org.clapper.curn.plugins
Class ArticleFilterPlugIn

java.lang.Object
  org.clapper.curn.plugins.ArticleFilterPlugIn

All Implemented Interfaces:

FeedConfigItemPlugIn, PlugIn, PostFeedParsePlugIn

public class ArticleFilterPlugIn
extends java.lang.Object
implements FeedConfigItemPlugIn, PostFeedParsePlugIn

The ArticleFilterPlugIn provides per-feed filtering capabilities. It can be used to filter out items (articles) that match one or more regular expressions.

The filtering syntax is (shamelessly) adapted from the rawdog RSS reader's article-filter plug-in. (rawdog is similar to curn: It's a command line-driven RSS reader, written in Python.) A feed filter is configured by adding an ArticleFilter property to the feed's configuration section. The property's value consists of one or more filter command sequences, separated by ";" characters. (The ";" must be surrounded by white space; see below.) Each filter command sequence is of this form:

show|hide [field 'regexp' [field 'regexp' ...]]

field can be one of:

• author: search the author field
• title: search the title field
• summary: search the summary, or description, field
• text: search the full content, if available
• category: search the article's category (or categories)
• any: search all fields

Each regular expression must be enclosed in single quotes.

For example:

 hide author 'Raymond Luxury-yacht' ; show author 'Arthur .Two-sheds. Jackson'
 

If the command is "hide", then the entry will be hidden if the specified field matches the regular expression. If the command is "show", then the entry will be shown if the field matches the regular expression. If there are no fields or regular expressions, then the command is a wildcard match. That is:

hide

is equivalent to:

hide any '.*'

and:

show

is equivalent to:

show any '.*'

Wildcard matches are useful in situations where you want to hide or show "everything but ...". See the examples, below, for details.

All filtering commands are processed, and the end result is what defines whether a given entry is suppressed or not. Regular expressions are matched in a case-blind fashion. The match logic also

• ignores any embedded newlines in article contents
• (temporarily) strips all HTML from the article text before matching

You can use multiple ArticleFilter parameters per feed (as long as they have unique suffixes. All filters are applied to each article to determine whether the article should be filtered out or not.

Examples

Some examples will help clarify the syntax.

For example, the following set of commands hide all articles with the phrase "mash-up" (because mash-ups bore me):

 ArticleFilter: hide any 'mash[- \t]?up'
 

The following, more complicated, entry hides everything by author "Joe Blow", unless the title has the word "rant" in it ('cause his rants are hilarious):

 ArticleFilter: hide author '^joe *blow$' ; show author '^joe *blow$' title rant
 

Finally, this example hides everything except articles by Moe Howard:

 ArticleFilter: hide ; show author '^moe *howard$'
 

Version:

$Revision: 6509 $

Constructor Summary

ArticleFilterPlugIn()
Default constructor (required).

Method Summary

java.lang.String	getPlugInName() Get a displayable name for the plug-in.
java.lang.String	getPlugInSortKey() Get the sort key for this plug-in.
void	initPlugIn() Initialize the plug-in.
boolean	runFeedConfigItemPlugIn(java.lang.String sectionName, java.lang.String paramName, CurnConfig config, FeedInfo feedInfo) Called immediately after curn has read and processed a configuration item in a "feed" configuration section.
boolean	runPostFeedParsePlugIn(FeedInfo feedInfo, RSSChannel channel) Called immediately after a feed is parsed, but before it is otherwise processed.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

ArticleFilterPlugIn

public ArticleFilterPlugIn()

Default constructor (required).

Method Detail

getPlugInName

public java.lang.String getPlugInName()

Get a displayable name for the plug-in.

Specified by:

getPlugInName in interface PlugIn

Returns:

the name

getPlugInSortKey

public java.lang.String getPlugInSortKey()

Get the sort key for this plug-in.

Specified by:

getPlugInSortKey in interface PlugIn

Returns:

the sort key string.

initPlugIn

public void initPlugIn()
                throws CurnException

Initialize the plug-in. This method is called before any of the plug-in methods are called.

Specified by:

initPlugIn in interface PlugIn

Throws:

CurnException - on error

runFeedConfigItemPlugIn

public boolean runFeedConfigItemPlugIn(java.lang.String sectionName,
                                       java.lang.String paramName,
                                       CurnConfig config,
                                       FeedInfo feedInfo)
                                throws CurnException

Called immediately after curn has read and processed a configuration item in a "feed" configuration section. All configuration items are passed, one by one, to each loaded plug-in. If a plug-in class is not interested in a particular configuration item, this method should simply return without doing anything. Note that some configuration items may simply be variable assignment; there's no real way to distinguish a variable assignment from a blessed configuration item.

Specified by:

runFeedConfigItemPlugIn in interface FeedConfigItemPlugIn

Parameters:

sectionName - the name of the configuration section where the item was found

paramName - the name of the parameter

config - the active configuration

feedInfo - partially complete FeedInfo object for the feed. The URL is guaranteed to be present, but no other fields are.

Returns:

true to continue processing the feed, false to skip it

Throws:

CurnException - on error

See Also:

CurnConfig, FeedInfo, FeedInfo.getURL()

runPostFeedParsePlugIn

public boolean runPostFeedParsePlugIn(FeedInfo feedInfo,
                                      RSSChannel channel)
                               throws CurnException

Called immediately after a feed is parsed, but before it is otherwise processed. This method can return false to signal curn that the feed should be skipped. For instance, a plug-in that filters on the parsed feed data could use this method to weed out non-matching feeds before they are downloaded. Similarly, a plug-in that edits the parsed data (removing or editing individual items, for instance) could use method to do so.

Specified by:

runPostFeedParsePlugIn in interface PostFeedParsePlugIn

Parameters:

feedInfo - the FeedInfo object for the feed that has been downloaded and parsed.

channel - the parsed channel data

Returns:

true if curn should continue to process the feed, false to skip the feed. A return value of false aborts all further processing on the feed. In particular, curn will not pass the feed along to other plug-ins that have yet to be notified of this event.

Throws:

CurnException - on error

See Also:

RSSChannel, FeedInfo