activity-logger

activity-logger Feature Module

Purpose

• logging a user's web page visit activity, interactions in each visit
• keep track of state related to visit activity per browser tab

How it works

Background script:

Keeps all the tab state (includes data about ongoing visits) managed, using tabs API events. The concept of visits comes into play with the tabs.onUpdated event, when a tab's URL is determined to be updated - this marks the beginning of a new visit. Logic behind the visit consists of 3 main stages:

1. initial page stub indexing
2. delayed page content indexing (optional)
3. final visit interations indexing

Stage 1 consists of indexing the page title, URL, and domain, and indexing the associated visit. This is done after the initial DOM is fully loaded. Note that the page data indexing can be skipped if there is a recent visit to the same URL. In this case, stage 2 will also be skipped, but the visit will be indexed.

Stage 2 consists of indexing the page text, favicon, screenshot, which is far more complex and space-consuming. page-storage and page-analysis modules will be invoked in this process to extract this data from the browser tab. This is done after the user has been active on the tab for at least 10s (managed by tabs state). Can be skipped.

Stage 3 consists of indexing the accumulated interaction data for this visit (tabs state). Data includes active time on page and scroll states. This is done when the user closes the tab or URL is determined to be changed (new visit).

Content script:

Simply handles deriving state from the scroll event on window and sending that to the background script. The background script uses this data to derive scroll px/% at any given time and also keep track of the max scroll along the y axis.