Update app.py

app.py

@@ -1,75 +1,78 @@
-import streamlit as st
 import easyocr
-import cv2
-import numpy as np
-from PIL import Image
-import io
+import requests
+from transformers import AutoModelForTokenClassification, AutoTokenizer
 from transformers import pipeline
 
-# Initialize OCR reader
+# Initialize EasyOCR for text extraction from medical report image
 reader = easyocr.Reader(['en'])
 
-# Initialize Hugging Face text classifier (use a pre-trained or custom model)
-classifier = pipeline("text-classification", model="distilbert-base-uncased")
-
-# Function to check if the image is clear
-def is_image_clear(image):
-    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
-    laplacian_var = cv2.Laplacian(gray, cv2.CV_64F).var()
-    return laplacian_var > 100  # Threshold to determine clarity
-
-# Function to extract text using OCR
-def extract_text_from_image(image):
-    result = reader.readtext(image)
-    text = " ".join([res[1] for res in result])
+# Function to extract text from image
+def extract_text_from_image(image_path):
+    result = reader.readtext(image_path)
+    text = ' '.join([item[1] for item in result])
     return text
 
-# Function to analyze the text for abnormalities
-def analyze_report(text):
-    result = classifier(text)
-    severity = "Normal"
-    explanation = "No abnormalities detected"
+# Load Med7 model for medical NER (Named Entity Recognition)
+model_name = "jeff1evesque/med7"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForTokenClassification.from_pretrained(model_name)
 
-    # Example logic to determine severity (you can expand this)
-    if 'elevated' in text or 'high' in text:
-        severity = "Moderate"
-        explanation = "Elevated levels detected (e.g., glucose, blood pressure)"
-    elif 'critical' in text:
-        severity = "Severe"
-        explanation = "Critical levels detected"
-    return severity, explanation
+# Initialize the NER pipeline with Med7 model
+nlp = pipeline("ner", model=model, tokenizer=tokenizer)
 
-# Streamlit UI
-st.title("Medical Report Analysis Chatbot")
+# Function to extract medical entities (test results) from text
+def extract_medical_values(text):
+    entities = nlp(text)
+    medical_data = {}
+    
+    for entity in entities:
+        if entity['entity_group'] == 'LAB_RESULT':  # Change based on Med7's NER labels
+            medical_data[entity['word']] = entity['score']
+    
+    return medical_data
 
-uploaded_file = st.file_uploader("Upload your medical report image", type=["jpg", "jpeg", "png"])
+# Function to check the values with LabTestAPI or similar API
+def validate_medical_value(test_name, test_value):
+    # Example of API call (replace with actual API)
+    api_url = "https://api.labtestapi.com/get_reference_range"
+    params = {
+        "test_name": test_name,
+        "value": test_value
+    }
+    response = requests.get(api_url, params=params)
+    if response.status_code == 200:
+        result = response.json()
+        return result['normal_range']
+    else:
+        return None
 
-if uploaded_file:
-    image = Image.open(uploaded_file)
-    st.image(image, caption="Uploaded Medical Report", use_column_width=True)
+# Function to analyze medical report
+def analyze_report(image_path):
+    # Step 1: Extract text from the medical report image
+    text = extract_text_from_image(image_path)
     
-    # Convert to OpenCV format for clarity check
-    open_cv_image = np.array(image)
-    open_cv_image = open_cv_image[:, :, ::-1].copy()  # Convert RGB to BGR
+    # Step 2: Extract medical values using Med7
+    medical_values = extract_medical_values(text)
     
-    if is_image_clear(open_cv_image):
-        st.success("Image clarity: Clear")
-        
-        # Extract text
-        text = extract_text_from_image(open_cv_image)
-        st.write("Extracted Text: ")
-        st.write(text)
-        
-        # Analyze report
-        severity, explanation = analyze_report(text)
-        st.write(f"Report Type: Medical Report")
-        st.write(f"Severity Level: {severity}")
-        st.write(f"Explanation: {explanation}")
-        
-        # Ask for doctor consultation
-        consultation = st.radio("Would you like to consult a doctor?", ("Yes", "No"))
-        if consultation == "Yes":
-            st.write("Consultation fee: $50")
-            st.write("Your report will be sent to the doctor for further diagnosis.")
-    else:
-        st.warning("Image clarity: Blurry. Please upload a clearer image.")
+    analysis_results = []
+    
+    for test_name, test_value in medical_values.items():
+        # Step 3: Validate test value against normal range
+        normal_range = validate_medical_value(test_name, test_value)
+        if normal_range:
+            if test_value < normal_range[0]:
+                analysis_results.append(f"{test_name} is low. Consider consulting a doctor.")
+            elif test_value > normal_range[1]:
+                analysis_results.append(f"{test_name} is high. Consult a doctor.")
+            else:
+                analysis_results.append(f"{test_name} is within the normal range.")
+        else:
+            analysis_results.append(f"Could not validate {test_name}.")
+    
+    return analysis_results
+
+# Example Usage
+image_path = "path_to_medical_report_image.png"
+analysis = analyze_report(image_path)
+for result in analysis:
+    print(result)