RSolr

A simple, extensible Ruby client for Apache Solr.

Documentation

The code docs www.rubydoc.info/gems/rsolr

Installation:

gem install rsolr

Example:

require 'rsolr'

# Direct connection
solr = RSolr.connect :url => 'http://solrserver.com'

# Connecting over a proxy server
solr = RSolr.connect :url => 'http://solrserver.com', :proxy=>'http://user:[email protected]:8080'

# Using an alternate Faraday adapter
solr = RSolr.connect :url => 'http://solrserver.com', :adapter => :em_http

# Using a custom Faraday connection
conn = Faraday.new do |faraday|
  faraday.response :logger                  # log requests to STDOUT
  faraday.adapter  Faraday.default_adapter  # make requests with Net::HTTP
end
solr = RSolr.connect conn, :url => 'http://solrserver.com'

# send a request to /select
response = solr.get 'select', :params => {:q => '*:*'}

# send a request to /catalog
response = solr.get 'catalog', :params => {:q => '*:*'}

When the Solr :wt is :ruby, then the response will be a Hash. This Hash is the same object returned by Solr, but evaluated as Ruby. If the :wt is not :ruby, then the response will be a String.

The response also exposes 2 attribute readers (for any :wt value), :request and :response. Both are Hash objects with symbolized keys.

The :request attribute contains the original request context. You can use this for debugging or logging. Some of the keys this object contains are :uri, :query, :method etc..

The :response attribute contains the original response. This object contains the :status, :body and :headers keys.

Request formats

By default, RSolr uses the Solr JSON command format for all requests.

RSolr.connect :url => 'http://solrserver.com', update_format: :json # the default
# or
RSolr.connect :url => 'http://solrserver.com', update_format: :xml

Timeouts

The read and connect timeout settings can be set when creating a new instance of RSolr, and will be passed on to underlying Faraday instance:

solr = RSolr.connect(:timeout => 120, :open_timeout => 120)

Retry 503s

A 503 is usually a temporary error which RSolr may retry if requested. You may specify the number of retry attempts with the :retry_503 option.

Only requests which specify a Retry-After header will be retried, after waiting the indicated retry interval, otherwise RSolr will treat the request as a 500. You may specify a maximum Retry-After interval to wait with the :retry_after_limit option (default: one second).

solr = RSolr.connect(:retry_503 => 1, :retry_after_limit => 1)

For additional control, consider using a custom Faraday connection (see above) using its ‘retry` middleware.

Querying

Use the #get / #post method to send search requests to the /select handler:

response = solr.get 'select', :params => {
  :q=>'washington',
  :start=>0,
  :rows=>10
}
response["response"]["docs"].each{|doc| puts doc["id"] }

The :params sent into the method are sent to Solr as-is, which is to say they are converted to Solr url style, but no special mapping is used. When an array is used, multiple parameters *with the same name* are generated for the Solr query. Example:

solr.get 'select', :params => {:q=>'roses', :fq=>['red', 'violet']}

The above statement generates this Solr query:

select?q=roses&fq=red&fq=violet

To paginate through a set of Solr documents, use the paginate method:

solr.paginate 1, 10, "select", :params => {:q => "test"}

The first argument is the current page, the second is how many documents to return for each page. In other words, “page” is the “start” Solr param and “per-page” is the “rows” Solr param.

The paginate method returns WillPaginate ready “docs” objects, so for example in a Rails application, paginating is as simple as:

<%= will_paginate @solr_response["response"]["docs"] %>

Method Missing

The RSolr::Client class also uses method_missing for setting the request handler/path:

solr.paintings :params => {:q=>'roses', :fq=>['red', 'violet']}

This is sent to Solr as:

paintings?q=roses&fq=red&fq=violet

This works with pagination as well:

solr.paginate_paintings 1, 10, {:q=>'roses', :fq=>['red', 'violet']}

Using POST for Search Queries

There may be cases where the query string is too long for a GET request. RSolr solves this issue by converting hash objects into form-encoded strings:

response = solr.music :data => {:q => "*:*"}

The :data hash is serialized as a form-encoded query string, and the correct content-type headers are sent along to Solr.

