OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

docbook-apps message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Re: [docbook-apps] WebHelp search - anybody working on improving?

I would be willing to help with this effort, but not lead it. If someone were willing to evaluate better alternatives and integrate the code into DocBook's webhelp, I could write the XSL templates that generate the index files.

Bob Stayton
Sagehill Enterprises

On 11/11/2016 1:45 AM, Fekete, Róbert wrote:

A quick google search shows lunr.js
(https://github.com/olivernn/lunr.js ), and a possibly improved for of
it called elasticlunrjs (http://elasticlunr.com/). Does anyone have an
idea about how to integrate one of these into webhelp? (Specifically,
how to generate the index file form the webhelp HTMLs? This post has
some pointers, but I'm not a developer, so I'm unsure how to get it
going: https://29a.ch/2014/12/03/full-text-search-example-lunrjs )



On Thu, Nov 10, 2016 at 9:21 PM, Jan Tosovsky <j.tosovsky@email.cz
<mailto:j.tosovsky@email.cz>> wrote:

    On 2016-11-10 Janice Manwiller wrote:
    > The WebHelp search is a source of frustration, mostly because it does
    > not support phrase searches...
    > Has there been any effort to improve the search? Has anyone else
    > implemented a custom search that supports phrase searches?

    In the current implementation there is no way to do it. When search
    index is built, the original content is split into words, from which
    kind of look-up table is created (which word in which file is present).

    When search phrase is entered, it is again split into separate words
    and each of them is searched in that look-up table.

    The result is the number of occurrences of the given word in the
    particual file, which is used for ordering the search results.

    In case of phrase searches the search index would have to store the
    full content. When performing the search all those content snippets
    would have to be processed using more complex algorithms.

    But instead of reinventing the wheel I believe there are some
    lightweight JavaScript ports of Lucene engine, which could be
    somehow integrated. However, I am not expert in this field.


    To unsubscribe, e-mail:
    For additional commands, e-mail:

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]