In the past, I have used Lucene collectors that calculated the
average/max score for each facet entry. Would like to replicate this logic
in ElasticSearch. My preference is to implement all custom code as plugins
since I do not want to maintain my own branch, but the facet infrastructure
is a bit complex and a re-working was said to be planned.
It should be easier to extend the existing code than it would be to create
a new facet type. Would such a feature be useful for the community in
general? Would only consider it if there would be a chance it can be merged
into the main branch.
I have also pull requests open on facet terms sorted by collators, waiting
for the pluggable facet API re-work. In the meantime, the move to Lucene 4
is of higher priority. I feel it is worth to add pull requests to mainline,
so it is easy for the Elasticsearch team to track all the suggestions and
get a picture of the whole story.
Jörg
On Wednesday, December 19, 2012 9:22:31 PM UTC+1, Ivan Brusic wrote:
In the past, I have used Lucene collectors that calculated the
average/max score for each facet entry. Would like to replicate this logic
in Elasticsearch. My preference is to implement all custom code as plugins
since I do not want to maintain my own branch, but the facet infrastructure
is a bit complex and a re-working was said to be planned.
It should be easier to extend the existing code than it would be to create
a new facet type. Would such a feature be useful for the community in
general? Would only consider it if there would be a chance it can be merged
into the main branch.
A pull request only makes sense if the feature would be used by the
community. I might start off by creating a plugin for one particular facet
(int terms) and see how that works.
I have also pull requests open on facet terms sorted by collators, waiting
for the pluggable facet API re-work. In the meantime, the move to Lucene 4
is of higher priority. I feel it is worth to add pull requests to mainline,
so it is easy for the Elasticsearch team to track all the suggestions and
get a picture of the whole story.
Jörg
On Wednesday, December 19, 2012 9:22:31 PM UTC+1, Ivan Brusic wrote:
In the past, I have used Lucene collectors that calculated the
average/max score for each facet entry. Would like to replicate this logic
in Elasticsearch. My preference is to implement all custom code as plugins
since I do not want to maintain my own branch, but the facet infrastructure
is a bit complex and a re-working was said to be planned.
It should be easier to extend the existing code than it would be to
create a new facet type. Would such a feature be useful for the community
in general? Would only consider it if there would be a chance it can be
merged into the main branch.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.