I'm using the elasticsearch-rails gem and the elasticsearch-model gem and writing a query that happens to be really huge just because of the way the gem accepts queries.
The query itself isn't very long, but it's the filters that are very, very long, and I need to pass variables in to filter out the results correctly. Here is an example:
def search_for(input, question_id, tag_id)
query = {
:query => {
:filtered => {
:query => {
:match => {
:content => input
}
},
:filter => {
:bool => {
:must => [
{
# another nested bool with should
},
{
# another nested bool with must for question_id
},
{
# another nested bool with must for tag_id
}
]
}
}
}
}
}
User.search(query) # provided by elasticsearch-model gem
end
For brevity's sake, I've omitted the other nested bools, but as you can imagine, this can get quite long quite fast.
Does anyone have any ideas on how to store this? I was thinking of a yml file, but it seems wrong especially because I need to pass in question_id and tag_id. Any other ideas?
If anyone is familiar with those gems and knows whether the gem's search method accepts other formats, I'd like to know that, too. Looks to me that it just wants something that can turn into a hash.
I think using a method is fine. I would separate the searching from the query:
def query_for(input, question_id, tag_id)
query = {
:query => {
...
end
search query_for(input, question_id, tag_id)
Also, I see that this search functionality is in the User model, but I wonder if it is belongs there. Would it make more sense to have a Search or Query model?
Related
I have a model Event that is connected to MongoDB using Mongoid:
class Event
include Mongoid::Document
include Mongoid::Timestamps
field :user_name, type: String
field :action, type: String
field :ip_address, type: String
scope :recent, -> { where(:created_at.gte => 1.month.ago) }
end
Usually when I use ActiveRecord, I can do something like this to group results:
#action_counts = Event.group('action').where(:user_name =>"my_name").recent.count
And I get results with the following format:
{"action_1"=>46, "action_2"=>36, "action_3"=>41, "action_4"=>40, "action_5"=>37}
What is the best way to do the same thing with Mongoid?
Thanks in advance
I think you'll have to use map/reduce to do that. Look at this SO question for more details:
Mongoid Group By or MongoDb group by in rails
Otherwise, you can simply use the group_by method from Enumerable. Less efficient, but it should do the trick unless you have hundreds of thousands documents.
EDIT: Example of using map/reduce in this case
I'm not really familiar with it but by reading the docs and playing around I couldn't reproduce the exact same hash you want but try this:
def self.count_and_group_by_action
map = %Q{
function() {
key = this.action;
value = {count: 1};
emit(key, value);
# emit a new document {"_id" => "action", "value" => {count: 1}}
# for each input document our scope is applied to
}
}
# the idea now is to "flatten" the emitted documents that
# have the same key. Good, but we need to do something with the values
reduce = %Q{
function(key, values) {
var reducedValue = {count: 0};
# we prepare a reducedValue
# we then loop through the values associated to the same key,
# in this case, the 'action' name
values.forEach(function(value) {
reducedValue.count += value.count; # we increment the reducedValue - thx captain obvious
});
# and return the 'reduced' value for that key,
# an 'aggregate' of all the values associated to the same key
return reducedValue;
}
}
self.map_reduce(map, reduce).out(inline: true)
# we apply the map_reduce functions
# inline: true is because we don't need to store the results in a collection
# we just need a hash
end
So when you call:
Event.where(:user_name =>"my_name").recent.count_and_group_by_action
It should return something like:
[{ "_id" => "action1", "value" => { "count" => 20 }}, { "_id" => "action2" , "value" => { "count" => 10 }}]
Disclaimer: I'm no mongodb nor mongoid specialist, I've based my example on what I could find in the referenced SO question and Mongodb/Mongoid documentation online, any suggestion to make this better would be appreciated.
Resources:
http://docs.mongodb.org/manual/core/map-reduce/
http://mongoid.org/en/mongoid/docs/querying.html#map_reduce
Mongoid Group By or MongoDb group by in rails
Let's say I have an app that handles a TODO list. The list has finished and unfinished items. Now I want to add two virtual attributes to the list object; the count of finished and unfinished items in the list. I also need these to be displayed in the json output.
I have two methods in my model which fetches the unfinished/finished items:
def unfinished_items
self.items.where("status = ?", false)
end
def finished_items
self.items.where("status = ?", true)
end
So, how can I get the count of these two methods in my json output?
I'm using Rails 3.1
The serialization of objects in Rails has two steps:
First, as_json is called to convert the object to a simplified Hash.
Then, to_json is called on the as_json return value to get the final JSON string.
You generally want to leave to_json alone so all you need to do is add your own as_json implementation sort of like this:
def as_json(options = { })
# just in case someone says as_json(nil) and bypasses
# our default...
super((options || { }).merge({
:methods => [:finished_items, :unfinished_items]
}))
end
You could also do it like this:
def as_json(options = { })
h = super(options)
h[:finished] = finished_items
h[:unfinished] = unfinished_items
h
end
if you wanted to use different names for the method-backed values.
If you care about XML and JSON, have a look at serializable_hash.
