app.py

import gradio as gr
import json
from reportlab.lib.pagesizes import letter
from reportlab.pdfgen import canvas
from reportlab.lib import colors
import tempfile

PAGE_WIDTH, PAGE_HEIGHT = letter

# -------------------
# PDF generation function
# -------------------
def generate_pdf(file):
    try:
        with open(file.name, "r", encoding="utf-8") as f:
            data = json.load(f)
    except Exception as e:
        return None, f"Error reading JSON: {e}"

    temp_pdf = tempfile.NamedTemporaryFile(delete=False, suffix=".pdf")
    pdf_path = temp_pdf.name
    temp_pdf.close()

    c = canvas.Canvas(pdf_path, pagesize=letter)
    page_created = False

    # Current vertical position
    current_y = PAGE_HEIGHT - 50  # start from top margin
    bottom_margin = 50

    for ad_size, ad_data in data.items():
        for item in ad_data:
            for line in item.get("text_lines", []):
                text = line.get("text", "")
                bbox = line.get("bbox")
                if not bbox:
                    continue
                page_created = True
                x1, y1, x2, y2 = bbox

                # Height of text box
                box_height = y2 - y1
                font_size = max(6, min(18, box_height * 0.8))

                # Check if the next line fits in current page, else start new page
                if current_y - font_size < bottom_margin:
                    c.showPage()
                    current_y = PAGE_HEIGHT - 50  # reset for new page

                pdf_x = x1
                pdf_y = current_y - box_height

                # Optional: draw bounding box
                c.setStrokeColor(colors.lightgrey)
                c.rect(pdf_x, pdf_y, x2 - x1, box_height, fill=0)

                if "<b>" in text:
                    c.setFont("Helvetica-Bold", font_size)
                    text = text.replace("<b>", "").replace("</b>", "")
                else:
                    c.setFont("Helvetica", font_size)

                c.setFillColor(colors.black)
                c.drawString(pdf_x, pdf_y + (box_height - font_size), text)

                current_y = pdf_y - 5  # move cursor down with small gap

    if not page_created:
        c.setFont("Helvetica", 12)
        c.drawString(50, 750, "No OCR text found")
        c.showPage()

    c.save()
    return pdf_path, "PDF generated successfully!"


# -------------------
# Gradio Interface
# -------------------
iface = gr.Interface(
    fn=generate_pdf,
    inputs=gr.File(file_types=[".json"], label="Upload JSON"),
    outputs=[gr.File(label="Download PDF"), gr.Textbox(label="Status")],
    title="JSON to PDF Converter",
    description="Upload a `results.json` file and download the generated PDF."
)

# Launch the app
iface.launch()