Creating PDF Reports using Python

In the data science world, sharing results in an organized and visually appealing format is crucial for effective communication. Often, we turn to PDF reports to achieve this goal. As Python is one of the most popular languages for data science, there are numerous open-source libraries to create these reports. In this blog post, we will explore some of the most notable libraries and compare their strengths and weaknesses. We will also provide examples using Plotly and Pandas to illustrate how each library performs in a typical data science scenario.

ReportLab

ReportLab is a well-established library for creating PDF documents in Python. Its primary focus is on generating dynamic, data-driven documents with precise control over layout and formatting. ReportLab offers a wide range of features, including support for vector graphics, images, and tables.


import pandas as pd
import plotly.express as px
from reportlab.lib.pagesizes import letter
from reportlab.lib import colors
from reportlab.platypus import SimpleDocTemplate, Table, TableStyle

# Create a sample DataFrame
data = pd.DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})

# Create a Plotly figure
fig = px.line(data, x=data.index, y='A')

# Save the figure as an image
fig.write_image("figure.png")

# Create a PDF document
doc = SimpleDocTemplate("report.pdf", pagesize=letter)

# Create a table from the DataFrame
table = Table(data.values.tolist())

# Format the table
table.setStyle(TableStyle([
    ('BACKGROUND', (0, 0), (-1, 0), colors.grey),
    ('TEXTCOLOR', (0, 0), (-1, 0), colors.whitesmoke),
    ('ALIGN', (0, 0), (-1, -1), 'CENTER'),
    ('FONTNAME', (0, 0), (-1, 0), 'Helvetica-Bold'),
    ('FONTSIZE', (0, 0), (-1, 0), 14),
    ('BOTTOMPADDING', (0, 0), (-1, 0), 12),
    ('BACKGROUND', (0, 1), (-1, -1), colors.beige),
    ('GRID', (0, 0), (-1, -1), 1, colors.black)
]))

# Add the table to the document
doc.build([table])

WeasyPrint

WeasyPrint is a library that converts HTML and CSS content to PDFs. It allows you to leverage your existing knowledge of HTML and CSS to design your data science reports. The library is capable of rendering complex page layouts, including support for headers, footers, and page numbers.

import pandas as pd
import plotly.express as px
from weasyprint import HTML

# Create a sample DataFrame
data = pd.DataFrame({'A': [1, 2, 3], 'B': [
4, 5, 6]})

# Create a Plotly figure
fig = px.line(data, x=data.index, y='A')

# Save the figure as an image
fig.write_image("figure.png")

# Create an HTML string containing the table
html_table = data.to_html()

# Create an HTML document with the figure and table
html_content = f"""
<!DOCTYPE html>
<html>
<head>
  <style>
    table {{
      width: 100%;
      border-collapse: collapse;
    }}
    table, th, td {{
      border: 1px solid black;
      text-align: center;
      padding: 8px;
    }}
  </style>
</head>
<body>
  <img src="figure.png" alt="Plotly Figure">
  {html_table}
</body>
</html>"""
Convert the HTML content to a PDF
HTML(string=html_content).write_pdf("report.pdf")

FPDF

‍
FPDF is a lightweight library for generating PDF documents from scratch. It focuses on simplicity and ease of use, providing basic functionality for text, images, and simple graphics. This library might be suitable for users who do not require advanced PDF features and prefer a minimalistic approach.

Pros:
- Lightweight and easy to use
- Basic functionality for text, images, and simple graphics

Cons:
- Limited feature set
- No native support for complex layouts or tables

import pandas as pd
import plotly.express as px
from fpdf import FPDF

# Create a sample DataFrame
data = pd.DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})

# Create a Plotly figure
fig = px.line(data, x=data.index, y='A')

# Save the figure as an image
fig.write_image("figure.png")

# Create a PDF document
pdf = FPDF()
pdf.add_page()

# Add the figure to the document
pdf.image("figure.png", x=10, y=10, w=190)

# Add the table to the document
pdf.set_font("Arial", size=12)
col_widths = pdf.get_string_width(str(max(data.max()))) + 6
row_height = pdf.font_size * 1.5
table_x = 10
table_y = 100

for i in range(len(data.columns)):
    pdf.set_xy(table_x + col_widths * i, table_y)
    pdf.cell(col_widths, row_height, data.columns[i], border=1)
for i in range(len(data.index)):
    for j in range(len(data.columns)):
        pdf.set_xy(table_x + col_widths * j, table_y + row_height * (i + 1))
        pdf.cell(col_widths, row_height, str(data.iloc[i, j]), border=1)

# Save the document as a PDF
pdf.output("report.pdf")

A Simpler Alternative: Datapane

After comparing these libraries, we can conclude that creating PDF reports from Python can be a complex and tedious process. However, there is an easier alternative to generate data science reports: Datapane. This library allows you to create interactive, shareable HTML reports with just a few lines of code.

import pandas as pd
import plotly.express as px
import datapane as dp

# Create a sample DataFrame
data = pd.DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})

# Create a Plotly figure
fig = px.line(data, x=data.index, y='A')

# Create a Datapane report with the figure and table
v = dp.View(
  dp.Plot(fig),
  dp.DataTable(data)
)

# Publish the report as an HTML file
dp.save_report(v, path='report.html', open=True)

‍
As you can see, with Datapane, we can quickly generate a visually appealing and interactive report with minimal effort. The output is an HTML file, which might be a more suitable format for some use cases, especially when you consider the interactive capabilities of web-based visualizations.

In this blog post, we explored three notable open-source libraries for creating PDF data science reports from Python: ReportLab, WeasyPrint, and FPDF. Each library has its strengths and weaknesses, depending on the complexity and design requirements of your reports.

However, if you find creating PDF reports cumbersome and prefer a more straightforward approach, Datapane is an excellent alternative for generating interactive and shareable HTML reports with minimal code.

As a data scientist, it is crucial to choose the right tool for your reporting needs. Carefully consider the trade-offs between the various libraries and their output formats to make an informed decision that best suits your requirements.