An experimental tool to help with navigating sites in order to capture a deeply nested page as a screenshot. This tool will allow you to capture pages that are under authentication. Page Capture Kit includes simplified logic to pre-fill forms with provided credentials.

Example:

page-capture-kit --scripts=/my/path/to/folder --output=/my/path/to/screenshots

Table of Contents

• Installation
• Motivation
• Scripts
  • Payload
  • Context
    • capture(filename, pattern)
    • click(selector, navigation)
    • end()
    • hover(selector)
    • login(username, password)
    • next()
    • open(url)
    • series(action1, action2, ..., actionX)
    • wait(condition)

Installation

Install from the NPM: npm install -g page-capture-kit

Or checkout this GitHub repository and link the package: npm link

Motivation

What if you need some data for your needs but service does not provide API to get it. What if every month you need invoice-like information from video services you are using. The motivation behind the tool to capture deeply nested pages with the information you can get in a different way other than navigating around the website and capturing a screenshot.

This tool does not include any fancy parsing logic or solutions to bypass CAPTCHA and other systems for the protection against unauthorized or bot access. But it can help with simple cases like to grab the last payment for Netflix or other services.

Scripts

Script is a concept for the set of instructions for Page Capture Kit to execute on. You write scripts in JavaScript and put them in some directory. Page Capture Kit will "play" them for you with the likely screenshot as output in the end.

The main helping tool in the script writing is Context. Context provides the set of predefined methods to help with actions like click items on the page, or visiting web pages.

Example of very basic script with Google (basic-google.script.js). Store this script somewhere and ask Page Capture Kit to play it:

class GoogleExample {
    visit(payload, context) {
        return context.series(
            context.open('https://www.google.com'),
            context.capture('Google'),
            context.end()
        );
    }
}

module.exports = GoogleExample;

# It will output PNG with Google page in the current directory where you run Page Capture Kit
page-capture-kit --scripts=/path/to/directory/with/basic-google --output=./ 

Technically scripts are JavaScript classes with a single method visit(payload, context), where payload - is potentially useful data, and context to help with actions for Page Capture Kit.

Payload

Simple Object with following properties:

• step <Number> - every time when visit is invoked step will increase. Counting starts from 0.

Context

Abstraction around possible actions with the current web page. There are three actions that help with orchestration around script execution: series(), next(), and end(). You should wrap all actions with series(), if you want to go to the next step, use next(), and do not forget to use end() in the very end of the script.

capture(filename, pattern)

Parameters:

• filename <String> - name of the screenshot
• pattern <String> - the format pattern for the date part in the filename. It's powered by Date Functions, follow documentation for the format. Default value is yyyy-MM-dd--HH-mm-ss.

click(selector, navigation)

Parameters:

• selector <String> - CSS selector or XPath for the element on the page
• navigation <Boolean> - checks if it's necessary to wait for the navigation on the page. Useful when links are clicked and browser will load a new page. Default: true.

end()

Use it in the end of all actions for the script. This action will terminate browse session.

hover(selector)

Parameters:

• selector <String> - CSS selector or XPath for the element on the page

login(username, password)

Do not store usernames and passwords in the scripts. Utilize environment variables. Please check examples directory for some ideas around credentials.

Parameters:

• username <String> - username credential for the service
• password <String> - password credentials for the service

next()

Asks Page Capture Kit to go to the next step. As a result visit() method in the script will be invoked with incremented step value. This technique could help to manage long and complex scripts.

open(url)

Parameters:

• url <String> - fully qualified URL to visit.

series(action1, action2, ..., actionX)

Groups context actions in the series of actions to execute in the context of the page.

wait(condition)

Parameters:

• condition <String> - it could be selector to wait for some element to appear on the page or it could be a timeout in milliseconds.