Hit CollectionsHit Collections
A hit collector is a way to collect hits (instances of org.polliwog.data.Hit), based on certain criteria, as polliwog is processing a log file.

Once collected the hits can be processed and sections/pages generated to display the information. For instance you may wish to collect "image hot-linking hits" (that is hits from other websites that link directly to images on your site).

Each hit collection defines the types of hits that it wants to collect.

The collector definitions are stored in file: data/hit-collectors.xml. This is an xml file that has the following elements/attributes:
Show the help for this table
collectorsYcollector+NONENONEThe root element, each child collector element defines a collection.
collectorNANYcollectorson(enum{nonPage, page, filtered},R)
Defines a particular collection. The class attribute defines the class that will represent the collection and implements the org.polliwog.collectors.HitCollector interface. The name attribute defines a unique name (within the list of hit collections) for the collection, it is this identifier that can be used in sections to identify the collection to use as the input. The on attribute defines the types of hit that should be collected, should be a comma-separated list of one of more of the values: nonPage, page, filtered. The usual value for this attribute is: nonPage, page.
Type of Hits
There are 3 types of hits that are relevant to hit collections. The types relate to how polliwog categorizes the hit. It should be noted that in terms of hit collections a hit will only be classified as one of the types. The types are:
  • Filtered - The hit is categorized as filtered if there is a hit filter in use and it rejects the hit, i.e. it has been filtered and won't contribute to any visit/pages statistics. To collect this type of hit, use a value of: filtered for the on attribute of the collector element.
  • Non page - The hit is categorized as not a page if the page collector decides that the hit does not constitute a page. To collect this type of hit, use a value of: nonPage for the on attribute of the collector element.
  • Page - The hit is categorized as a page if the page collector decides that the hit does constitute a page. To collect this type of hit, use a value of: page for the on attribute of the collector element.
Basic Hit Collector
Whilst it is possible to provide your own implementation of the HitCollector interface a basic implementation that allows you to define criteria for collecting hits via a External site, opens in new window JoSQL WHERE clause is available with the org.polliwog.collectors.BasicHitCollector class.

To use this class just use a value of:
for the class attribute of the collector element.

The WHERE clause is then provided by the content of the collector element. Remember that the class for any accessors is: Hit. Any valid WHERE clause can be used, the expression used will be evaluated to a boolean true/false value. This value is returned from the accept(org.polliwog.data.Hit) method.

Find all hits that have a request parameter: search for pages.
<collector on="page"
  get (requestParameters, "search") != NULL
Find hot linked image hits (replace www.mysite.com with the name of your website).
<collector on="nonPage,page"
  refererURI NOT LIKE 'http://www.mysite.com'
  refererURI.scheme LIKE 'http%'
  requestURI.path IN $LIKE ('%.gif', '%.jpg')