app.py

from transformers import BlipProcessor, BlipForConditionalGeneration
from PIL import Image
import gradio as gr
import torch
from datetime import datetime
from fpdf import FPDF
from simple_salesforce import Salesforce
import requests
import json
import os
from dotenv import load_dotenv


# Load environment variables from .env file
load_dotenv()

# Salesforce credentials
SF_USERNAME = os.getenv('SF_USERNAME')
SF_PASSWORD = os.getenv('SF_PASSWORD')
SF_SECURITY_TOKEN = os.getenv('SF_SECURITY_TOKEN')

# Initialize Salesforce connection
sf = Salesforce(username=SF_USERNAME, password=SF_PASSWORD, security_token=SF_SECURITY_TOKEN)

# Load BLIP model and processor
processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-base")
model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-base")
model.eval()
device = "cuda" if torch.cuda.is_available() else "cpu"
model.to(device)

# Inference function to generate captions dynamically based on image content
def generate_captions_from_image(image):
    if image.mode != "RGB":
        image = image.convert("RGB")
    
    # Preprocess the image and generate a caption
    inputs = processor(image, return_tensors="pt").to(device, torch.float16)
    output = model.generate(**inputs, max_new_tokens=50)
    caption = processor.decode(output[0], skip_special_tokens=True)
    
    return caption

# Function to create PDF and upload to Salesforce
def create_and_upload_pdf(dpr_content):
    # Create PDF instance using FPDF
    pdf = FPDF()
    pdf.set_auto_page_break(auto=True, margin=15)
    pdf.add_page()

    # Set title and content for the PDF
    pdf.set_font("Arial", size=12)
    pdf.cell(200, 10, txt="Daily Progress Report", ln=True, align='C')
    pdf.ln(10)  # Add space between lines

    # Add the content of the DPR text to the PDF
    pdf.multi_cell(0, 10, dpr_content)

    # Save PDF to a file (temporary storage)
    pdf_output_path = "/tmp/dpr_report.pdf"
    pdf.output(pdf_output_path)

    # Upload PDF to Salesforce as ContentVersion
    with open(pdf_output_path, 'rb') as pdf_file:
        pdf_data = pdf_file.read()

    # Prepare request to Salesforce
    content_version_payload = {
        "Title": "Daily Progress Report",
        "PathOnClient": "DPR_Report.pdf",
        "VersionData": pdf_data
    }

    content_version_url = f"{sf.base_url}/services/data/v50.0/sobjects/ContentVersion/"
    headers = {
        "Authorization": f"Bearer {sf.session_id}",
        "Content-Type": "application/json"
    }

    response = requests.post(content_version_url, headers=headers, files={"VersionData": pdf_data}, data=json.dumps(content_version_payload))

    if response.status_code == 201:
        content_version = response.json()
        pdf_url = f"/sfc/servlet.shepherd/version/download/{content_version['Id']}"
        return pdf_url
    else:
        raise Exception("Error uploading PDF to Salesforce: " + response.text)

# Function to generate the daily progress report (DPR) text
def generate_dpr(files):
    dpr_text = []
    current_time = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
    
    # Add header to the DPR
    dpr_text.append(f"Daily Progress Report\nGenerated on: {current_time}\n")
    
    # Process each uploaded file (image)
    for file in files:
        # Open the image from file path
        image = Image.open(file.name)  # Using file.name for filepath
        
        if image.mode != "RGB":
            image = image.convert("RGB")
        
        # Dynamically generate a caption based on the image
        caption = generate_captions_from_image(image)
        
        # Generate DPR section for this image with dynamic caption
        dpr_section = f"\nImage: {file.name}\nDescription: {caption}\n"
        dpr_text.append(dpr_section)

    # Join the DPR sections to form the complete report
    dpr_content = "\n".join(dpr_text)

    # Create and upload the PDF for this report in Salesforce
    pdf_url = create_and_upload_pdf(dpr_content)

    # Create a new Salesforce record for this report
    new_report_data = {
        "Report_Date__c": current_time,  # Set Report Date to current time
        "Detected_Activities__c": dpr_content,  # Set the activities field to the generated text
        "Photo_Uploads__c": ", ".join([file.name for file in files]),  # Upload file names as Photo Uploads
        "PDF_URL__c": pdf_url  # Set the PDF URL field with the generated PDF link
    }

    # Insert new Daily Progress Report record into Salesforce
    new_report = sf.Daily_Progress_Reports__c.create(new_report_data)

    return f"New report created in Salesforce with PDF URL: {pdf_url}"

# Gradio interface for uploading multiple files and displaying the text-based DPR
iface = gr.Interface(
    fn=generate_dpr,
    inputs=gr.Files(type="filepath", label="Upload Site Photos"),  # Handle batch upload of images
    outputs="text",  # Display the DPR as text in the output section
    title="Daily Progress Report Generator",
    description="Upload up to 10 site photos. The AI model will dynamically detect construction activities, materials, and progress and generate a text-based Daily Progress Report (DPR).",
    allow_flagging="never"  # Optional: Disable flagging
)

iface.launch()