[hypermail] Indexing with Swish-e

From: Bill Moseley <moseley_at_hank.org_at_hypermail-project.org>
Date: Thu, 5 Feb 2004 15:05:26 -0800
Message-ID: <20040205230526.GB6799_at_hank.org>

I noticed this page on indexing with Swish-e:


Here's another way if using Perl.

The Swish-e (http://swish-e.org) package comes with a script for indexing hypermail archives and includes instructions on usage. Below I've included part of the documentation.

That script is in use here:


The search results page is not that pretty, but it's just using the default templates. I'm often frustrated when searching archives that I can't limit by date or author. So besides being able to just search by name, email, title, etc., you can do a search like:

  hypermail name=moseley

to find posts with "hypermail" in either the title or body, and also only messages from moseley.

Here's the (un-proof read) instructions:

    index_hypermail.pl - Parse Hypermail archive for indexing with Swish-e

    Using an example data structure like this:


    Create the hypermail archive:

$ cd hypermail
$ hypermail -i -d archive < messages.mbox

    Create a swish-e config file:

$ cd search
$ cat swish.conf

        # config for indexing hypermail v2.1.8 archives

        IndexDir ./index_hypermail.pl
        SwishProgParameters ../archive

        MetaNames swishtitle name email
        PropertyNames name email
        IndexContents HTML* .html
        StoreDescription HTML* <body> 100000
        UndefinedMetaTags  ignore

    Copy index_hypermail.pl to the current directory. Swish-e installs     index_hypermail.pl in the $prefix/share/doc/swish-e/examples/prog-bin     directory, where $prefix is typically "/usr/local" or simply "/usr" on     some distributions.

$ cp /usr/local/share/doc/swish-e/example/prog-bin/index_hypermail .


    Index the documents:

$ swish-e -c swish.conf -S prog

    Now create the search interface:

$ cp /usr/local/lib/swish-e/swish.cgi .
$ cat .swishcgi.conf

        $ENV{TZ} = 'UTC'; # display dates in UTC format

        return {
            title           => "Search the Foo List Archive",
            display_props   => [qw/ name email swishlastmodified /],
            sorts           => [qw/swishrank swishtitle email swishlastmodified/],
            metanames       => [qw/swishdefault swishtitle name email/],
            name_labels     => {
                swishrank           =>  'Rank',
                swishtitle          =>  'Subject Only',
                name                =>  "Poster's Name",
                email               =>  "Poster's Email",
                swishlastmodified   =>  'Message Date',
                swishdefault        =>  'Subject & Body',

            highlight       => {
                package         => 'SWISH::PhraseHighlight',

                xhighlight_on    => '<font style="background:#FFFF99">',
                xhighlight_off   => '</font>',

                meta_to_prop_map => {   # this maps search metatags to display properties
                    swishdefault    => [ qw/swishtitle swishdescription/ ],
                    swishtitle      => [ qw/swishtitle/ ],
                    email           => [ qw/email/ ],
                    name            => [ qw/name/ ],
                    swishdocpath    => [ qw/swishdocpath/ ],

    Setup web server (OS/web server dependent):

        /var/www # ln -s /path/to/hypermail/search
        /var/www # ln -s /path/to/hypermail/archive

    and maybe tell apache to run the script:

$ cat .htaccess

        Deny from all
        <files swish.cgi>
            Allow from all
            SetHandler cgi-script
            Options +ExecCGI

Bill Moseley
Received on Fri 06 Feb 2004 07:38:16 AM GMT

This archive was generated by hypermail 2.3.0 : Sat 13 Mar 2010 03:46:12 AM GMT GMT