_found"]:
+			return ":large_orange_circle: You likely have anaemia, but there is not enough information to determine the cause. Visit a doctor if you have any concerns of blood loss."
+		mcv = hbdict["mcv"]["test_value"]
+		if mcv < hbdict["mcv"]["test_ref_min"]: #microcytic
+			if not hbdict["rbc_count"]["test_found"]:
+				return ":large_orange_circle: You likely have anaemia, but there is not enough information to determine the cause. Visit a doctor if you have any concerns of blood loss."
+			mentzer = mcv/ hbdict["rbc_count"]["test_value"]
+			if mentzer > 13:
+				output_phrase = ":large_orange_circle: You likely have anaemia, which could be due to iron deficiency. This can be caused by minor bleeding e.g. menstruation, or from the gastrointestinal tract. Visit your doctor for an assessment to rule out sources of bleeding. Consider taking iron supplements and foods rich in iron, such as green leafy vegetables, meat, especially red meat (beef, mutton, pork), seafood, and organs (kidney, liver)."
+			else :
+				output_phrase = ":large_orange_circle: You likely have anaemia, which could be related to genetic conditions such as thalassaemia. Visit your doctor for an assessment." 
+		elif mcv > hbdict["mcv"]["test_ref_max"]: # macrocytic
+			output_phrase = ":large_orange_circle: You likely have anaemia, which may be caused by low folate (vitamin B9) or vitamin B12 levels. Other causes may include chronic alcohol intake, thyroid problems, and liver problems. Consider taking more foods high in folate, such as broccoli, spinach, and brown rice, and foods high in vitamin B12, such as meat, milk, cheese and eggs."
+		else: #normocytic 
+			output_phrase = ":large_orange_circle: You likely have anaemia, which may be associated with chronic illness such as kidney disease, or chronic inflammatory conditions."
+	else:
+		output_phrase = ":large_green_circle: You likely do not have anaemia."
+	return output_phrase
+
+#print (f"here is result: {anaemia_analysis (testdict)}")

bmi.py

@@ -0,0 +1,43 @@
+test_results = {
+	"height":{
+		"test_found":False,
+		"test_value":1.5,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		},
+	"weight":{
+		"test_found":False,
+		"test_value":70,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		}
+}
+
+
+def bmi_advice(test_results):
+    print ("in bmi advice")
+    output_string = "I didn't catch that"
+    weight = test_results["weight"]["test_value"]
+    height = test_results["height"]["test_value"]
+    if weight > 0 and height > 0:
+        bmi = round(weight / pow(height,2), 1)
+    else:
+        return ":large_yellow_circle: You need both weight and height to calculate BMI."
+    weightloss = False
+    if bmi >=27.5: 
+        output_string = ":large_orange_circle:  You are obese. Your BMI is " + str(bmi)
+        weightloss = True
+    elif bmi >= 23:
+        output_string = ":large_orange_circle:  You are overweight. Your BMI is " + str(bmi)
+        weightloss = True
+    elif bmi < 18.5:
+        output_string = ":large_orange_circle:  You are underweight. Your BMI is " + str(bmi)+ " Consider increasing food intake, for example, by taking smaller, frequent healthy meals. Increase protein intake by taking more lean meats, fish, eggs, dairy, legumes and nuts. Do strength training to build up muscles."
+    else:
+        output_string = ":large_green_circle: You have a healthy BMI. Your BMI is " + str(bmi)
+    if weightloss:
+        output_string += "Choose healthier choices that are lower in fat (e.g. lean meat, low-fat dairy products), lower or no sugar (e.g. unsweetened beverages, fresh fruits), and higher in fibre (e.g. whole-meal bread, brown rice). Look out for alternatives that are lower in calories. Reduce your meal sizes by consuming ¾ of your usual. Do some moderate-intensity aerobic physical activities such as brisk walking, swimming or cycling for a minimum of 150-300 minutes weekly. If you're just starting out, accumulating shorter bouts of 10-minute exercise is a good start too."
+    return output_string
+
+#print ("advice " + bmi_advice(test_results))

