Yes, we need more ES tutorials, and "introduction-to" articles.
On Thu, 2011-05-19 at 16:27 -0700, Theral Mackey wrote:
From my perspective also as a new user it drops you off with a running
ES server. All fine and dandy but it leaves out the part about how to
integrate a crawler, or feed content other than via running curl PUTs
(which it specifically states is not the way to do in production),
This kinda depends what language you want to use. There are several
APIs: most used are Java, Perl, Ruby, PHP, Python, and you may want to
This is why the examples are explained using curl - it's the "universal
language" that all of the APIs (except Java) translate their requests
You may want to read through some of the blog entries on
http://www.elasticsearch.org/blog/ - it will give you some taste for how
to use ES.
Also, here are a few examples in different languages:
and severely lacks any configuration guide (just some hints on a few
curl commands or that it can use jquery...
The point about ElasticSearch is that you shouldn't configure anything.
At least, not until you actually have to. ES has got very good
defaults, and generally just works out of the box. When you have a
better idea of your encounter a problem, that's the time to see what you
I would rather have the ability to know what Im doing than just see
wow, jquery! posted everywhere in the docs). I have gone and started
using OpenSearchServer (its gui is intuitive enough to figure out most
of what it can and cant do, even with a similar lack of docs) instead
until I can either gather enough info to get ES working with a crawler
and front-end (ed-head) or it gets updated and I can follow a more
complete guide. Even better would be a ES-Nutch how-to. I have seen
enough discussion about getting the projects working together, and the
upcoming 3.1 nutch release might have more for that, but any sort of
"this is where it looks for plugins or crawlers or other integration
stuff" would be most helpful. Docs right now seem more centered for
devs to use to imbed it in existing code rather than as a pure
stand-alone (with minimal crawler and frontend) search engine.
Yes - it is early days and kimchy has been focused on developing a
powerful, easy-to-integrate search server which developers from any
language can use in their applications, in a way that makes sense for
It is now up to us, the community, to develop reusable "pluggable"
applications and write about specific use cases.
It sounds like you are looking for a prepackaged search service, rather
than a highly customisable search engine.
Perhaps have a look at Karmi's field notes which explain how he setup ES
search for the elasticsearch.org website:
Your use case is one of many, and so far nobody has written that
particular app, although the easy part would be implementing search with
PS See my next email for an example in Perl