A Python library to execute Lighthouse audits and export results of the running performance tests against different URL's. Aims to track code quality and improve user experience.
- Python 3 (not compatible with Python 2 at the moment)
- Lighthouse
- Google Chrome (>= Chrome 59 for headless support)
- Clone the repository or download the project as a zip file from this Github page
- Install Lighthouse as a Node command line tool:
npm install -g lighthouse
# or use yarn:
yarn global add lighthouse
Note: Lighthouse requires Node 6 or later. It is recommended to install the current Long-Term Support version of Node.
Create virtualenv to target the Python 3 interpreter, e.g.:
python3.6 -m venv ~/.venvs/pointloma3.6source ~/.venvs/pointloma3.6/bin/activate
pip install -r requirements.txt
python pointloma [-h] [-r RUNS] [-o OUTPUT_PATH] [-v] url
positional arguments:
url url to test against
optional arguments:
-h, --help show this help message and exit
-r RUNS, --runs RUNS number of test runs
-o OUTPUT_PATH, --output-path OUTPUT_PATH
path to csv file output
-v, --verbose increase output verbosity
-a AUTH_MODULE, --auth-module AUTH_MODULE
authentication module to use
Point Loma supports testing URLs which are behind authentication with custom auth modules. Currently, a single auth module is supported, kolibri, but additional ones can be written for specific use cases.
kolibri auth module enables authentication to Kolibri instances and its code is extensively commented so it should be rather straightforward to use it to write your own custom authentication module.
To be able to use kolibri (or custom auth module) you should add the following environment variables with the user credentials, e.g. in your ~/.bashrc:
POINTLOMA_USERNAME=yourusernamePOINTLOMA_PASSWORD=yourpassword
You could also simply prepend environment variables to the pointloma CLI command, but of course, in that case you would have to type it every time you use it (adding it to your ~/.bashrc or similiar script would persist it), e.g.:
POINTLOMA_USERNAME=yourusername POINTLOMA_PASSWORD=yourpassword python pointloma --auth-module kolibri http://kolibridemo.learningequality.org/user
python pointloma -r 3 https://example.com
python pointloma -r 3 --output-path /tmp/example.csv http://localhost:8000
python pointloma -v https://example.com
Environment variables added to e.g. ~/.bashrc:
python pointloma --auth-module kolibri http://kolibridemo.learningequality.org/user
Environment variables prepended to the command:
POINTLOMA_USERNAME=yourusername POINTLOMA_PASSWORD=yourpassword python pointloma --auth-module kolibri http://kolibridemo.learningequality.org/user
Resulting output is a comma delimited csv file with the following columns:
- Timestamp
- First Meaningful Paint [ms]
- First Interactive [ms]
- Consistently Interactive [ms]
- Speed Index [ms]
- Estimated Input Latency [ms]
The csv file with the results will be written to one of the two following locations:
- path specified by
-oor--output-pathCLI arguments - to the
outputdirectory under the thepointlomacodebase root directory, e.g.:output/results_1987_08_20__11_22_33__123456.csv
The same output path can be used for multiple test runs (by using the -o or --output-path CLI options) as the results will simply be appended to the specified csv file.
This approach can be useful when testing a single URL with different code states or repository branches and wanting to gather the results in a single csv file for easier processing.
- Automatically generate graphs from the resulting csv
- Test multiple URLs by specifying file with the URLs listed
- Integrate with Health Inspector (https://github.com/learningequality/auto-screenshots)