Update app.py

app.py

@@ -1,83 +1,43 @@
-import fitz  # PyMuPDF
-import re
 import gradio as gr
-import pandas as pd
-import matplotlib.pyplot as plt
-import io
-
-# Normal reference ranges
-normal_ranges = {
-    "Hemoglobin": (13.5, 17.5),  # g/dL
-    "WBC": (4000, 11000),        # /µL
-    "Platelets": (150000, 450000),  # /µL
-    "RBC": (4.7, 6.1),           # million/µL
-    "Creatinine": (0.6, 1.3),    # mg/dL
-    "Glucose": (70, 99),         # mg/dL (fasting)
-}
-
-def extract_text(file):
-    try:
-        doc = fitz.open(stream=file.read(), filetype="pdf")
-        text = ""
-        for page in doc:
-            text += page.get_text()
-        return text
-    except Exception as e:
-        return f"❌ Error: {str(e)}"
-
-def analyze_report(file):
-    text = extract_text(file)
-    if "❌ Error" in text:
-        return text, None
-
-    results = {}
-    for test, (low, high) in normal_ranges.items():
-        pattern = rf"{test}[^0-9\-]*([\d\.]+)"
-        match = re.search(pattern, text, re.IGNORECASE)
-        if match:
-            value = float(match.group(1))
-            status = "Normal" if low <= value <= high else "Abnormal"
-            results[test] = {"value": value, "low": low, "high": high, "status": status}
-
-    if not results:
-        return "No known lab values found in report.", None
+import fitz  # PyMuPDF
+from transformers import pipeline
 
-    # Create dataframe
-    df = pd.DataFrame([
-        {
-            "Test": test,
-            "Value": data["value"],
-            "Normal Range": f'{data["low"]} - {data["high"]}',
-            "Status": data["status"]
-        }
-        for test, data in results.items()
-    ])
+# Load summarization model
+summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
 
-    # Plot chart
-    fig, ax = plt.subplots(figsize=(6, 4))
-    colors = ["green" if row["Status"] == "Normal" else "red" for _, row in df.iterrows()]
-    ax.bar(df["Test"], df["Value"], color=colors)
-    ax.set_title("Medical Report Values")
-    ax.set_ylabel("Measured Value")
-    ax.tick_params(axis='x', rotation=45)
+def extract_text_from_pdf(pdf_file):
+    doc = fitz.open(stream=pdf_file.read(), filetype="pdf")
+    text = ""
+    for page in doc:
+        text += page.get_text()
+    return text
 
-    buf = io.BytesIO()
-    plt.tight_layout()
-    plt.savefig(buf, format="png")
-    buf.seek(0)
+def simplify_summary(summary):
+    # Convert technical language to a more friendly explanation
+    return "🩺 Here's what the report says in simple words:\n\n" + summary.replace("\n", " ")
 
-    return df, buf
+def process_report(pdf_file):
+    try:
+        text = extract_text_from_pdf(pdf_file)
+        if len(text.strip()) == 0:
+            return "❌ Couldn't extract text from the PDF.", ""
+        
+        summary = summarizer(text, max_length=300, min_length=60, do_sample=False)[0]["summary_text"]
+        explanation = simplify_summary(summary)
+        return summary, explanation
+    except Exception as e:
+        return f"❌ Error: {str(e)}", ""
 
+# Gradio Interface
 demo = gr.Interface(
-    fn=analyze_report,
-    inputs=gr.File(label="Upload Medical PDF Report"),
+    fn=process_report,
+    inputs=gr.File(label="Upload your Medical Report PDF"),
     outputs=[
-        gr.Dataframe(label="Extracted Values"),
-        gr.Image(type="pil", label="Visualization")
+        gr.Textbox(label="Summarized Report", lines=10),
+        gr.Textbox(label="Explanation in Simple Terms", lines=10)
     ],
-    title="🩺 Medical Report Analyzer",
-    description="Upload a medical lab report (PDF). Extracted lab values will be compared with normal reference ranges. Abnormal values are flagged and visualized in red."
+    title="📋 Medical Report Analyzer",
+    description="Upload a medical report and get a simplified summary using GPT-powered summarization.",
 )
 
-if __name__ == "__main__":
-    demo.launch()
+demo.launch()