app.py

import gradio as gr
from PIL import Image
from datetime import datetime
import pytz
from ocr_engine import extract_weight, extract_unit_from_text
from simple_salesforce import Salesforce
import base64
import re
import os

# Salesforce credentials
SF_USERNAME = "Autoweightlogger@sathkrutha.com"
SF_PASSWORD = "autoweight@32"
SF_TOKEN = "UgiHKWT0aoZRX9gvTYDjAiRY"

# Connect to Salesforce
sf = Salesforce(username=SF_USERNAME, password=SF_PASSWORD, security_token=SF_TOKEN)

def restore_decimal(text):
    if re.fullmatch(r"\d{5}", text):
        return f"{text[:2]}.{text[2:]}"
    elif re.fullmatch(r"\d{4}", text):
        return f"{text[:2]}.{text[2:]}"
    return text

def process_image(image):
    if image is None:
        return "❌ No image provided", "", None, gr.update(visible=True)
    try:
        result = extract_weight(image)
        weight, raw_text = result if isinstance(result, tuple) else (result, "")

        print("🧠 Final OCR Result:", weight)
        print("🔤 OCR Raw Text:", raw_text)

        ist = pytz.timezone('Asia/Kolkata')
        timestamp = datetime.now(ist).strftime("%Y-%m-%d %H:%M:%S IST")

        if not weight or (isinstance(weight, str) and weight.startswith("Error")):
            return f"❌ OCR Error: {weight}", "", image, gr.update(visible=True)

        match = re.search(r'(\d{1,3}\.\d{1,3})\s*(kg|g)?', weight)
        if match:
            numeric_value = float(match.group(1))
            unit = match.group(2) if match.group(2) else extract_unit_from_text(raw_text) or "kg"
        else:
            cleaned = re.sub(r"[^\d]", "", weight)
            decimal_fixed = restore_decimal(cleaned)
            try:
                numeric_value = float(decimal_fixed)
                unit = extract_unit_from_text(raw_text) or "kg"
            except:
                return f"❌ Could not extract number | OCR: {weight}", "", image, gr.update(visible=True)

        image_path = "snapshot.jpg"
        image.save(image_path)

        record = sf.Weight_Log__c.create({
            "Captured_Weight__c": numeric_value,
            "Captured_Unit__c": unit,
            "Captured_At__c": datetime.now(ist).isoformat(),
            "Device_ID__c": "DEVICE-001",
            "Status__c": "Confirmed"
        })

        with open(image_path, "rb") as f:
            encoded_image = base64.b64encode(f.read()).decode("utf-8")

        content = sf.ContentVersion.create({
            "Title": f"Snapshot_{timestamp}",
            "PathOnClient": "snapshot.jpg",
            "VersionData": encoded_image
        })

        content_id = sf.query(
            f"SELECT ContentDocumentId FROM ContentVersion WHERE Id = '{content['id']}'"
        )['records'][0]['ContentDocumentId']

        sf.ContentDocumentLink.create({
            "ContentDocumentId": content_id,
            "LinkedEntityId": record['id'],
            "ShareType": "V",
            "Visibility": "AllUsers"
        })

        return f"{numeric_value} {unit}", timestamp, image, gr.update(visible=False)
    except Exception as e:
        return f"Error: {str(e)}", "", None, gr.update(visible=True)

with gr.Blocks(css=".gr-button {background-color: #2e7d32 !important; color: white !important;}" ) as demo:
    gr.Markdown("""
    <h1 style='text-align: center; color: #2e7d32;'>📷 Auto Weight Logger</h1>
    <p style='text-align: center;'>Upload or capture a digital weight image. Detects weight using AI OCR.</p>
    <hr style='border: 1px solid #ddd;' />
    """)

    with gr.Row():
        image_input = gr.Image(type="pil", label="📁 Upload or Capture Image")

    detect_btn = gr.Button("🚀 Detect Weight")

    with gr.Row():
        weight_out = gr.Textbox(label="📦 Detected Weight", placeholder="e.g., 75.5 kg", show_copy_button=True)
        time_out = gr.Textbox(label="🕒 Captured At (IST)", placeholder="e.g., 2025-07-01 12:00:00")

    snapshot = gr.Image(label="📸 Snapshot Preview")
    retake_btn = gr.Button("🔁 Retake / Try Again", visible=False)

    detect_btn.click(fn=process_image, inputs=image_input, outputs=[weight_out, time_out, snapshot, retake_btn])
    retake_btn.click(fn=lambda: ("", "", None, gr.update(visible=False)),
                     inputs=[], outputs=[weight_out, time_out, snapshot, retake_btn])

if __name__ == "__main__":
    demo.launch()