Generating PDFs in Python with the ReportLab Toolkit (notes/outline for 2010 Jan 11 presentation to TuPLE : edited post-presentation version)

---

What is it?

A library for generating PDFs directly from Python code

Formerly called ReportLab PDF Library

http://www.reportlab.org

• for ReportLab Inc. open source

What are some alternatives?

• ReportLab Toolkit (aka ReportLab PDF Library)
  • FLOSS
  • cross-platform
  • since ca. 2001
  • opening existing PDFs to extract stuff requires non-FLOSS add-on.
  • I happen to have used a version of this. :-)

• pyPdf (at http://pybrary.net/pyPdf)
  • FLOSS
    • (‘a modified BSD’)
  • pure Python —> should be cross-platform
  • newer: 2006-
  • additionally,
    tries to let you open existing PDFs and read data out of them.
    • [Joel in audience commented that pyPdf is oriented more around taking
      an existing PDF and changing it.]

• pisa (aka XHTML2PDF?)
  • Generates PDF from HTML and CSS.
  • Uses ReportLab Toolkit AND pyPdf.
  • GPL

• z3c/pdftemplate ?
  • really a repackaging of ReportLab Toolkit?

• PyODConverter
  • an interface to OpenOffice.
  • LGPL

• at least one Windows-only proprietary $ solution using a COM interface.

• hack up interfaces to other languages for libraries e.g. PHP, Ruby.

Installation

In Feb 2008 I got my version from
http://cvs.sourceforge.net/cgi-bin/cvsweb.cgi/reportlab/README?cvsroot=reportlab

Old-style (as of Version 2.1, late 2007, early 2008):

• simple & easy manual installation
• fully documented
  • all changes you perform on file system are fully documented
• IF YOU WILL USE IMAGES, also requires the “Python Imaging Library”.

New-style:

• setup script
• Windows installer for Windows

License

• basically a post-1999 BSD license.
• It’s open source.
• You have to include (a copy of) the copyright notice and license
  with any redistribution
  • even in binary form
    • can be in accompanying
      manual, documentation, “other materials provided…”

What’s in this Library?

pdfgen is low-level; Platypus is higher-level.

pdfgen

provides features such as

• canvas (to draw on)
• cursors (to write text)
• text
  • can have e.g. fonts, origins, transforms (e.g. rotation, scaling)
• graphical/geometric paths
  • e.g. rectangles, ellipses, Bezier curves
• line drawing, fills…
• colors
• images

I ended up using Platypus for my layout.

Platypus

Formatting tools include

• document templates
  • page templates
    • non-flowing (fixed location on page) parts
      • e.g. can have header, footer, page number, watermarks…
    • frame into which ‘flowables’ go
      • e.g. text, tables, diagrams, hard page breaks…
• you feed a ‘story’ of ‘flowables’ into document template
  to generate the document.

My project used a table-based layout method

• for each page, a table (a flowable)
  • fills the page
  • holds everything for easy layout
• makes sense for my calendar/datebook with specific layout desired
• Platypus tables can calculate their height and width
• so code can even measure, fiddle things to get desired size/fit.
  • I used one ‘fiddle factor’,
    but suspect quick & dirty style
    of my relevant function was to blame.