app.py

import gradio as gr
import torch
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch.nn.functional as F

# Mock ICD and CPT data (replace with actual API calls or datasets)
def fetch_icd_codes(query):
    # Mock ICD codes for demonstration
    return [
        {"code": "R50.9", "description": "Fever, unspecified"},
        {"code": "A00", "description": "Cholera"},
        {"code": "J06.9", "description": "Acute upper respiratory infection, unspecified"}
    ]

def fetch_cpt_codes(query):
    # Mock CPT codes for demonstration
    return [
        {"code": "99213", "description": "Office or other outpatient visit"},
        {"code": "87804", "description": "Infectious agent detection by immunoassay"},
        {"code": "85025", "description": "Complete blood count (CBC)"}
    ]

# Load tokenizer and model
tokenizer = AutoTokenizer.from_pretrained("emilyalsentzer/Bio_ClinicalBERT")
model = AutoModelForSequenceClassification.from_pretrained("emilyalsentzer/Bio_ClinicalBERT", num_labels=1000)  # Adjust num_labels as needed

# Prediction function
def predict_codes(text):
    if not text.strip():
        return "Please enter a medical summary."
    
    # Tokenize input
    inputs = tokenizer(
        text,
        return_tensors="pt",
        max_length=512,
        truncation=True,
        padding=True
    )
    
    # Get predictions
    model.eval()
    with torch.no_grad():
        outputs = model(**inputs)
        logits = outputs.logits
    
    # Get probabilities
    probs = F.softmax(logits, dim=1)
    
    # Get top 3 predictions
    top_k = torch.topk(probs, k=3)
    
    # Fetch ICD and CPT codes using mock functions
    icd_results = fetch_icd_codes(text)
    cpt_results = fetch_cpt_codes(text)
    
    # Format results
    result = "Recommended ICD-10 Codes:\n"
    for i, code in enumerate(icd_results[:3]):  # Show top 3 ICD codes
        result += f"{i+1}. {code.get('code', 'Unknown')}: {code.get('description', 'No description')}\n"
    
    result += "\nRecommended CPT Codes:\n"
    for i, code in enumerate(cpt_results[:3]):  # Show top 3 CPT codes
        result += f"{i+1}. {code.get('code', 'Unknown')}: {code.get('description', 'No description')}\n"
    
    return result

# Create Gradio interface
iface = gr.Interface(
    fn=predict_codes,
    inputs=gr.Textbox(
        lines=5,
        placeholder="Enter medical summary here...",
        label="Medical Summary"
    ),
    outputs=gr.Textbox(
        label="Predicted Codes",
        lines=10
    ),
    title="AutoRCM - Medical Code Predictor",
    description="Enter a medical summary to get recommended ICD-10 and CPT codes.",
    examples=[
        ["Patient presents with blood pressure 150/90. Complains of occasional headaches. History of hypertension."],
        ["Patient has elevated blood sugar levels. A1C is 7.8. History of type 2 diabetes."],
        ["Patient complains of chronic lower back pain, worse with movement. No radiation to legs."]
    ],
    allow_flagging="never"  # Disable caching
)

# Launch the interface
iface.launch(share=True)