pesco-ncr/backend/app/services/pdf.py

"""Printable NCR PDF (WeasyPrint) — a clean single-document rendering of the
complete NCR for hard-copy travelers and audits.

WeasyPrint is imported lazily so environments without the Pango/Cairo system
libraries (e.g. unit tests) can still import the app.
"""
from functools import partial
from pathlib import Path

import anyio
from jinja2 import Environment, FileSystemLoader, select_autoescape

from app.domain import STAGE_LABELS, Stage
from app.models import Ncr
from app.models.base import utcnow
from app.services.storage import attachment_abs_path

_TEMPLATES_DIR = Path(__file__).resolve().parent.parent / "templates"

_env = Environment(
    loader=FileSystemLoader(_TEMPLATES_DIR),
    autoescape=select_autoescape(["html"]),
)


def _render_html(ncr: Ncr, related_ncr_numbers: list[str]) -> str:
    images = []
    other_files = []
    for att in ncr.attachments:
        entry = {
            "filename": att.original_filename,
            "uploaded_by": att.uploaded_by.display_name,
            "uploaded_at": att.uploaded_at,
            "size_kb": max(1, att.size_bytes // 1024),
        }
        path = attachment_abs_path(att.stored_path)
        if att.is_image and path.is_file():
            entry["src"] = path.as_uri()
            images.append(entry)
        else:
            other_files.append(entry)

    template = _env.get_template("ncr_pdf.html")
    return template.render(
        ncr=ncr,
        stage_label=STAGE_LABELS[Stage(ncr.stage)],
        stage_labels=STAGE_LABELS,
        Stage=Stage,
        images=images,
        other_files=other_files,
        related_ncr_numbers=related_ncr_numbers,
        generated_at=utcnow(),
    )


def _html_to_pdf(html: str) -> bytes:
    from weasyprint import HTML  # lazy: needs Pango/Cairo system libs

    return HTML(string=html).write_pdf()


async def render_ncr_pdf(ncr: Ncr, related_ncr_numbers: list[str] | None = None) -> bytes:
    html = _render_html(ncr, related_ncr_numbers or [])
    # WeasyPrint rendering is CPU-bound; keep it off the event loop.
    return await anyio.to_thread.run_sync(partial(_html_to_pdf, html))