Pydfy: Turn DataFrames into PDF Reports

Pydfy makes creating beautiful PDF reports in Python as simple as you hope it to be. In practice, this is often not the case as:

• Not the whole team knows LaTeX
• The support for this in BI tooling is lacking
• Webpages are usually not tuned for PDF exports
• PDF builders are too complex to get something out the door by the end of the week

Hence, pydfy boils it down to the data and some text, with just enough CSS to beat matplotlib in terms of design. Fortunately, that bar was not high.

Installation

First, make sure tailwindcss is in your PATH, which for Linux means:

curl -sLO https://github.com/tailwindlabs/tailwindcss/releases/latest/download/tailwindcss-linux-x64 chmod +x tailwindcss-linux-x64 sudo mv tailwindcss-linux-x64 /usr/local/bin/tailwindcss

Check https://github.com/tailwindlabs/tailwindcss/releases/latest to find the matching binary for your system.

Also make sure to install Chromium:

apt-get update && apt-get install -y chromium-driver  # Example using apt-get brew install --cask chromium                          # Homebrew example for macOS

Now run:

pip install pydfy

And you should be good to go. Optionally install the extra dependencies as well:

pip install "pydfy[pandas,matplotlib]"

Usage

Example

import pydfy.models as pf
import pandas as pd
salesdf = pd.readcsv("sales.csv") orderslastweek = salesdf[salesdf["week"] == 10] avgprice = salesdf["price"].mean()
pdf = pf.PDF(     pf.Title(text="Sales Report"),     [pf.Number(orderslastweek, title="Orders last week"), pf.Number(avg_price, title="Average price")],     [pf.Table(salesdf, "Sales table"), salesdf.plot("hist")], ) pdf.render()

The main entrypoint is defined, in this context, by the pf.PDF function, which accepts the rows of the PDF as arguments and wraps a pf._PDF model for ease of use. A row here can either be a single component or an iterable of components, as shown in the example above.

Overview of Components

Component(templatepath=..., colspan=...)  # Abstract component, args apply to all other components
Table(rows=..., title=..., headers=...) Number(number=..., tiele=...)
Image(path=...) Figure(figure=...)
Title(text=...) Section(text=...) Paragraph(content=...) PageBreak(...)

Output location

By default, this renders a PDF file called out.pdf in the current working directory. To control the location of the build files, you can do:

# Change the name of the rendered report and store the intermediate html and css files in the /build directory
pdf.render(out="report.pdf", build_dir="/build")
Save everything in the /data directory
pdf.render(out="/data/report.pdf", build_dir="/data")
Only save the pdf in the /data directory
os.environ["PYDFY_OUT"] = "/data/report.pdf"
pdf.render()

Custom Components

You can create custom components with custom html templates as follows:

from dataclasses import dataclass, field
import pydfy.models as pf
@dataclass class TwoNumbers(pf.Component):     number_1: int     number_2: int     title: str
templatepath: str = field(default="twonumbers.html", kw_only=True)

Where the contents of two_numbers.html is:

<div class="pf-two-numbers border-solid border border-gray-100">     <div class="pf-two-numbers-title bg-gray-200 text-lg font-bold text-center align-top">{{ component.title }}</div>     <div class="pf-two-numbers-content mt-2 text-center">         <div class="grid grid-cols-2">             <div class="text-3xl mb-2">{{ "{:,}".format(component.number_1) }}</div>             <div class="text-3xl mb-2">{{ "{:,}".format(component.number_2) }}</div>         </div>     </div> </div>

To see a working example, check out examples/custom.

Custom CSS

You can add custom CSS to the compilation step:

pdf.render(css="blue.css")

Content of my.css, where we need to copy the contents of base.css first:

/ Tailwind setup / @tailwind base;  @tailwind components; @tailwind utilities; @media print { body { -webkit-print-color-adjust: exact; } } @media print {  html, body { width: 210mm; height: 297mm; font-size: 8pt; } } @media print { thead {display: table-row-group;} } @page { size: A4; margin: 1cm; }

.pf-number-content {     color: blue; }

To see a working example, check out examples/iris.

Using Docker

You could use Docker in the development process as follows:

docker build -t pydfy . cd examples/iris docker run -v $PWD:/pydfy pydfy > out.pdf  # The docker image is configured to run main.py and write to stdout

Contributing

poetry install --with dev --all-extras