Python / invoices
Invoice PDFs in Python
Post your invoice HTML to a render API from an async endpoint that returns a streaming response, then stream the response with httpx.stream and yield chunks. No browser binary in your Python deployment.
Why invoices become PDFs
An invoice is a legal record of a demand for payment, and in most jurisdictions it has to be retained in a form that cannot change after issue. A page that re-renders from live data is not that. That is the difference between a document and a view of a database, and it is why finance teams and anyone billing customers keep asking for this.
What a invoice has to carry
Before any of the code below matters, the template has to produce a document that is actually a invoice. These are the fields that make it one, and the ones a reader or an auditor will look for first.
- an invoice number that is sequential and never reused
- the issue date, and separately the payment due date
- supplier and customer legal names with addresses, not trading names
- tax registration numbers for both parties where the jurisdiction requires them
- per-line quantity, unit price, tax rate and line total
- the tax subtotal broken out by rate, then the gross total
The Python implementation
httpx is the client most projects here already have. The render happens in an async endpoint that returns a streaming response, and you stream the response with httpx.stream and yield chunks.
python
import os, httpx
from fastapi import FastAPI
from fastapi.responses import StreamingResponse
app = FastAPI()
@app.get("/invoice/{invoice_id}.pdf")
async def invoice_pdf(invoice_id: str):
html = render_invoice(invoice_id)
client = httpx.AsyncClient(timeout=60.0)
request = client.build_request(
"POST",
"https://api.pdfpipe.xyz/v1/pdf",
headers={"Authorization": f"Bearer {os.environ['PDFPIPE_KEY']}"},
json={"html": html, "options": {"format": "A4", "printBackground": True}},
)
response = await client.send(request, stream=True)
# stream=True keeps a long invoice off the worker's heap.
return StreamingResponse(
response.aiter_bytes(),
media_type="application/pdf",
background=response.aclose,
)The CSS that makes a invoice page correctly
The layout problem specific to this document is that line items are unbounded, so the table has to break across pages while the totals block stays whole and the column headers repeat. These rules handle it.
css
/* Line items run past one page. Repeat the header, keep the
totals block whole, and never strand a single row. */
thead { display: table-header-group; }
tfoot { display: table-footer-group; }
tr { break-inside: avoid; }
.totals {
break-inside: avoid;
break-before: auto;
}
@page {
size: A4;
margin: 18mm 16mm 22mm;
}What goes wrong in Python
requests holds the entire body in memory before you touch it. httpx.stream does not, and on documents past a few megabytes that difference decides whether your worker survives a burst.
What people try first
Most Python projects reach for WeasyPrint, which is excellent at CSS Paged Media but needs Cairo, Pango and GDK-PixBuf present in the image, and diverges from a browser on modern layout. That works until it is running on more than one machine, at which point the browser becomes the thing you operate rather than the thing you use.
Where the invoice lives afterwards
Rendering is the short part. Invoices are kept for the statutory retention period, which in most jurisdictions runs to six years or more. That means the file has to be stored, not regenerated, and the stored copy has to be the one that was sent.
Getting the document right
- Check the total. A rounding difference between what you display and what you charge is the kind of error that surfaces during an audit rather than on the day.
- Generation is triggered by an order completing, or a billing cycle closing, so size the timeout for that path rather than for a health check.
- Volume arrives in bursts, so queue the render rather than doing it inside the request that triggered it.
- Backgrounds are painted by default here, so a design that uses colour needs nothing set. Only an explicit print_background of false turns them off.
Frequently asked
Do I need Chromium installed to generate invoices from Python?
No. The render happens over HTTP, so your Python deployment stays the size it is now. That is the main reason to use an API rather than WeasyPrint, which is excellent at CSS Paged Media but needs Cairo, Pango and GDK-PixBuf present in the image, and diverges from a browser on modern layout.
How do I stop a long invoice using all the memory?
Stream the response with httpx.stream and yield chunks. requests holds the entire body in memory before you touch it. httpx.stream does not, and on documents past a few megabytes that difference decides whether your worker survives a burst.
What has to be on a invoice?
At minimum: an invoice number that is sequential and never reused; the issue date, and separately the payment due date; supplier and customer legal names with addresses, not trading names. The one to get right before anything else is the total. A rounding difference between what you display and what you charge is the kind of error that surfaces during an audit rather than on the day.
Do I need to store the generated invoices?
Depends on the document, and this one has a clear answer: invoices are kept for the statutory retention period, which in most jurisdictions runs to six years or more. That means the file has to be stored, not regenerated, and the stored copy has to be the one that was sent.
Can I keep my existing invoice template?
Yes, if it produces HTML. Whatever renders your invoice view today can render the same markup for the PDF, which is why the CSS above is the only new thing you write.
Related
Other Python documents, and the same invoice in other stacks.
100 free documents a month, no card.