How to debug/analyze a morelikethis query?


(vinhphu1711) #1

Hi,
I'm playing with the More Like This query
http://www.elasticsearch.org/guide/reference/query-dsl/mlt-query.html
.
There're quite a couple of parameters to tune, mainly to select the
right tokens in the source document to be compared against the rest in
the index.

I'm curious to see if there's any convenient way to look at the tokens
of the "like text" after analyzed by ElasticSeach (i.e: after stop
word removal, tokenizing, tf idf filtering...). The idea is similar to
the Admin Analyze API: http://www.elasticsearch.org/guide/reference/api/admin-indices-analyze.html

This would be very helpful for one to tweak the parameters in a
MoreLikeThis query. Is there already such an API for MoreLikeThis ?

Thank you in advance.

LTVP


(vinhphu1711) #2

After searching a while, I saw this ticket: https://github.com/elasticsearch/elasticsearch/issues/1412
It's exactly what I'd like to ask for.

LTVP

On May 8, 2012, at 5:35 PM, Phu Le wrote:

Hi,
I'm playing with the More Like This query
http://www.elasticsearch.org/guide/reference/query-dsl/mlt-query.html
.
There're quite a couple of parameters to tune, mainly to select the
right tokens in the source document to be compared against the rest in
the index.

I'm curious to see if there's any convenient way to look at the tokens
of the "like text" after analyzed by ElasticSeach (i.e: after stop
word removal, tokenizing, tf idf filtering...). The idea is similar to
the Admin Analyze API: http://www.elasticsearch.org/guide/reference/api/admin-indices-analyze.html

This would be very helpful for one to tweak the parameters in a
MoreLikeThis query. Is there already such an API for MoreLikeThis ?

Thank you in advance.

LTVP


(vinhphu1711) #3

Hi, any update on that? Or do you have any idea of an efficient way to debug and tweak a morelikethis query?

On May 8, 2012, at 6:05 PM, Phu Le wrote:

After searching a while, I saw this ticket: https://github.com/elasticsearch/elasticsearch/issues/1412
It's exactly what I'd like to ask for.

LTVP

On May 8, 2012, at 5:35 PM, Phu Le wrote:

Hi,
I'm playing with the More Like This query
http://www.elasticsearch.org/guide/reference/query-dsl/mlt-query.html
.
There're quite a couple of parameters to tune, mainly to select the
right tokens in the source document to be compared against the rest in
the index.

I'm curious to see if there's any convenient way to look at the tokens
of the "like text" after analyzed by ElasticSeach (i.e: after stop
word removal, tokenizing, tf idf filtering...). The idea is similar to
the Admin Analyze API: http://www.elasticsearch.org/guide/reference/api/admin-indices-analyze.html

This would be very helpful for one to tweak the parameters in a
MoreLikeThis query. Is there already such an API for MoreLikeThis ?

Thank you in advance.

LTVP


(system) #4