Feature Request: tags_with_weight for subsets

[airy]

Post.limit(10).tags_with_weight and Post.tags_with_weight returns the same results. Having filtered tags_with_weight would be realy nice.

thanks

[wilkerlucio]

Hi airy,

Yes, they return the same... You can't limit or sort or do any of these things for this method, because it uses a low level collection for this data, in fact, after each save this data is cached on an internal collection for that.

If you can send a good implementation that covers that and keeps performance, I will be happy to accept :)

[airy]

thanks for your confidence but i am not good enough on mongodb to write implementation :)

i found some usefull information here : http://www.mongodb.org/display/DOCS/MapReduce#MapReduce-Overview

there is a query option as well as limit and scope.

[wilkerlucio]

In fact, the cached table data is created using the MapReduce, you can take a look here: https://github.com/wilkerlucio/mongoid_taggable/blob/master/lib/mongoid/taggable.rb#L78

It should not be so hard to finish the implementation of query methods to suporte the context, but need to dig a little on MongoId source to see how internally you can use the current search context.

Currently Im most dedicated on other projects, but if you have some time, I mean read the MongoId source and do this implementation will be a good learning experience :)

[hayksaakian]

any progress on this front?

scoped/contextual tags would be really convenient

+1

[loicginoux]

For now you can use the aggregate method from mongoid. This works for me:

# you define here your query
criteria = YourModel.where(my_attribute: "a value")
# and you count the tags
result = YourModel.collection.aggregate([
    { "$match" => criteria.selector},
    { "$project" => YourModel.only(:tags_array).options.fields},
    { "$unwind" => "$tags_array"},
    { "$group" =>  { "_id" => "$tags_array", 
      "count" => { "$sum" => 1 }
    }}
  ])

p results # print [{"_id"=>"bee", "count"=>1},  {"_id"=>"ant", "count"=>1},  {"_id"=>"food", "count"=>2}]

and you can add a sorting at the end of the pipeline if you want them by counts:

{ "$sort" => {"count" => -1,"_id" => 1}}