Superseded

PythonReports is superseded by s-reports, a banded report formatter with a template format that grew out of this one. New reports should be written for s-reports.

Existing templates and printouts can be carried over with the pr2sr command; see Migrating to s-reports below.

Introduction

PythonReports is a toolkit aimed to build database reports in Python® programs. The toolkit includes the report builder and printout renderer for PDF output.

Report builder applies a template to a sequence of uniform data objects and produces a printout structure that can be saved in a file and/or rendered to PDF.

It is possible to render printouts for other presentation formats such as a specific printer language or a GUI toolkit display. There used to be printing output for wxPython and screen display for Tkinter and wxPython, but they fell out of use and their support is discontinued.

Requirements

Current version of PythonReports has only been tested with Python v3.12. Earlier versions of Python3 may or may not work; Python2 is not supported starting from PythonReports v1.0.0.

Report building and printout rendering require the ReportLab Toolkit. Reports containing images additionally require Python Imaging Library: Pillow (preferred) or PIL.

Templates in RSON format (introduced in v0.7.0) require rsonlite.

QR barcodes (introduced in v0.8.0) require qrcode.

Quick Start

A report is built from a list of data dictionaries or data objects having uniform structure. An example of such data is a list of rows fetched from an SQL server.

Report layout is defined by a template file which can have either XML format or a textual format based on rsonlite.

Report builder applies loaded template to a list of data objects and produces a printout structure. Printout can be saved in XML format.

PDF writer creates an output file from a printout structure.

There are example templates in the test directory: sakila.prt is in XML format, and sakila.prtr is the same template in RSON format. The data file for those templates is distributed as sakila.zip with the project releases.

This is an example of building a PDF from that data:

import pickle
from PythonReports.api import Builder, parse_rson, write_pdf

with open("sakila.dat", "rb") as data_file:
    data = pickle.load(data_file, encoding="latin1")[:200]
template = parse_rson("sakila.prtr")
build = Builder(template)
printout = build.run(data)
with open("sakila.prp", "wb") as printout_file:
    printout.write(printout_file)
# It is necessary to call .validate() on built or loaded printout
# before creating PDF output.
printout.validate()
write_pdf(printout, "sakila.pdf")

Migrating to s-reports

The pr2sr command converts PythonReports files to s-reports ones. It is installed with PythonReports, and python -m PythonReports.migrate runs it as well:

pr2sr template sakila.prt                  # writes sakila.kdl
pr2sr template sakila.prt --records payments.jsonl
pr2sr printout sakila.prp                  # writes sakila.srp.jsonl
pr2sr printout sakila.prp -o sakila.pdf    # rendered by s-reports

pr2sr template reads a template in either format, XML or RSON, and writes an s-reports template in KDL. It needs nothing beyond PythonReports, and rsonlite for RSON. pr2sr printout reads a printout and writes an s-reports printout, as NDJSON (.srp.jsonl), CBOR (.srp.cbor), or PDF (.pdf), so that archived printouts can be rendered by s-reports. It needs s-reports itself:

pip install PythonReports[sr]

Where s-reports is installed, pr2sr template also checks the converted template with it, and exits with status 1 if s-reports does not accept it.

What the conversion could not carry over unchanged is listed on standard error, and in the converted template as comments that begin with migrate:. These are the things to look at most often:

• Record fields. s-reports declares the fields of the data records and their types. The converter declares every field the expressions read. With --records, a JSON array or NDJSON file of sample records, the types come from the data; without it, they are guessed from how the expressions use the fields. A field that the expressions test for None or for truth is declared nullable.
• Data. s-reports reads its records from JSON rather than from Python objects, so the program that built reports has to write its data as JSON instead.
• Expressions. s-reports evaluates Starlark rather than Python. The converter rewrites what has a Starlark counterpart, and notes what has none, such as names from import elements.
• Deferred fields. s-reports sizes a field or a bar code that has evaltime by a placeholder text. Where the template gave none, the converter writes a guess.
• Values of None. PythonReports prints nothing for them, and s-reports prints None. A field that shows a variable prints only when the value is not None.
• Inline subreports. s-reports runs an inline subreport without a page header or footer of its own.