Auditing script for producing desktop and mobile screenshots, markup validator screenshots and lighthouse reports given a URL
- Node v7.6.0 or above
- Clone this Git repository into a new directory
After cloning the git repo, execute the following lines on the command line in the root directory:
npm init -ynpm install puppeteer lighthouse w3cjs psi
To run the auditing tool, all you need is the URL of the home page of the website/staging site and a Directory Name which will store the results of the audit. The following should then be typed into the terminal:
node audit.js --url <URL> --dir <Directory Name>
The tool produces a lighthouse report for every page that it finds. Each report takes time to produce so reports can be produced in parallel. You can change how many reports are produced in parallel in the settings.json file under the name "parallel". The default value is 5 lighthouse istances
If the website you are auditing requires a login to access all of the pages, there are some settings you can change so the tool automatically logins before any operations are performed. These settings (which are found in the settings.json file) include:
- loginPage : The page where the user logins (i.e /login or /auth/login)
- userName : The value of the username/email that the user would use to login (i.e admin or admin@email.com)
- userNameCSS : A unique CSS selector that identifies the text input where the user would enter their username (i.e "input[id='username']")
- password : The value of the password that the user would use to login (i.e admin or password)
- passwordCSS : A unique CSS selector that identifies the text input where the user would enter their password (i.e "input[id='password']")
- loginButtonCSS : A unique CSS selector that identifies the button which the user would click after their username and password has been entered (i.e "input[id='login']")
To exclude a link to the logout page, add the logout page to the list of exclusions in the settings.josn file.
To include this feature, append the --login tag like in the text below:
node audit.js --url <URL> --dir <Directory Name> --login
If you suspect some links are redirecting the page to a different/old staging site or an old site, you can use another feature provided by this tool to find these links. Since the auditing tool checks every href in every anchor tag, any href that contains a substring of the old site's URL can be logged and saved as a JSON file.
To use this feature, specify a substring of the old site's URL in the settings.json file under the name "linkString" and append the --link tag like so:
node audit.js --url <URL> --dir <Directory Name> --link
What will be returned will be a JSON file containing an array of objects where each object represents a link. The object's page is where the link was found and the object's href is the href attribute of the link. The JSON file can be found in the directory created by the tool.
To view the JSON in a nice readable format, open the following page http://chris.photobooks.com/json/default.htm and paste the generated JSON into the input.
Sometimes there are pages which can't be accessed from the home page or can't be accessed from clicking a link (For example a search page which requires the user to type a string into a text bar and hit search). These pages can be searched through as well by adding them to the pages array in the settings.json file. All pages should start with the character '/'.
To include this feature, append the --pages tag like so:
node audit.js --url <URL> --dir <Directory Name> --pages
Some pages might need a timeout to load before they display links. For instance if a page has to make an request to retrieve data from an SQL database, it might load up before loading any links causing the tool not to navigate the links that haven't been loaded .
To fix this there is an option in the settings.json file called timeout with two attributes called page and time. This option waits for x amount of milliseconds for page's containing a string s, where x equals the 'time' attribute and s equals the 'page' attribute.
To include this feature, append the --timeout tag like so:
node audit.js --url <URL> --dir <Directory Name> --timeout
To exclude pages that you do not want the auditing tool to navigate to, simply add the page to the exlusions array in the settings.json file
In a lot of websites, filler text such as 'Lorem Ipsum' is used to fill content in pages where the content is not known yet. You can use the setting findText to finds these pages as the setting returns an array of pages containing the text defined in the setting. This can be useful to find on what pages contain filler text. The default value for this setting is "lorem ipsum".
Results of the auditing tool can be found in the four directories of the directory created by the tool. The four folders are as follows:
- Images : Contains screenshots of all of the pages
- Lighthouse : Contains lighthouse reports in html of all of the pages
- MarkupValidator : Contains markup validator reports in html for all of the pages
- MobileImages : Contains mobile screenshots of all of the pages
A google page speed report is also created for all of your pages. To enable this you need an api key from google which can be obtained from the following link https://developers.google.com/speed/docs/insights/v2/first-app. This key should then be copied into the settings.json file under the name 'pageSpeedKey'. To enable the page speed report use the following tag --pageSpeed like so:
node audit.js --url <URL> --dir <Directory Name> --pageSpeed