Sails-hook-solr NPM

Sails Hook Solr

This hook allows any Sails.js Model defined to easily integrate to a Solr http://lucene.apache.org/solr/ server by exposing functionalities built on top of the http://lbdremy.github.io/solr-node-client/ as Model's class methods and Model's properties.

Solr Configuration

Bellow you can see the default config.client object which is used to create a client instance in order to connect to Solr server.

Any parameter can be override by setting them in your own project.

Look for createClient() in http://lbdremy.github.io/solr-node-client/code/solr.js.html for a list of the available parameters.

    defaults: {
      solr: {
        client: {
          host: 'localhost', // - IP address or host address of the Solr server
          port: '8983', // - port of the Solr server
          path: '/solr', // - root path of all requests
          secure: false, // - if true HTTPS will be used instead of HTTP
          bigint: false // - if true JSONbig serializer/deserializer will be used instead
        }
      }
    },

During the creation of the sorl clients, this hook sets the core property (not listed above) as the Model's name taken from the sails.models object, so each solr client will connect to its corresponding sorl core.

Solr Client

The solr client is created and exposed for all Models so you can harness the full power of the http://lbdremy.github.io/solr-node-client/.

var sorlClient =  Model.solrClient

With solrClient in hands you can perform any method provided by solr-node-client.

Class Methods

In order to speed up the integration of the Sails.js projects and Solr, this hook exposes a few useful Class Methods to the models:

updateSolrQuery(opts, callback)

This method retrieves all records from a given query (opts), populates all associations and update these documents to the solr corresponding index/core.

This hook uses find(opts).populateAll() to fetch data.

@param {object} opts

More information on available query parameters in http://sailsjs.org/#!/documentation/concepts/ORM/Querylanguage.html?q=query-language-basics

@param {function} callback

The callback function is treated as a standard callback node function receiving an err and res objects.

Example of successful response res:

{
	responseHeader: {
		status: 0,
		QTime: 61
	},
	docsUpdated: 10
}

Below an example on how to proper call this method:

Model.updateSolrQuery({
	where: { name: 'foo' },
	skip: 20,
	limit: 10
}, function(err, res){
	if ( err ) {
		console.log(err)
	}
	console.log(res.docsUpdated)
});

updateToSolr(records, cb)

This method updates a set of records to the solr corresponding index/core.

The motivation behind this method is to give the user the ability to built custom queries taking the most advantage of its database implementation and adapter to gather the intended data.

Let's say, for instance, that the user is not satisfied by the performance of the default Waterline find() implementation and wants to use the query or native methods to build custom queries on SQL and NoSQL databases. One would simply gather the data as desired and pass it on to this method to update the corresponding index.

Below an example on how to proper call this method:

Model.updateToSolr(docsToUpdate, function(err, res){
	if ( err ) {
		console.log(err)
	}
	console.log(res.docsUpdated)
});

deleteSolrQuery(query, cb)

// TODO

deleteSolrById(id, cb)

// TODO

Schemas

But wait ?! How does the sails-hook-solr knows which fields we want to upload to solr index? One might only want to upload a few fields to relieve the burden of its solr server, or even denormalise some associations copying only the informations needed from the relation.

To better integrate with solr and give more power to the user, the hook apply an adapted concept of solr called Schema.

The idea of the schema configuration is to tell the hook how does one want to map the fields from the application model to solr index.

As seen bellow this configuration is set in config.solr.schemas[Model].

Attention !!! In order to use both class methods above, you have got to define a schema for the desired model, otherwise an error will be thrown

module.exports.solr = {
	schemas: {
		model: {
			id: {},
			field1: {},
			field2_field: { copy: true, from: 'othermodel', field: 'field' },
			field3_more: { copy: true, from: 'othermodel', field: function( record, model, callback ){
				sails.models[model]
				.findOneById(record.field.id)
				.exec(function (err, r){
					if ( err ) {
						return callback(err);
					}

					if ( r && r.course && r.course.length ) {
						return callback(null, _.last( _.sortBy(r.course, 'endyear') ).name);
					}

					callback(null, null);
				})
			}},
		}
	}
}

The content of a model's schema is an key-value {Object} with the bellow characteristics:

Key - Valid field of the model - Non-existing field if the key's pair has copy property set to true
Value {Object} - copy - Paired key is not a valid model field, and the content is going to be taken from from field - from - It is the field where the content will be taken from and is required if copy is set to true. - field - {String} - It is the field that determines which field will be retrieve for associations (modeland collection) - {Function} - It is an asynchronous function to be called upon assignment process which receives as parameter current record, Model's name and callback function