app.py

import openai
import gradio as gr
import os
import PyPDF2
from docx import Document
from PIL import Image
import pytesseract
import json

# Load your OpenAI API key and prompt from environment variables
openai.api_key = os.getenv("HF_API_KEY_OPENAI")
preset_prompt = os.getenv("HF_PROMPT_HEART_HEALTH")

# Risk Disclaimer to be added at the end of the report
risk_disclaimer = f'''
Risk Disclaimer:
This AI-generated summary report provided is intended solely for use as a decision support tool and should not be considered as medical diagnosis. These reports are designed to assist individuals and healthcare professionals and should be used in conjunction with clinical judgment and other diagnostic methods. The AI system does not replace the expertise of qualified medical practitioners. Users should not make any healthcare decisions based solely on AI-generated reports without consulting with a licensed healthcare provider.
'''

# Function to extract text from a PDF
def extract_text_from_pdf(pdf_file):
    reader = PyPDF2.PdfReader(pdf_file)
    text = ""
    for page in reader.pages:
        text += page.extract_text()
    return text

# Function to extract text from a DOCX file
def extract_text_from_docx(docx_file):
    doc = Document(docx_file)
    text = ""
    for paragraph in doc.paragraphs:
        text += paragraph.text
    return text

# Function to extract text from an image using OCR
def extract_text_from_image(image_file):
    image = Image.open(image_file)
    text = pytesseract.image_to_string(image)
    return text

# Function to detect file type and extract text accordingly
def extract_text(file):
    file_type = file.name.split('.')[-1].lower()
    if file_type == 'pdf':
        return extract_text_from_pdf(file)
    elif file_type == 'docx':
        return extract_text_from_docx(file)
    elif file_type in ['jpg', 'jpeg', 'png']:
        return extract_text_from_image(file)
    elif file_type == 'txt':
        return file.read().decode('utf-8')
    else:
        return "Unsupported file format."

# Function to interact with the GPT-4o-mini model via the OpenAI API
def heart_health_bot(file):
    if file is None:
        return "Please upload a PDF, DOCX or text file for analysis."

    context = extract_text(file)
    if "Unsupported file format" in context:
        return context
    
    # Use the secret prompt directly in the API call
    response = openai.chat.completions.create(
        model="gpt-4o-mini",
        messages=[
            {"role": "system", "content": preset_prompt},
            {"role": "user", "content": context}
        ],
        max_tokens=3000,  # Increased to capture the complete analysis
        temperature=0.7,
    )
    
    # Extract the response text and remove any leading/trailing whitespace
    result = response.choices[0].message.content.strip()

    # Append the risk disclaimer to the result
    result_with_disclaimer = f"{result}\n\n{risk_disclaimer}"

    # Create a JSON output
    output_json = {
        "result": result_with_disclaimer
    }
    
    # Save the JSON output to a file
    json_file = "heart_health_report.json"
    with open(json_file, "w") as f:
        json.dump(output_json, f)

    return result_with_disclaimer, json_file

# Create the Gradio interface
iface = gr.Interface(
    fn=heart_health_bot,
    inputs=[
        gr.File(label="Upload a PDF, DOCX, image, or text file for analysis")  # File upload input
    ],
    outputs=[
        "text",  # Output type for the bot's response
        gr.File(label="Download JSON Report")  # JSON download link
    ],
    title="AIWO 181 AI Report (https://aiwo.com/)",
    description=(
        "Upload your 181 bloodwork clinical data to get a recommendation for better health outcomes.\n\n"
        "Developed by Venkat Srinivasan.\n"
        "Email: venkat.srinivasan@gbs-alumni.com\n"
        "LinkedIn: [Venkataraghavan Srinivasan](https://www.linkedin.com/in/venkataraghavansrinivasan/)\n\n"
        "Upload a document (PDF, DOCX, image, or text) to receive a heart health report."
    )
)

# Launch the Gradio interface
iface.launch()