Skip to content

Latest commit

 

History

History
94 lines (70 loc) · 4.48 KB

README.md

File metadata and controls

94 lines (70 loc) · 4.48 KB
Raw

YaCy Grid Component: Crawler

The YaCy Grid is the second-generation implementation of YaCy, a peer-to-peer search engine. A YaCy Grid installation consists of a set of micro-services which communicate with each other using the MCP, see https://github.com/yacy/yacy_grid_mcp

Purpose

The Crawler is a microservices which can be deployed i.e. using Docker. When the Crawler Component is started, it searches for a MCP and connect to it. By default the local host is searched for a MCP but you can configure one yourself.

What it does

The Crawler then does the following:

while (a Crawl Contract is in the queue crawler_pending) do
   - read the target url from the contract
   - check against the search index if the url is registered in the transaction index as 'to-be-parsed'. If not, continue
   - load the url content from the assets (it must have been loaded before! - that is another process)
   - parse the content and create a YaCy JSON object with that content
   - place the YaCy JSON within a contract in the index_pending queue
   - extract all links from the YaCy JSON
   - check the validity of the links using the crawl contract
   - all remaining urls are checked against the transaction index, all existing urls are discarded
   - write an index entry for the remaining urls with status 'to-be-loaded'
   - and these remaining urls are placed onto the loader_pending queue
   - the status of the target url is set to to-be-indexed
od

Required Infrastructure (Search Index, Asset Storage and Message Queues)

This requires an transaction index with the following information:

  • URL (as defined with https://tools.ietf.org/html/rfc3986)
  • crawlid (a hash)
  • status (to-be-loaded, to-be-parsed, to-be-indexed, indexed) As long as a crawl process is running, new urls (as discovered in the html source of a target url) must be written to the transaction index before the target url has a status change (from to-be-parsed to to-be-indexed). This makes it possible that the status of a crawl job and the fact that it has been terminted can be discovered from the transaction index.
  • if all status entries for a single crawlid are indexed then the crawl has been terminated. The Crawl process needs another database index, which contains the crawl description. The content must be almost the same as describe in http://www.yacy-websuche.de/wiki/index.php/Dev:APICrawler

Every loader and parser microservice must read this crawl profile information. Because that information is required many times, we omit a request into the cawler index by adding the crawler profile into each contract of a crawl job in the crawler_pending and loader_pending queue.

The crawl is therefore controlled by those queues:

  • loader_pending queue: entries which the yacy_grid_loader process reads. This process loads given resources and writes them to the asset storage.
  • crawler_pendingqueue: entries which the yacy_grid_crawler process reads. This process loads the content from the asset storage, parses the content and creates new loader_pending tasks.

The required indexes are:

  • a crawl profile index
  • a transaction index which reflects the crawl status
  • a search index

The microservices will create these indexes on their own using the MCP component.

Installation: Download, Build, Run

At this time, yacy_grid_crawler is not provided in compiled form, you easily build it yourself. It's not difficult and done in one minute! The source code is hosted at https://github.com/yacy/yacy_grid_crawler, you can download it and run loklak with:

> git clone --recursive https://github.com/yacy/yacy_grid_crawler.git

If you just want to make a update, do the following

> git pull origin master
> git submodule foreach git pull origin master

To build and start the crawler, run

> cd yacy_grid_crawler
> gradle run

Please read also https://github.com/yacy/yacy_grid_mcp/edit/master/README.md for further details.

Contribute

This is a community project and your contribution is welcome!

  1. Check for open issues or open a fresh one to start a discussion around a feature idea or a bug.
  2. Fork the repository on GitHub to start making your changes (branch off of the master branch).
  3. Write a test that shows the bug was fixed or the feature works as expected.
  4. Send a pull request and bug us on Gitter until it gets merged and published. :)

What is the software license?

LGPL 2.1

Have fun!

@0rb1t3r