Seems a lot of work to create such query, I thought it will be much
easier, as they say Elasticsearch is so ready to use...
On 29 Cze, 14:36, Rafał Kuć r....@solr.pl wrote:
Hello!
There are two plugins for Polish for Elasticsearch. One can be found
here:GitHub - elastic/elasticsearch-analysis-stempel: Stempel (Polish) Analysis Plugin for Elasticsearch
and the second one is here:
The first one is from Elasticsearch team and uses Stempel library for
handling Polish. The second one uses Morfologik analyzer, which is
better in terms of the number of words properly handled (at least from
my experience).If you want to see some comparison on how those libraries work, take a
look athttp://solr.pl/2012/04/02/solr-4-0-i-mozliwosci-analizy-jezyka-polski...
Its not based on Elasticsearch, but on Solr, but you will be able to
see how those libraries compare to each other.--
Regards,
Rafał Kuć
Sematext ::http://sematext.com/:: Solr - Lucene - Nutch - ElasticsearchSomething like this, mayby not so much like finding bought when typing
buy, but I'm searching in our beautifull Polish languageso when I
will tyle "szkolenie" I would also want to find "szkoleń",
"szkoleniowy", "szkolenia" and so on...
On 29 Cze, 14:21, Rafał Kuć r....@solr.pl wrote:Hello!
You mean that when you type in plural form, you also find singular and
all the cases like buy, bought should be find when typing buy ? Then
you also should look at analysis, because its a matter of the correct
language analysis:Elasticsearch Platform — Find real-time answers at scale | Elastic...
.--
Regards,
Rafał Kuć
Sematext ::http://sematext.com/::Solr - Lucene - Nutch - ElasticsearchAAX ->> AAA I meant similarity so when I will write CAT I will always
fins CATS etc.
On 29 Cze, 14:07, Rafał Kuć r....@solr.pl wrote:Hello!
Use boolean query with OR to get all documents with at least one term,
than use phrase query to boost the documents that have the words close
to each other (look at the phrase sloop to see how you control that)
and finally use boosting query to boost the documents that have all
the terms in the field you are searching on. Of course all of that
should be combined into a single query for Elasticsearch to do all the
score calculation for you.I don't know what you want to achieve when you say you want to have
results of 'AAX' when querying for 'AAA'.--
Regards,
Rafał Kuć
Sematext ::http://sematext.com/::Solr- Lucene - Nutch - ElasticsearchHeh yeah I know that it's kind of problem with answers like "you can
or you can or you can" and it all depends on what you need... problem
is that it's hard to start using this search engine because of lack of
documentation tuts etc
I would like to use it only in one case I'm searching the documents
using only one field always and I want to build the most standard
search engine for frazes that will return results from best similarity
to lower and so on...
so if I have a fraze like "AAA BBB CCC" search results will be like:
AAA BBB CCC
AAX BBB CCC
AAA BBB
AAX BBB
AAA
AAX
On 29 Cze, 13:28, Rafał Kuć r....@solr.pl wrote:Hello!
Its hard to give a general query that will work in all situations, I
don't think it's even possible. You should think what fields in your
index are important to you (and use boosting on those), you should
think if you want phrase queries to be used to boost documents that
have them present. You may also want to have some fields boost score
of your documents (for example number of likes or recent date) so you
may want to use custom score query. And many, many more 'You may'.You should think what you want Elasticsearch to return and how and
this way construct your queries.--
Regards,
Rafa³ Kuæ
Sematext ::http://sematext.com/::Solr-Lucene - Nutch - Elasticsearchanyone? I'm looking for best configuration of query_string for
searching results?