Latest commit 1d7e763 Apr 13, 2017 @matthiask PDFDocument v3.2



This is a wrapper for ReportLab which allows easy creation of PDF documents:

from io import BytesIO
from pdfdocument.document import PDFDocument

def say_hello():
    f = BytesIO()
    pdf = PDFDocument(f)
    pdf.h1('Hello World')
    pdf.p('Creating PDFs made easy.')
    return f.getvalue()

Letters and reports

PDFDocument comes with two different PDF templates, letters and reports. The only difference is the layout of the first page: The letter has an additional frame for the address at the top and a smaller main content area.

Usage is as follows:

# Or:

The letter generates default styles using 9 point fonts as base size, the report uses 8 points. This can be changed by calling pdf.generate_style again.

There exists also a special type of report, the confidential report, the only differences being that the confidentiality is marked using a red cross at the top of the first page and a watermark in the background.


The call to pdf.generate_style generates a set of predefined styles. (Yes it does!) That includes the following styles; this list is neither exhaustive nor a promise:


Most of the time you will not use those attributes directly, except in the case of tables. Convenience methods exist for almost all styles as described in the next chapter.


All content passed to the following methods is escaped by default. ReportLab supports a HTML-like markup language, if you want to use it directly you'll have to either use only pdf.p_markup or resort to creating pdfdocument.document.MarkupParagraph instances by hand.


pdf.h1, pdf.h2, pdf.h3


pdf.p, pdf.p_markup, pdf.small, pdf.smaller

Unordered lists




Various elements, pdf.hr_mini, pdf.spacer, pdf.pagebreak, pdf.start_keeptogether, pdf.end_keeptogether, pdf.next_frame,


pdf.table, pdf.bottom_table

Canvas methods

Canvas methods work with the canvas directly, and not with Platypus objects. They are mostly useful inside stationery functions. You'll mostly use ReportLab's canvas methods directly, and only resort to the following methods for special cases.

pdf.confidential, pdf.draw_watermark, pdf.draw_svg

Additional methods

pdf.append, pdf.restart

Django integration

PDFDocument has a few helpers for generating PDFs in Django views, most notably pdfdocument.utils.pdf_response:

from pdfdocument.utils import pdf_response

def pdf_view(request):
    pdf, response = pdf_response('filename_without_extension')
    # ... more code

    return response

The SVG support uses svglib by Dinu Gherman. It can be found on PyPI: <>