Skip to content
Permalink
Branch: master
Find file Copy path
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
147 lines (100 sloc) 7.14 KB

Memex Explorer Crawler Guide

memex-explorer uses two crawlers, :ref:`ache-section` and :ref:`nutch-section`.

Crawler Overview

Both crawlers have their own unique designs, and both use the data they collect in unique ways.

There is some commonality between the two, however. They both require a list of URLs to crawl, called a seeds list, and they both share similar interactivity with the Crawler Control Buttons.

This section will go over the common elements of the two crawlers.

Creating a Seeds List

The common point between the two crawlers is that they both use the same kind of seeds list for their crawling. The seeds list is comprised of a list of urls separated by line breaks. Both Nutch and Ache use them in different ways, and the result you get directly from the crawlers is different for each of them. Here is a sample seeds list:

http://www.reddit.com/r/aww
http://gizmodo.com/of-course-japan-has-an-island-where-cats-outnumber-peop-1695365964
http://en.wikipedia.org/wiki/Cat
http://www.catchannel.com/
http://mashable.com/category/cats/
http://www.huffingtonpost.com/news/cats/
http://www.lolcats.com/

Simply put, the seeds list should contain pages that are relevant to the topics you are searching. Both Nutch and Ache provide insight into the relevance of your seeds list, but in different ways.

For the purposes of memex-explorer, the extension and name of your seeds list does not matter. It will be automatically renamed and stored according to the specifications of the crawler.

Seeds lists are created on the seeds page, and seeds lists can be created from the add crawl page.

Crawler Control Buttons

Here we have an overview of the buttons available to each crawler for controlling the crawlers. The buttons behave differently depending on which one you are using.

These are the buttons available for Ache:

_static/img/ache-buttons.png

These are the buttons available for Nutch:

_static/img/nutch-buttons.png

Options Button

Symbolized by the "pencil" icon. This allows you to change various settings on the crawl. See Crawl Settings.

Start Button

Symbolized by the "play" button. This will start the crawler for you, and will display the status as "starting" immediately after pressing it, and "running" after the crawl has been started.

Stop Button

Symbolized by the "stop" button. Stops the crawl.

In the case of Ache, the crawler stops immediately. In the case of Nutch, the crawler stops after it has finished the current process. However, the data on the current round of the crawl will be lost.

Restart Button

Symbolized by the "refresh" icon. Restarts the current crawl. This button is only available after the crawl has stopped.

With Ache, it will immediately start a brand new Ache crawl, deleting all of the previous crawl information. With Nutch, it will start a new crawler round, using the information gathered by the crawl in the previous round.

Get Crawl Log

This button will let you download the log of the current running crawl. This allows you to see the progress of the crawl and any errors that may be occurring during the crawl. This is only available for Ache crawls.

CCA Export

This button is Nutch only. It allows you to export your crawl data into the CCA format.

Rounds Input

Nutch only. This allows you to specify how many rounds you want the crawl to run. You can press the stop button at any time and it will stop when it is done with the current round.

Crawl Settings

The crawl settings page allows you to delete the crawl, as well as change the name or description of the crawl. It is accessed by clicking the "pencil" icon next to the name of the crawl.

_static/img/crawl_settings.png

Here you can change the name or description of the crawl. You can also delete the crawl.

Nutch

Nutch is developed by Apache, and has an interface with Elasticsearch. All Nutch crawls create Elasticsearch indices by default.

With Nutch, you can define how long you want to crawl by setting the number of rounds to crawl. You can keep track of the overall crawl time and the sites currently being crawled by looking at the Nutch crawl visualizations.

The number of pages left to crawl in a Nutch round increases significantly after each round. You might pass it a seeds list of 100 pages to crawl, and it can find over 1000 pages to crawl for the next round. Because of this, Nutch is a much easier crawler to get running.

Memex Explorer currently uses the Nutch REST API for running all crawls.

Nutch Dashboard

Memex explorer recently added features for monitoring the status of Nutch crawls. You can now get real-time information about which pages Nutch is currently crawling, and information about the duration of the crawl.

_static/img/nutch-dashboard.png

Statistics

Nutch will tell you how many pages have been crawled after the current round has finished.

_static/img/nutch_stats.png

Ache

Ache is developed by NYU. Ache is different from Nutch because it requires a crawl model to be created before you can run a crawl (see Building a Crawl Model). Unlike Nutch, Ache can be stopped at any time. However, if you restart an Ache crawl, it will erase all the data from the previous crawl.

Ache Dashboard

_static/img/ache-dashboard.png

_static/img/ache_stats.png

Plots

Memex Explorer uses Bokeh for its plots. There are two plots available for analyzing Ache crawls, Domain Relevance and Harvest Rate.

The Domain Relevance plot sorts domains by the number of pages crawled, and adds information for relevancy of that domain to your crawl model. This plot helps you understand how well your model fits.

The Harvest Rate plot shows the overall performance of the crawl in terms how many pages were relevant out of the total pages crawled.

Statistics

Like Nutch, Ache also collects statistics for its crawls, and allows you to see the head of the seeds list.

Harvest rate reflects the relevance to the model of the pages crawled. In this case, 58% of the pages crawled were relevant according to the model.

Ache Specific Buttons

Ache has a "Download Relevant Pages" button, which will allow you download which pages Ache has found to be relevant to your seeds list and your crawl model.

Building a Crawl Model

Ache requires a crawl model to run. For information on how to build crawl models, see the Ache readme.

For more detailed information on Ache, head to the Ache Wiki.

You can’t perform that action at this time.