OData4

NOTE: starting with version 0.9.1, this gem has been renamed to frodata. All further development will continue in the new repository. This repository will stay up for historical reasons, but now new features will be added to it.

The OData4 gem provides a simple wrapper around the OData Version 4.0 API protocol. It has the ability to automatically inspect compliant APIs and expose the relevant Ruby objects dynamically. It also provides a set of code generation tools for quickly bootstrapping more custom service libraries.

This gem supports OData Version 4.0. Support for older versions is not a goal.

If you need a gem to integration with OData Version 3, you can use James Thompson's original OData gem, upon which this gem is based.

Add this line to your application's Gemfile:

gem 'odata4'

And then execute:

$ bundle

Or install it yourself as:

$ gem install odata4

The OData4 gem provides a number of core classes, the two most basic ones are the OData4::Service and the OData4::ServiceRegistry. The only time you will need to worry about the OData4::ServiceRegistry is when you have multiple OData4 services you are interacting with that you want to keep straight easily. The nice thing about OData4::Service is that it automatically registers with the registry on creation, so there is no manual interaction with the registry necessary.

To create an OData4::Service simply provide the location of a service endpoint to it like this:

  OData4::Service.new('http://services.odata.org/V4/OData/OData.svc')

You may also provide an options hash after the URL. It is suggested that you supply a name for the service via this hash like so:

  OData4::Service.new('http://services.odata.org/V4/OData/OData.svc', name: 'ODataDemo')

For more information regarding available options and how to configure a service instance, refer to Service Configuration below.

This one call will setup the service and allow for the discovery of everything the other parts of the OData4 gem need to function. The two methods you will want to remember from OData4::Service are #service_url and #name. Both of these methods are available on instances and will allow for lookup in the OData4::ServiceRegistry, should you need it.

Using either the service URL or the name provided as an option when creating an OData4::Service will allow for quick lookup in the OData4::ServiceRegistry like such:

  OData4::ServiceRegistry['http://services.odata.org/V4/OData/OData.svc']
  OData4::ServiceRegistry['ODataDemo']

Both of the above calls would retrieve the same service from the registry. At the moment there is no protection against name collisions provided in OData4::ServiceRegistry. So, looking up services by their service URL is the most exact method, but lookup by name is provided for convenience.

Typically the metadata file of a service can be quite large. You can speed your load time by forcing the service to load the metadata from a file rather than a URL. This is only recommended for testing purposes, as the metadata file can change.

  service = OData4::Service.new('http://services.odata.org/V4/OData/OData.svc', {
    name: 'ODataDemo',
    metadata_file: "metadata.xml",
  })

The OData protocol does not deal with authentication and authorization at all, nor does it need to, since HTTP already provides many different options for this, such as HTTP Basic or token authorization. Hence, this gem does not implement any special authentication mechanisms either, and relies on the underlying HTTP library (Faraday) to take care of this.

You can customize request headers with the :connection option key. This allows you to e.g. set custom headers (such as Authorization) that may be required by your service.

  service = OData4::Service.new('http://services.odata.org/V4/OData/OData.svc', {
    name: 'ODataDemo',
    connection: {
      headers: {
        "Authorization" => "Bearer #{access_token}"
      }
    }
  })

You may also set up authorization by directly accessing the underlying Faraday::Connection object (as explained in Advanced Customization below). This allows you to make use of Faraday's authentication helpers, such as basic_auth or token_auth.

For instance, if your service requires HTTP basic authentication:

  service = OData4::Service.new('http://services.odata.org/V4/OData/OData.svc', {
    name: 'ODataDemo'
  })
  service.connection.basic_auth('username', 'password')

You may also use these helpers when passing a block to the constructor (see second example below).

Under the hood, the gem uses the Faraday HTTP library to provide flexible integration of various Ruby HTTP backends.

There are several ways to access the underlying Faraday::Connection:

If you already have a Faraday::Connection instance that you want the service to use, you can simply pass it to the constructor instead of the service URL as first parameter. In this case, you'll be setting the service URL on the connection object, as shown below:

  conn = Faraday.new('http://services.odata.org/V4/OData/OData.svc') do |conn|
    # ... customize connection ...
  end

  service = OData4::Service.new(conn, name: 'ODataDemo')

NOTE: if you use this method, any options set via the :connection options key will be ignored.

Alternatively, the connection object is also yielded by the constructor, so you may customize it by passing a block argument. For instance, if you wanted to use Typhoeus as your HTTP library:

  service = OData4::Service.new('http://services.odata.org/V4/OData/OData.svc', {
    name: 'ODataDemo'
  }) do |conn|
    conn.adapter :typhoeus
  end

IMPORTANT

Please be aware that if you use this method to customize the connection, you must ALWAYS specify an adapter:

  service = OData4::Service.new('http://services.odata.org/V4/OData/OData.svc', {
    name: 'ODataDemo'
  }) do |conn|
    conn.basic_auth('username', 'password')
    conn.adapter Faraday.default_adapter
  end

Otherwise, your requests WILL fail!

Once instantiated, you can request various information about the service, such as the names and types of entity sets it exposes, or the names of the entity types (and custom datatypes) it defines.

