nutch-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Dennis Kubes (JIRA)" <j...@apache.org>
Subject [jira] Updated: (NUTCH-614) Order Inlinks by OPIC score of parent page
Date Tue, 26 Feb 2008 22:14:51 GMT

     [ https://issues.apache.org/jira/browse/NUTCH-614?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Dennis Kubes updated NUTCH-614:
-------------------------------

    Attachment: NUTCH-614-2-20080226.patch

Very, very messy patch.  This is a first cut at both allowing urls to be ordered by their
OPIC score and to allow those inlinks to be indexed according to those scores, meaning field
boosts for anchor text.  This patch removes the default document boost for nutch and changes
the index filter api to pass in the document or another relevant boost, which must now be
handled by the indexing filters.  The base indexing filter has been changed.  Because of this
many different plugins have been touched and not all of them have been correctly updated for
current behavior.  Inlinks inherit their score from their "coming from" parent's OPIC score.
 In the current cut of the anchor filter, anchors with no text or zero score are not indexed.

> Order Inlinks by OPIC score of parent page
> ------------------------------------------
>
>                 Key: NUTCH-614
>                 URL: https://issues.apache.org/jira/browse/NUTCH-614
>             Project: Nutch
>          Issue Type: Improvement
>    Affects Versions: 0.9.0
>         Environment: All
>            Reporter: Dennis Kubes
>            Assignee: Dennis Kubes
>             Fix For: 0.9.0, 1.0.0
>
>         Attachments: NUTCH-614-1-20080219.patch, NUTCH-614-2-20080226.patch
>
>
> Currently when saving inlinks there is a max number of inlinks (configurable) which get
saved and very little logic goes into deciding which inlinks get saved.  This patch uses the
OPIC score of the encompassing page to set a score for each inlink.  Inlinks are then reverse
sorted according to score and the best inlinks are saved first.  The logic behind this is
that pages with higher OPIC scores should have better links which they are pointing to.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message