Posted on

Indexing my web site for search: part 4

Why do I get the feeling this will be a never-ending blog series? Perhaps because my javascript skills are still at the novice level?

I’ve finished my tweaking of the search interface for now. When a search is started from a non-search page, I give the user the search page with results. Once on the search page, up to 5 previous searches will be saved (in browser memory via jQuery calls and some other javascript), and may be flipped through by clicking the search results title (“N matches found for …”). I may change this number in the future. At the moment it has little effect on browser memory consumption, and is very handy for flipping through search results for different search terms.

You can check out the search page here.

My javascript foo is getting better, as is my CSS foo. It won’t be long before I move from procedural javascript to OO javascript (as unintuitive as its syntax is for OO, it works).

I like jQuery and the yui library.

Posted on

Indexing my web site for search: part 3

I now have my indexer working on albums and photos in the photo gallery. It needs some additional features, but it’s functional. So now I have indexing and searching for my HTML content, php content, blog posts and photos. This should be sufficient to get me by for a while; I’ll tweak it as needed.

The indexer (dwmsiteindex) uses a configuration file that permits skipping some directories when searching for HTML and php content. It is also used for database settings for the blog database and the gallery.

Posted on

Indexing my web site for search

I’ve been looking for site indexing code to support searching my web site.  Yes, you can certainly use Google for some of my content.  But Google won’t typically see content that’s not linked in.  Then there’s the data hidden away in databases instead of HTML.

Looking around at this time, nothing really excites me.  I want something fairly light weight, fast, and capable of indexing content from my HTML, PHP, blogs and photos.  I also want something that easily integrates into the look and feel of my site, today and 2 years from now.

Conceptually, I like xapian.  There’s also the ‘omega’ package built on xapian.  However, omega’s ‘omindex’ doesn’t work for me; I’ve got considerable textual content in PHP files and ‘omindex’ skips PHP.  It also doesn’t handle WordPress blogs, gallery3 databases, etc.  I’m also not terribly fond of the output of omega (though that’s not difficult to change).

I’ve decided I’ll write my own indexer and search using xapian as the back end.  In fact I’ve already written the code to index HTML and PHP files, and have a design for the code to index my blog posts.  I have a test search program that emits simple HTML.  All looks good so far; after indexing a handful of pages, searches yield appropriate document weights and rankings.