For example:

Get a list of available entity types

  service.entity_types
  # => [
  #   "ODataDemo.Product",
  #   "ODataDemo.FeaturedProduct",
  #   "ODataDemo.ProductDetail",
  #   "ODataDemo.Category",
  #   "ODataDemo.Supplier",
  #   "ODataDemo.Person",
  #   "ODataDemo.Customer",
  #   "ODataDemo.Employee",
  #   "ODataDemo.PersonDetail",
  #   "ODataDemo.Advertisement"
  # ]

Get a list of entity sets

  service.entity_sets
  # => {
  #   "Products"       => "ODataDemo.Product",
  #   "ProductDetails" => "ODataDemo.ProductDetail",
  #   "Categories"     => "ODataDemo.Category",
  #   "Suppliers"      => "ODataDemo.Supplier",
  #   "Persons"        => "ODataDemo.Person",
  #   "PersonDetails"  => "ODataDemo.PersonDetail",
  #   "Advertisements" => "ODataDemo.Advertisement"
  # }

Get a list of complex types

  service.complex_types
  # => ["ODataDemo.Address"]

Get a list of enum types

  service.enum_types
  # => ["ODataDemo.ProductStatus"]

For more examples, refer to usage_example_specs.rb.

When it comes to reading data from an OData4 service the most typical way will be via OData4::EntitySet instances. Under normal circumstances you should never need to worry about an OData4::EntitySet directly. For example, to get an OData4::EntitySet for the products in the ODataDemo service simply access the entity set through the service like this:

  service = OData4::Service.new('http://services.odata.org/V4/OData/OData.svc')
  products = service['ProductsSet'] # => OData4::EntitySet

OData4::EntitySet instances implement the Enumerable module, meaning you can work with them very naturally, like this:

  products.each do |entity|
    entity # => OData4::Entity for type Product
  end

You can get a list of all your entity sets like this:

  service.entity_sets

Some versions of Microsoft CRM do not support count.

  products.count

You can you the following methods to grab a collection of Entities:

  products.each do |entity|
    ...
  end

The first entity object returns a single entity object.

  products.first

first(x) returns an array of entity objects.

  products.first(x)

  service['ProductsSet']['<primary key of entity>']

With certain navigation properties expanded (i.e. eagerly loaded):

  # Eagerly load a single navigation property
  service['ProductsSet', expand: 'Categories']

  # Eagerly load multiple navigation properties
  service['ProductsSet', expand: ['Categories', 'Supplier']]

  # Eagerly load ALL navigation properties
  service['ProductsSet', expand: :all]

OData4::Entity instances represent individual entities, or records, in a given service. They are returned primarily through interaction with instances of OData4::EntitySet. You can access individual properties on an OData4::Entity like so:

  product = products.first # => OData4::Entity
  product['Name']  # => 'Bread'
  product['Price'] # => 2.5 (Float)

Individual properties on an OData4::Entity are automatically typecast by the gem, so you don't have to worry about too much when working with entities. The way this is implemented internally guarantees that an OData4::Entity is always ready to save back to the service or OData4::EntitySet, which you do like so:

  service['Products'] << product # Write back to the service
  products << product        # Write back to the Entity Set

You can get a list of all your entities like this:

  service.entity_types

Reading, parsing and instantiating all properties of an entity can add up to a significant amount of time, particularly for those entities with a large number of properties. To speed this process up all properties are lazy loaded. Which means it will store the name of the property, but will not parse and instantiate the property until you want to use it.

You can find all the property names of your entity with

  product.property_names

You can grab the parsed value of the property as follows:

  product["Name"]

or, you can get a hold of the property class instance using

  product.get_property("Name")

This will parse and instantiate the property if it hasn't done so yet.

By default, we use strict property validation, meaning that any property validation errors in the data will raise an exception. However, you may encounter OData implementations in the wild that break the specs in strange and surprising ways (shocking, I know!).

Since it's often better to get some data instead of nothing at all, you can optionally make the property validation lenient. Simply add strict: false to the service constructor options. In this mode, any property validation error will log a warning instead of raising an exception. The corresponding property value will be nil (even if the property is declared as not allowing NULL values).

  service = OData4::Service.new('http://services.odata.org/V4/OData/OData.svc', strict: false)
  # -- alternatively, for an existing service instance --
  service.options[:strict] = false

OData4::Query instances form the base for finding specific entities within an OData4::EntitySet. A query object exposes a number of capabilities based on the System Query Options provided for in the OData V4.0 specification. Below is just a partial example of what is possible:

  query = service['Products'].query
  query.where(query[:Price].lt(15))
  query.where(query[:Rating].gt(3))
  query.limit(3)
  query.skip(2)
  query.order_by("Name")
  query.select("Name,CreatedBy")
  query.inline_count
  results = query.execute
  results.each {|product| puts product['Name']}

The process of querying is kept purposely verbose to allow for lazy behavior to be implemented at higher layers. Internally, OData4::Query relies on the OData4::Query::Criteria for the way the where method works. You should refer to the published RubyDocs for full details on the various capabilities:

• OData4::Query
• OData4::Query::Criteria

[x] Lenient property validation [ ] Write support (create/update/delete) [ ] Support for invoking Operations (Functions/Actions) [ ] Property facets [ ] Annotations

1. Fork it (https://github.com/[my-github-username]/odata/fork)
2. Create your feature branch (git checkout -b my-new-feature)
3. Commit your changes (git commit -am 'Add some feature')
4. Push to the branch (git push origin my-new-feature)
5. Create a new Pull Request

Many thanks go to James Thompson, who wrote the original OData (Version 3.0) gem.

Also, I would like to thank W+R Studios for generously allowing me to work on Open Source software like this. If you want to work on interesting challenges with an awesome team, check out our open positions.