autioch/page-ripper

Work in progress

Page ripper

---

_ Work in progress _

---

Allows to "download" a page. This is usefull in few cases:

• there's no access to the "admin panel" for whatever reason,
• and old page is about to vanish,
• need to access the resources in offline mode,
• in worst case, malicious idea of copying/reusing other people's hard work :(

Configuration

The main export is an asynchronous function pageRipper, which requires configuration object:

• dataPath should be folder, in which folders for each page will be created,
• dbPath should be folder, in which sqlite db file will be stored,
• parsePost should be a function that will extract information from the page html,
• startingPages optional string array, that should be url of the first page to crawl.
• requestPause optional number, that will describe pause between each page request.

parsePost

Function that must be passed in configuration. It will be used to extract information from the retrieved post html.

It will accept three arguments:

• cheerio instance with loaded post body,
• post url,
• raw response html body.

It must return an object, with optional properties:

• id, uniquified if it already exists by appending __(count) to it,
• folderName, name of dir in which assets will be stored; if it's missing, no assets will be downloaded,
• nextUrls, array of urls, that be added to the queue and persisted in the database,
• imageUrls array of urls for images.

The object can contain any other properties, they will be all persisted in the database.

Usage

Cli

1. Download this repo
2. Install dependencies: npm i.
3. In file config.js provide valid config.
4. Execute start: npm run start.

NPM module

This app is currently unavailable in NPM registry.

TODO

1. Move all normalizations of urls to filenames to the crawler, instead of relying on the valid parsePost result.
2. Decide what should be unique key in the posts array, url or id?

• if id, then how it can by forsed/ensured to be unique by the parser?
• if url, then how should be treated query parameters, ports, protocols? Which part of url should be "unique"?
• maybe both fields should combine into unique property?