Sending HEAD Requests

There may be cases where you’d like to send a HEAD request to Solr:

solr.head("admin/ping").response[:status] == 200

Sending HTTP Headers

Solr responds to the request headers listed here: wiki.apache.org/solr/SolrAndHTTPCaches To send header information to Solr using RSolr, just use the :headers option:

response = solr.head "admin/ping", :headers => {"Cache-Control" => "If-None-Match"}

Building a Request

RSolr::Client provides a method for building a request context, which can be useful for debugging or logging etc.:

request_context = solr.build_request "select", :data => {:q => "*:*"}, :method => :post, :headers => {}

To build a paginated request use build_paginated_request:

request_context = solr.build_paginated_request 1, 10, "select", ...

Updating Solr

Updating is done using native Ruby objects. Hashes are used for single documents and arrays are used for a collection of documents (hashes). These objects get turned into simple XML “messages”. Raw XML strings can also be used.

Single document via #add

solr.add :id=>1, :price=>1.00

Multiple documents via #add

documents = [{:id=>1, :price=>1.00}, {:id=>2, :price=>10.50}]
solr.add documents

The optional :add_attributes hash can also be used to set Solr “add” document attributes:

solr.add documents, :add_attributes => {:commitWithin => 10}

Raw commands via #update

solr.update data: '<commit/>', headers: { 'Content-Type' => 'text/xml' }
solr.update data: { optimize: true }.to_json, headers: { 'Content-Type' => 'application/json' }

When adding, you can also supply “add” xml element attributes and/or a block for manipulating other “add” related elements (docs and fields) by calling the xml method directly:

doc = {:id=>1, :price=>1.00}
add_attributes = {:allowDups=>false, :commitWithin=>10}
add_xml = solr.xml.add(doc, add_attributes) do |doc|
  # boost each document
  doc.attrs[:boost] = 1.5
  # boost the price field:
  doc.field_by_name(:price).attrs[:boost] = 2.0
end

Now the “add_xml” object can be sent to Solr like:

solr.update :data => add_xml

Deleting

Delete by id

solr.delete_by_id 1

or an array of ids

solr.delete_by_id [1, 2, 3, 4]

Delete by query:

solr.delete_by_query 'price:1.00'

Delete by array of queries

solr.delete_by_query ['price:1.00', 'price:10.00']

Commit / Optimize

solr.commit, :commit_attributes => {}
solr.optimize, :optimize_attributes => {}

Response Formats

The default response format is Ruby. When the :wt param is set to :ruby, the response is eval’d resulting in a Hash. You can get a raw response by setting the :wt to “ruby” - notice, the string – not a symbol. RSolr will eval the Ruby string ONLY if the :wt value is :ruby. All other response formats are available as expected, :wt=>‘xml’ etc..

Evaluated Ruby:

solr.get 'select', :params => {:wt => :ruby} # notice :ruby is a Symbol

Raw Ruby:

solr.get 'select', :params => {:wt => 'ruby'} # notice 'ruby' is a String

XML:

solr.get 'select', :params => {:wt => :xml}

JSON (default):

solr.get 'select', :params => {:wt => :json}

RSolr Google Group – The RSolr discussion group
rsolr-ext – An extension kit for RSolr
rsolr-direct – JRuby direct connection for RSolr
rsolr-nokogiri – Gives RSolr Nokogiri for XML generation.
SunSpot – An awesome Solr DSL, built with RSolr
Blacklight – A “next generation” Library OPAC, built with RSolr
java_bin – Provides javabin/binary parsing for RSolr
Solr – The Apache Solr project
solr-ruby – The original Solr Ruby Gem!

Note on Patches/Pull Requests

Fork the project.
Make your feature addition or bug fix.
Add tests for it. This is important so I don’t break it in a future version unintentionally.
Commit, do not mess with rakefile, version, or history (if you want to have your own version, that is fine but bump version in a commit by itself I can ignore when I pull)
Send me a pull request. Bonus points for topic branches.

Contributors

Nathan Witmer
Magnus Bergmark
shima
Randy Souza
Mat Brown
Jeremy Hinegardner
Denis Goeury
shairon toledo
Rob Di Marco
Peter Kieltyka
Mike Perham
Lucas Souza
Dmitry Lihachev
Antoine Latter
Naomi Dushay

Author

Matt Mitchell <[email protected]>