README

Automating the boring stuff.

Managing a website can be difficult. You can end up with hundreds of pages, and making sure all of them are functional after an event, say a deployment, peaks of traffic, editorial changes, ads or javascripts behaving badly, ... Creating software and maintaining websites and apps is hard... regression testing should not.

Before you start, Clone / fork this repository and go into the folder. There are two ways of using this tool: With Vagrant or Docker.

Vagrant

Download and install Vagrant.

Requirements

• PHP 7.0
• Sqlite
• A (maybe big) list of urls to crawl. Robots.txt and Sitemaps detected automatically

Installation

1. Download and unzip the package:

curl -L -# -C - -O "https://github.com/alex-moreno/glitcherbot/archive/main.zip"
unzip main.zip
cd glitcherbot-main

1. Run: composer install

2. Make a copy of your config.php

cp config.sample.php config.php

Usage

Create a .csv which contains a list of urls to iterate over (see example.csv).

If using Acquia Site Factory, a command is supplied to generate a list of sites from a sites.json file. You'll need to:

1. Download the sites.json in your Acquia Cloud subscription

scp [subscription][ENV].[ENV]@[subscription][ENV].ssh.enterprise-g1.acquia-sites.com:/mnt/files/[subscription][ENV]/files-private/sites.json ./sites-dev.json

1. Vagrant up if you want to use the crawler inside the virtual machine (recommended).

2. Run the crawl against that json

php bin/visual_regression_bot.php acquia:acsf-crawl-sites sites.json

You can see all available commands by running:

php bin/visual_regression_bot.php

For help with a specific command use:

php bin/visual_regression_bot.php help <command>

Whilst debugging, increase verbosity by adding a number of -v flags.

-v : verbose -vv : very verbose -vvv : debug

Configuration

There are some settings that you can configure, like the headers that you'll send to the site or the concurrency that you want to use.

Move your config.sample.php into config.php and adapt to your needs. For example:

<?php

return [
  'headers' => [
    'User-Agent'   => 'GlitcherBotScrapper/0.1',
    'Accept'       => 'application/json',
  ],
  'http_errors' => false,
  'connect_timeout' => 0, // wait forever
  'read_timeout' => 0,
  'timeout' => 0, // wait forever
  'concurrency' => 60,
];

Note: The higher the concurrency is configured, the more sites it will run on each step, but be careful, php is fast (contrary to popular belief), it could send high load to a site and put it in trouble. Big power means bigger responsibility.

Stand Alone Tool

To run the regression tool as a stand along interface you need to point your webserver at the html/ directory in the repo.

PHP Web Server

A composer script has been included to aid with testing of the tool. To run this use the command.

composer start

Then navigate to the following address in your browser.

http://0.0.0.0:8000/

Docker

A docker setup has been included to aid with the running or the tool.

Download and install Docker

Starting the Docker container

This command will start the containers

make up

Building the tool (only need to run once)

This command will check if the config file exists and create one if needed. Then it will install all Composer dependencies.

make build

Crawling

This command will use sample-sites.csv as source of urls to Crawl by default.

make crawl

To run the command with a different file, use the syntax

make crawl SITES_CSV=path_to_sites_csv

or for Acquia json files:

make crawl-acquia SITES_JSON=sitesd8-prod.json

Keep in mind that the crawl runs within the container, so path_to_sites_csv needs to be relative to the container.

Web interface

Opens the tool on the browser.

make open

Stopping the container

make stop

Chained commands.

You can run all commands at once, for example the following command will start the containers, build, craw and open the browser.

make up build crawl open

Parameters

--include_sitemaps=yes

This will include all sitemaps in the website, if they are referenced from the robots.txt

--force_sitemaps=yes

Using makefile and Docker:

make crawl SITES_CSV=sample-sites.csv INCLUDE_SITEMAPS=yes

make crawl SITES_CSV=sample-sites.csv FORCE_SITEMAPS=yes