chatgpt_values.py

@@ -0,0 +1,123 @@
+import streamlit as st
+import pandas as pd
+import numpy as np
+import os
+import time
+import base64
+import json
+import cv2
+from openai import OpenAI
+from paddleocr import PaddleOCR
+
+
+template_prompt = """
+Extract the items from the health screening result listed below into a json format, using the example json template. Ignore other items not listed in the json template. Output data types are "test_found" (True/False), "test_value" (float), "test_unit" (string), "test_ref_min" (float), "test_ref_max" (float). If test item is not found, output False for "test_found", and output False for the other values. If test is found but reference max, reference min, or reference range is not found, output False for "test_ref_min" and "test_ref_max". 
+Convert height to metres and weight to kg.
+
+Example output json template
+{
+	"ldl_cholesterol": {
+		"test_found":False,
+		"test_value":False,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		},
+	"hdl_cholesterol": {
+		"test_found":True,
+		"test_value":False,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		},
+	"total_cholesterol": {
+		"test_found":False,
+		"test_value":False,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		},
+	"mcv":{
+		"test_found":False,
+		"test_value":False,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		},
+	"hb":{
+		"test_found":False,
+		"test_value":False,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		},
+    "rbc_count":{
+		"test_found":False,
+		"test_value":False,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		},
+	"glucose":{
+		"test_found":False,
+		"test_value":False,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		},
+	"hba1c":{
+		"test_found":False,
+		"test_value":False,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		},
+	"systolic_bp":{
+		"test_found":False,
+		"test_value":False,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		},
+	"diastolic_bp":{
+		"test_found":False,
+		"test_value":False,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		},
+	"height":{
+		"test_found":False,
+		"test_value":False,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		},
+	"weight":{
+		"test_found":False,
+		"test_value":False,
+		"test_unit":False,
+		"test_ref_min":False,
+		"test_ref_max":False
+		}
+}
+
+Health screening result:
+"""
+
+def extract_values(client,extracted_text):
+    extract_start_time = time.time()
+    extract_prompt = f"{template_prompt} {extracted_text}"
+    response = client.chat.completions.create(
+        model="gpt-4-turbo-preview", #gpt-3.5-turbo-1106
+	seed=4022024,
+        response_format={ "type": "json_object" },
+        messages=[
+            {"role": "system", "content": "You are a helpful assistant designed to output JSON."},
+            {"role": "user", "content": extract_prompt}
+        ]
+    )
+    test_results = json.loads(response.choices[0].message.content)
+    extract_end_time = time.time()
+    extract_time = int(extract_end_time - extract_start_time)
+    return response,test_results,extract_time

