Skip to content

Edinburgh-Genome-Foundry/pdf_reports

Repository files navigation

PDF Reports Logo

GitHub CI build status https://coveralls.io/repos/github/Edinburgh-Genome-Foundry/pdf_reports/badge.svg?branch=master

PDF Reports (complete documentation here) is a Python library to create nice-looking PDF reports from HTML or Pug templates. It features modern-looking components (via the Semantic UI framework) and provides routines to embed tables or plots in the documents.

Note that only Python 3.x is officially supported, although with the right version of weasyprint the library can also run on 2.x.

Example of use

Your Pug template file template.pug may look like this (see a full example):

#sidebar I am the text in the sidebar.

h1 {{ title }}

.ui.piled.segment
  p Oh hi there ! I am some text in a cool box.

Your Python code will be as follows:

from pdf_reports import pug_to_html, write_report
html = pug_to_html("template.pug", title="My report")
write_report(html, "example.pdf")

And your final result may look like this (PDF file):

https://github.com/Edinburgh-Genome-Foundry/pdf_reports/raw/master/screenshot.png

See also this example embedding some python code in the template to create figures and tables on the flight.

Other features

Preloading CSS and SCSS

PDF Reports provides a preload_stylesheet method which can be used to load and parse a CSS file. It also works with SCSS files (which will automatically be compiled to CSS) but this requires libsass installed (for instance via pip install libsass). Here is an example:

from pdf_reports import pug_to_html, write_report, preload_stylesheet

css = preload_stylesheet('style.scss')
html = pug_to_html("template.pug", title="My report", my_name='Zulko')
write_report(html, "example.pdf", extra_stylesheets=[css])

Using a ReportWriter

The ReportWriter class allows to define default templates, styles, and variable names. It can be used to avoid repeating yourself across your application:

from pdf_reports import ReportWriter

# DEFINE A WRITER WITH DEFAULT TEMPLATE AND VALUES
report_writer = ReportWriter(
    default_stylesheets=["style.css"],
    default_template="template.pug",
    title="My default title",
    version="0.1.2"
)

# THEN LATER IN YOUR CODE:
html = report_writer.pug_to_html(my_name="Zulko", my_organization="EGF")
report_writer.write_report(html, "example_reportwriter.pdf")

Markdown support

As a feature of PyPugJS, markdown is supported in the Pug templates.

div
  :markdown
    This is some markdown text. Here is a [link](http://example.com/).

    - this is a bullet point list
    - Second item
    - Etc.

PDF tools

Some useful functions for generating reports are available from inside the Pug templates under pdf_tools. For instance, pdf_tools.figure_data() to embed matplotlib images, or pdf_tools.dataframe_to_html() to turn Pandas dataframes into HTML, and style them nicely with Semantic UI. Have a look at the docs, or this example.

JupyterPDF

The JupyterPDF class eases report templates writing by embedding PDF files in Jupyter notebooks (using the browser's interactive PDF viewer).

from pdf_reports import JupyterPDF

# Build your PDF

# At the end of the notebook cell:
JupyterPDF("path_to_your.pdf")

Notes

The core of the library consists of just a few lines of Python, using pypugjs to parse Pug templates, optionally including stylesheets from the Semantic UI CSS framework, and finally calling weasyprint for PDF generation. Please refer to the Weasyprint documentation for the customization of templates. For instance, to customize the page margins and numbering the Weasyprint way, add this to your SCSS code:

@page {
    margin: 1cm 0 2cm 0cm;
    @bottom-center {
        content: "Page " counter(page) " / " counter(pages);
        font-family: 'Lato';
    }
}

Using Semantic UI implies that (1) the Lato font family should be installed on your machine, otherwise the results will look less good, and (2) the first time that write_pdf is called in a Python session, if using the default Semantic UI style, the parsing of the CSS will add a 3-second overhead to the function calls (but there will be no overhead for the next calls in that session).

Installation

You can install the library via PIP:

pip install pdf_reports

Alternatively, you can unzip the sources in a folder and type:

python setup.py install

Note: the package depends on the WeasyPrint Python package. If there are any issues, see installation instructions in the WeasyPrint documentation.

If Pango is not available in your conda environment (which pango-view), but otherwise it's installed, then install it with conda install anaconda::pango

PDF Reports has been tested on Ubuntu 22.04. If you have an older GNU/Linux distribution (e.g. Ubuntu 18.04) which doesn't have the latest Pango that is required by the latest WeasyPrint, then installing an older WeasyPrint (<=52) may help: pip install weasyprint==52

Note: on some Debian systems you may need to first install libffi-dev (apt install libffi-dev). The package name may be libffi-devel on some systems.

Note: on macOS, you may need to first install pango with: brew install pango

License = MIT

This open-source software project was originally written at the Edinburgh Genome Foundry by Zulko and released on Github under the MIT licence (Copyright 2018 Edinburgh Genome Foundry). Everyone is welcome to contribute !