unitreport 0.0.1

A small unittest-based tool for generating single page html reports in Python.

:page_facing_up: unitreport

UnitReport is a small unittest-based tool for generating single page html reports in Python. The reports can include matplotlib figures and html tables. It is designed to be minimal and fit into the unittesting framework, allowing users to make assertions about their data (e.g. data quality) before generating figures.

Getting Started

You can install the library using,

pip3 install unitreport

There are usage examples in test_plots.py and test_tables.py. You need to create test cases using the unittest Python library, and utilise unitreport's decorators:

import unittest
import unitreport

import seaborn as sns
import pandas as pd

class TestExample(unittest.TestCase):
    """Example test suite producing plots and tables in a report using unitreport."""

    dataset: pd.DataFrame = None

    @classmethod
    def setUpClass(cls):
        """Load dataset on setup."""
        cls.dataset = sns.load_dataset("fmri")

    @unitreport.plotting
    def test_timepoint_vs_signal_by_event(self):
        """*fMRI data:* timepoint versus signal for stim and cue events."""
        # you can still run assertions to check data quality before plotting
        self.assertEqual(self.dataset.shape, (1064, 5))

        # plotting decorator will call plt.savefig() to generate the plot
        sns.relplot(
            x="timepoint",
            y="signal",
            hue="event",
            style="event",
            kind="line",
            data=self.dataset,
        )
        # could also return a figure & .savefig() will be called on returned object

    @unitreport.tabling
    def test_region_counts(self):
        """*fMRI data:* table summary description of timepoints and signals."""
        # you can still run assertions to check data quality before making table
        self.assertEqual(self.dataset.shape, (1064, 5))

        return self.dataset.describe().to_html()

You can run the tests using,

python3 -m unitreport

This will discover and run the tests (which could be across multiple test files), generate the report and save it to the current directory.

For extra parameters you can run the following,

python3 -m unitreport -h

usage: __main__.py [-h] [--pattern PATTERN] [--templates_dir TEMPLATES_DIR]

optional arguments:
  -h, --help            show this help message and exit
  --pattern PATTERN     File patterns to discover test cases in.
  --templates_dir TEMPLATES_DIR
                        Path to jinja2 templates directory including
                        index.html and main.css

Built With

• unittest - underlying testing framework
• Jinja2 - rendering and generating the report