With Rails 4, you can do the following -
render json: #my_object.to_json(:methods => [:finished_items, :unfinished_items])
Hope this helps somebody who is on the later / latest version
Another way to do this is add this to your model:
def attributes
super.merge({'unfinished' => unfinished_items, 'finished' => finished_items})
end
This would also automatically work for xml serialization.
http://api.rubyonrails.org/classes/ActiveModel/Serialization.html
Be aware though, you might want use strings for the keys, since the method can not deal with symbols when sorting the keys in rails 3. But it is not sorted in rails 4, so there shouldn't be a problem anymore.
just close all of your data into one hash, like
render json: {items: items, finished: finished, unfinished: unfinished}
I just thought I'd provide this answer for anyone like myself, who was trying to integrate this into an existing as_json block:
def as_json(options={})
super(:only => [:id, :longitude, :latitude],
:include => {
:users => {:only => [:id]}
}
).merge({:premium => premium?})
Just tack .merge({}) on to the end of your super()
This will do, without having to do some ugly overridings. If you got a model List for example, you can put this in your controller:
render json: list.attributes.merge({
finished_items: list.finished_items,
unfinished_items: list.unfinished_items
})
As Aswin listed above, :methods will enable you to return a specific model's method/function as a json attribute, in case you have complex assosiations this will do the trick since it will add functions to the existing model/assossiations :D it will work like a charm if you dont want to redefine as_json
Check this code, and please notice how i'm using :methods as well as :include [N+Query is not even an option ;)]
render json: #YOUR_MODEL.to_json(:methods => [:method_1, :method_2], :include => [:company, :surveys, :customer => {:include => [:user]}])
Overwritting as_json function will be way harder in this scenario (specially because you have to add the :include assossiations manually :/
def as_json(options = { })
end
If you want to render an array of objects with their virtual attributes, you can use
render json: many_users.as_json(methods: [:first_name, :last_name])
where first_name and last_name are virtual attributes defined on your model
I currently have the following:
#threads = current_user.threads.includes(:user, :thread_members)
I then take threads and do the following:
#threads.each do |thread|
thread_members = thread.thread_members_active(current_user)
#threadList << {
:id => thread.id,
:uuid => thread.uuid,
:user_id => thread.user.id,
:last_activity_at => thread.last_activity_at,
:user_count => thread_members.length,
:user_photos => thread_members.collect { |thread_member|
{
:id => thread_member.user.id,
:photo => thread_member.user.photo(:thumb),
:name => thread_member.user.full_name
}
},
:caption => thread.caption
}
end
The issue here is that every EACH loop, rails is hitting the DB for the same basic records. Rails sees to be caching as I see CACHE in the log but it's mighty messy. Leaves me wishing I could do some type of includes so there wasn't so many db requests.
Any ideas on how this can be optimized? Something around including all the users in one db hit?
Thanks
If you don't want any DB queries in the loop, you have to define everything that's used there in the named associations that are included, so instead of a thread_members_active method you'd define a thread_members_active association which has the same behavior. Note that the association also needs to use includes on user. Can't give you more right now, but maybe that helps a bit.
Edit: Check out the "Eager loading of associations" part of this doc:
http://api.rubyonrails.org/classes/ActiveRecord/Associations/ClassMethods.html
Need some help with how to use atomic modifiers on an embedded document.
To illustrate, let's assume I've got a collection that looks like this.
Posts Collection
{
"_id" : ObjectId("blah"),
"title" : "Some title",
"comments" : [
{
"_id" : ObjectId("bleh"),
"text" : "Some comment text",
"score" : 0,
"voters" : []
}
]
}
What I'm looking to do with MongoMapper/MongoDB is perform an atomic update on a specific comment within a post document.
Something like:
class Comment
include MongoMapper::EmbeddedDocument
# Other stuff...
# For the current comment that doesn't have the current user voting, increment the vote score and add that user to the voters array so they can't vote again
def upvote!(user_id)
collection.update({"comments._id" => post_id, "comments.voters" => {"$ne" => user_id}},
{"$inc" => {"comments.score" => 1}, "$push" => {"comments.voters" => user_id}})
end
end
That's basically what I have now and it isn't working at all (nothing gets updated). Ideally, I'd also want to reload the document / embedded document but it seems as though there may not be a way to do this using MongoMapper's embedded document. Any ideas as to what I'm doing wrong?
Got this working for anyone that's interested. Two things I was missing
Using $elemMatch to search objects within an array that need to satisfy two conditions (such as _id = "" AND voters DOES NOT contain the user_id)
Using the $ operator on the $inc and $push operations to ensure I'm modifying the specific object that's referenced by my query.
def upvote!(user_id)
# Use the Ruby Mongo driver to make a direct call to collection.update
collection.update(
{
'meanings' => {
'$elemMatch' => {
'_id' => self.id,
'voters' => {'$ne' => user_id}
}
}
},
{
'$inc' => { 'meanings.$.votes' => 1 },
'$push' => { 'meanings.$.voters' => user_id }
})
end
I'm returning a complex result of indeterminate size that I will need to handle again and again, so I'm wondering what is a good way to package it?
something like this
loop>>>
#results = { external_id => { :name => name, :type => type } }
or
#results = [ { :external_id => external_id, :name => name, :type => type } ]
or?
end>>>>
and if it ends up being a hash of a hash, do i just use merge?
Thanks
I ended up with an array of a hash... and it works fine.