diabetes.py

@@ -0,0 +1,77 @@
+testdict = {
+    "glucose":{
+		"test_found":True,
+		"test_value":6.5,
+		"test_unit":"mmol/l",
+		"test_ref_min":False,
+		"test_ref_max":False
+		},
+	"hba1c":{
+		"test_found":True,
+		"test_value":6.4,
+		"test_unit":"%",
+		"test_ref_min":False,
+		"test_ref_max":False
+		}
+}
+
+test_attributes = {
+    "age" : 40, 
+    "sex" :"male", #0 for male, 1 for female 
+    "smoker" : True, 
+    "stroke" : False,
+    "diabetes" : True,
+    "ckd" : False,
+    "heart_attack" : False ,
+    "race" :"indian",
+    "systolic_blood_pressure" : 120,
+    "on_BP_meds" : False
+}
+
+
+def get_dm_advice(inputattributes, inputdict):
+    print ("in dm advice")
+    output_phrase = "I didn't catch that."
+    lifestyle = False
+    #known dm
+    a1c_value = inputdict["hba1c"]["test_value"] if inputdict["hba1c"]["test_found"] else 0 
+    glucose_value = inputdict["glucose"]["test_value"] if inputdict["glucose"]["test_found"] else 0 
+    if inputattributes["diabetes"]:
+        if a1c_value >0:
+            if inputattributes["age"] >= 80:
+                a1ctarget = 8
+            elif inputattributes["age"] <=40:
+                a1ctarget = 6.5 
+            else:
+                a1ctarget = 7.0 
+            if a1c_value > a1ctarget:
+                output_phrase = ":large_orange_circle:  Your HbA1c is above target, your diabetes can be controlled better. Aim for a HbA1c below 7."
+                lifestyle = True
+            else:
+                output_phrase = ":large_green_circle:  Your HbA1c is below 7, which is within expected range for a diabetic."
+        if glucose_value >0:
+            if glucose_value > 6:
+                output_phrase += "  \nYour fasting glucose level is slightly high. Please consult your doctor for your specific glucose targets. Having a HbA1c measurement may be helpful to better evauate your diabetes control."
+            else:
+                output_phrase += "  \nYour fasting glucose level is within range, <6."
+    else:
+        #diagnose dm
+        if a1c_value >= 6.5 or glucose_value >= 7:
+            output_phrase = ":large_orange_circle:  You likely have diabetes. Consult a doctor for advice, you may need to be started on medications."
+            lifestyle = True
+        elif a1c_value >= 6.1 or glucose_value >=6.1:
+            output_phrase = ":large_yellow_circle:  You likely have prediabetes."
+            if a1c_value >= 5.7:
+                output_phrase += " Your HbA1c is >6.0." 
+            if glucose_value >=5.6:
+                output_phrase += " Your fasting glucose is >6.0. If this was a non-fasting sample, it may be difficult to assess. A fasting sample is preferred."
+            lifestyle = True
+        else:
+            output_phrase = ":large_green_circle: You likely do not have diabetes."
+    if lifestyle:
+        output_phrase += " Eat a healthy balanced diet - using My Healthy Plates (filling a quarter of the plate with wholegrains, quarter with good sources of protein, and half with fruit and vegetables). Maintain a healthy weight BMI ranging from 18.5 to 22.9 kg/m2. Exercise regularly, aiming for 150 minutes of moderate-intensity activity per week or 20 minutes of vigorous-intensity activity 3 or more days a week). Limit alcohol intake to 1 drink per day for females, and 2 drinks per day for males."
+    if inputattributes["smoker"]:
+        output_phrase += " You are highly encouraged to quit smoking."
+    return output_phrase 
+
+#print(f"advice {get_dm_advice(test_attributes, testdict)}")

image_loading.py

@@ -0,0 +1,36 @@
+import streamlit as st
+import pandas as pd
+import numpy as np
+import os
+import time
+import base64
+import json
+import cv2
+import re
+from openai import OpenAI
+from paddleocr import PaddleOCR
+
+def load_image():
+    uploaded_file = st.file_uploader(label='Upload your test results image below:')
+    if uploaded_file is not None:
+        image_data = uploaded_file.getvalue()
+        st.image(image_data, caption='', use_column_width=True)  # Adjust width for mobile screens
+        return image_data
+
+def remove_nric(text):
+    pattern = r'[STFG]\d{7}[A-Z]'
+    replaced_text = re.sub(pattern, ' ', text)
+    return replaced_text
+
+def extract_text(image,ocr_model):
+    ocr_start_time = time.time()
+    result = ocr_model.ocr(image)
+    result = result [0] #idk why this needs a result[0] instead of result for Github
+    extracted_text = ''
+    for idx in range(len(result)):
+        txt = result[idx][1][0]
+        extracted_text += str(txt) + " "
+    extracted_text_clean = remove_nric(extracted_text)
+    ocr_end_time = time.time()
+    ocr_time = int(ocr_end_time - ocr_start_time)
+    return extracted_text_clean, ocr_time