Update app.py

app.py

@@ -1,45 +1,85 @@
 import streamlit as st
-from PyPDF2 import PdfReader
-from transformers import pipeline
 from PIL import Image
+import google.generativeai as genai
 
-# Load Hugging Face model
-model = pipeline("text-generation", model="Ketak-ZoomRx/Drug_Prompt_Ollama_67k",torch_dtype="auto", trust_remote_code=True, use_fast=True, device_map={"": "cuda:0"},)
+# Configure Google Generative AI
+genai_api_key = "AIzaSyCOEqA_IZlpWCHhMOGaDJ3iJjl5cRmzKgQ"
+genai.configure(api_key=genai_api_key)
 
-def extract_text_from_pdf(pdf_file):
-    reader = PdfReader(pdf_file)
-    text = ""
-    for page in reader.pages:
-        text += page.extract_text()
-    return text
+# Initialize Gemini model
+@st.cache_resource
+def load_gemini_model():
+    model = genai.GenerativeModel("gemini-1.5-flash")
+    return model
 
-def get_lab_report_interpretation(text):
-    result = model(text)
-    return result[0]['label']
+# Function to extract text from the image using Gemini model
+def extract_text_from_image(uploaded_file, model):
+    # Open the uploaded file as a PIL image
+    image = Image.open(uploaded_file).convert("RGB")
+    
+    # Generate content using the Gemini model with the image
+    response = model.generate_content(["Extract text from this medical report:", image])
+    extracted_text = response.text.strip()
+    
+    return extracted_text
 
-def process_image(image_file):
-    img = Image.open(image_file)
-    return img
+# Function to interpret the extracted text in layman's language
+def interpret_medical_report(extracted_text, model):
+    # Provide interpretation in layman's terms
+    prompt = (
+        f"The following is a medical report text:\n\n"
+        f"{extracted_text}\n\n"
+        "Please interpret this report for 7th grader and non native english speaker, "
+        "explaining the main findings in as short as possible"
+    )
+    response = model.generate_content([prompt])
+    interpretation = response.text.strip()
+    
+    return interpretation
 
-# Streamlit UI
-st.title("Medical Lab Report Analyzer")
+# Function to provide recommendations based on the extracted text
+def provide_recommendations(extracted_text, model):
+    # Provide recommendations
+    prompt = (
+        f"Based on the medical report text below:\n\n"
+        f"{extracted_text}\n\n"
+        "What recommendations would you give to the patient for managing their health?"
+        "Provide brief suggestions that are easy to understand for someone without medical knowledge."
+    )
+    response = model.generate_content([prompt])
+    recommendations = response.text.strip()
+    
+    return recommendations
 
-# Upload image and PDF
-uploaded_pdf = st.file_uploader("Upload your PDF Medical Lab Report", type="pdf")
-uploaded_image = st.file_uploader("Upload Image of Medical Report", type="jpg png jpeg")
+# Streamlit UI for the web app
+def main():
+    st.title("Medical Report Analyzer Using Gemini Model")
+    st.write("Upload an image of a medical report")
 
-if uploaded_pdf is not None:
-    st.subheader("Extracted Text from PDF:")
-    pdf_text = extract_text_from_pdf(uploaded_pdf)
-    st.write(pdf_text)
-    
-    st.subheader("Lab Report Interpretation:")
-    interpretation = get_lab_report_interpretation(pdf_text)
-    st.write(interpretation)
+    # Load the Gemini model
+    model = load_gemini_model()
 
-elif uploaded_image is not None:
-    st.image(process_image(uploaded_image), caption="Uploaded Image", use_column_width=True)
-    st.warning("For text extraction from image, please upload a PDF.")
-    
-else:
-    st.info("Please upload a medical PDF or image for analysis.")
+    # File uploader for medical report image
+    uploaded_image = st.file_uploader("Upload Medical Report Image", type=["png", "jpg", "jpeg"])
+
+    if uploaded_image is not None:
+        image = Image.open(uploaded_image).convert("RGB")
+        st.image(image, caption="Uploaded Medical Report Image", use_container_width=True)
+
+        if st.button("Analyze Report"):
+            with st.spinner("Processing image and analyzing report..."):
+                # Extract text from image
+                extracted_text = extract_text_from_image(uploaded_image, model)
+
+                # Interpret the extracted text
+                st.subheader("Interpretation in Layman's Language:")
+                interpretation = interpret_medical_report(extracted_text, model)
+                st.text(interpretation)
+
+                # Provide health recommendations
+                st.subheader("Health Recommendations:")
+                recommendations = provide_recommendations(extracted_text, model)
+                st.text(recommendations)
+
+if __name__ == "__main__":
+    main()