up

app.py

@@ -1,12 +1,9 @@
 import streamlit as st
 import fitz  # PyMuPDF
-from transformers import pipeline
 from gtts import gTTS
 import tempfile
 import os
-import qrcode
-from PIL import Image
-import io
+from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline
 
 # Streamlit page setup
 st.set_page_config(page_title="Health Report Analyzer", page_icon="🩺")
@@ -23,10 +20,36 @@ def extract_text(file):
         text += page.get_text()
     return text
 
-# Load Hugging Face model for medical explanation
+# Load Medical Domain Model (Bio_ClinicalBERT)
 @st.cache_resource
-def load_explainer():
-    return pipeline("text2text-generation", model="google/flan-t5-large")
+def load_medical_explainer():
+    tokenizer = AutoTokenizer.from_pretrained("emilyalsentzer/Bio_ClinicalBERT")
+    model = AutoModelForSequenceClassification.from_pretrained("emilyalsentzer/Bio_ClinicalBERT")
+    return pipeline("text-classification", model=model, tokenizer=tokenizer)
+
+# Doctor recommendation logic
+def suggest_doctor(term):
+    term = term.lower()
+    if "kidney" in term or "creatinine" in term:
+        return "You should consider seeing a **nephrologist**."
+    elif "hemoglobin" in term:
+        return "You might want to consult a **hematologist** if your hemoglobin is outside the normal range."
+    elif "liver" in term or "bilirubin" in term:
+        return "Consult a **hepatologist** for liver-related concerns."
+    elif "sugar" in term or "glucose" in term:
+        return "A **diabetologist** or **endocrinologist** is recommended for sugar level issues."
+    return "Please consult a **general physician** for further evaluation."
+
+# Normal ranges (very basic examples)
+def get_reference_range(term):
+    term = term.lower()
+    if "hemoglobin" in term:
+        return "Normal range: 13.8 to 17.2 g/dL for men, 12.1 to 15.1 g/dL for women."
+    elif "creatinine" in term:
+        return "Normal range: 0.74 to 1.35 mg/dL for men, 0.59 to 1.04 mg/dL for women."
+    elif "bilirubin" in term:
+        return "Normal range: 0.1 to 1.2 mg/dL."
+    return "Range data not available for this term."
 
 # Main logic
 if uploaded_file:
@@ -37,52 +60,32 @@ if uploaded_file:
     st.markdown("### 📄 Health Report Content")
     st.write(text_data)
 
-    explainer = load_explainer()
-
-    # Store extracted text in session
+    explainer = load_medical_explainer()
     st.session_state['report_text'] = text_data
 
-    # Chatbot
     st.subheader("💬 Ask About Any Medical Term or Part of the Report")
 
     user_question = st.text_input("Enter a medical term or question (e.g. 'CT scan', 'Explain creatinine'):")
 
     if st.button("Get AI Explanation") and user_question:
         with st.spinner("Thinking..."):
-
-            # Enhanced prompt for better answers
-            prompt = (
-                f"You are a friendly and experienced medical assistant. "
-                f"Explain this term in very simple language. "
-                f"Provide a short definition and mention why it's important or what it means in a report.\n\n"
-                f"Question: {user_question}"
-            )
-
-            response = explainer(prompt, max_length=300)[0]['generated_text']
+            explanation = f"{user_question.capitalize()} is a common term in medical reports."
+            range_info = get_reference_range(user_question)
+            suggestion = suggest_doctor(user_question)
 
         st.success("Explanation:")
-        st.write(response)
+        st.write(explanation)
+
+        st.info(f"📏 Reference Range: {range_info}")
+        st.warning(f"👨‍⚕️ Doctor Suggestion: {suggestion}")
 
         # Text-to-speech using gTTS
-        tts = gTTS(text=response)
+        tts = gTTS(text=f"{explanation}. {range_info}. {suggestion}")
         temp_audio = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3")
         tts.save(temp_audio.name)
 
-        audio_file = open(temp_audio.name, 'rb')
-        audio_bytes = audio_file.read()
-        st.audio(audio_bytes, format='audio/mp3')
-
-        # Generate QR Code for the explanation text
-        if st.button("📱 Generate QR Code to Share Explanation"):
-            qr = qrcode.QRCode(version=1, box_size=10, border=5)
-            qr.add_data(response[:300])  # Keep text short for QR
-            qr.make(fit=True)
-
-            img = qr.make_image(fill='black', back_color='white')
-
-            buf = io.BytesIO()
-            img.save(buf)
-            st.image(buf.getvalue(), caption="Scan to Read Explanation", use_column_width=False)
+        with open(temp_audio.name, 'rb') as audio_file:
+            st.audio(audio_file.read(), format='audio/mp3')
 
 else:
     st.info("Upload a PDF Health Report to begin.")