Added support for API slugs #14344

abellotti · 2017-03-15T16:41:41Z

added slug method on application record instances
returns ":collection/:id" if instance class matches a collection class or any of its descendants, nil otherwise.
declared as a virtual_column so it's accessible from API.
Added specs

Closes Issue #11635

abellotti · 2017-03-15T16:42:12Z

/cc @mkanoor @gmcculloug @gtanzillo as promised.

@Fryguy please review. Thanks!!

gmcculloug

A few comments, but overall this looks good. Maybe add a negative test for a new record that does not have an ID value yet for a valid class.

gmcculloug · 2017-03-15T17:38:54Z

app/models/application_record.rb

+  # API Support
+  virtual_column :slug, :type => :string
+
+  def slug


@Fryguy Should this logic be moved into a separate file and included here like the Ar* modules above.

Is slug a descriptive enough method name. Is url_slug or href_slug more useful? Just a thought.

I prefer url_slug, myself.

gmcculloug · 2017-03-15T17:47:47Z

app/models/application_record.rb

+  virtual_column :slug, :type => :string
+
+  def slug
+    collection = Api::CollectionConfig.new.name_for_subklass(self.class)


Suggest returning early if id is nil so you do not spend time in name_for_subklass for new records that do not have an ID yet.

def slug return nil unless id collection = Api::CollectionConfig.new.name_for_subklass(self.class) "#{collection}/#{id}" if collection end

gmcculloug · 2017-03-15T19:05:31Z

cc @bzwei

mkanoor · 2017-03-17T13:32:57Z

@abellotti would it make sense to add another function like fetch_by_slug that fetches the object based on the slug that we could use internally.

A typical use case would be we have a slug for a provider that looks like providers/99 and we can fetch the ExtManagementSystems.

This code might already exist that the REST API uses to load the AR object

abellotti · 2017-03-17T14:01:53Z

We can provide something like that in a separate PR. API internally does not use slugs, so we'd write something similar to resource_search but simply takes the slug.

abellotti · 2017-03-17T14:58:37Z

/cc @gmcculloug @gtanzillo I went with href_slug since it pertains to the unique identifier of the href presented by the API. Thanks.

gmcculloug · 2017-03-17T15:03:52Z

Thanks, looks good. @Fryguy Any thoughts here?

Fryguy · 2017-03-17T17:43:36Z

lib/api/collection_config.rb

@@ -73,6 +73,13 @@ def name_for_klass(resource_klass)
      @cfg.detect { |_, spec| spec[:klass] == resource_klass.name }.try(:first)
    end

+    def name_for_subklass(resource_klass)


I don't like subklass bit. you can just say subclass here and it won't break anything (as well as resource_class for that matter). klass is only necessary when standalone because it's a keyword.

Fryguy · 2017-03-17T17:47:01Z

lib/extensions/ar_slug.rb

+  extend ActiveSupport::Concern
+
+  included do
+    virtual_column :href_slug, :type => :string


I feel like we should namespace ourselves here and call it api_href_slug and the module itself ArApiSlug or ArApiHrefSlug or something like that, because we get slugs in the providers world for other purposes and it can get confusing. e.g. ManageIQ::Providers::Redhat::Instance#href_slug. Is that the ManageIQ href slug or is that the native RHV href slug?

What do you think?

Interesting point, I wonder if the namespace should go beyond api as they have API's too. Maybe we need Miq. i.e. "miq_api_href_slug" or maybe "miq_href_slug" is sufficient in this case.

miq_href_slug 👍

after discussing with @Fryguy, we're keeping it as href_slug as it's the most relevant/natural name w.r.t. slug for the href attribute, as compared to href and miq_href_slug.

In a separate PR though (replication), we should check/enforce that no href or href_slug attributes are named as such as it would clash with the API.

@abellotti @Fryguy Did you also discuss the module name which is mentioned in this chain? Are you keeping ArSlug or should that be changing in this PR?

haven't discussed that. I'm game to changing it if that's desired, is there a preference for ArHrefSlug maybe ? @Fryguy ?

Yeah, I think make it match the method, so ArHrefSlug

Fryguy · 2017-03-17T17:51:43Z

lib/extensions/ar_slug.rb

+    def href_slug
+      return unless id
+      collection = Api::CollectionConfig.new.name_for_subklass(self.class)
+      "#{collection}/#{id}" if collection


I feel like this should be a builder method in Api namespace, that we just call here. Thinking something like...

# here def href_slug Api.build_href_slug(self.class, id) end # there Not sure the best module, but that's up to you module Api def self.build_href_slug(klass, id) return unless id collection = Api::CollectionConfig.new.name_for_subklass(klass) "#{collection}/#{id}" if collection end end

This way the knowledge for building slugs generically belongs with the API and the API specs, irrespective of having an AR instance (i.e. it can also be called (and tested) with only a type/id pair and not an instance). And the knowledge of how to build it for AR instance then is just a simple call with some defaults from the instance.

Additionally, then name_for_subklass actually just becomes a private method

Does this assume that all objects that are configured in the API have a primary collection?

- added slug method on application record instances - returns ":collection/:id" if instance class matches a collection class or any of its descendants, nil otherwise. - declared as a virtual_column so it's accessible from API. - Added specs Closes Issue ManageIQ#11635

…ovements on nil id case.

…me_for_subclass

miq-bot · 2017-03-21T20:22:31Z

Checked commits abellotti/manageiq@2f3c950~...f0de33b with ruby 2.2.6, rubocop 0.47.1, and haml-lint 0.20.0
6 files checked, 0 offenses detected
Everything looks good. 🍰

abellotti added api enhancement labels Mar 15, 2017

abellotti assigned Fryguy Mar 15, 2017

abellotti force-pushed the api_slugs branch 2 times, most recently from a3a5062 to dcdf739 Compare March 15, 2017 16:46

gmcculloug reviewed Mar 15, 2017

View reviewed changes

gmcculloug mentioned this pull request Mar 15, 2017

Add manageiq to the extra_var before launching a job #14354

Merged

abellotti force-pushed the api_slugs branch from dcdf739 to ced1a96 Compare March 17, 2017 14:56

Fryguy reviewed Mar 17, 2017

View reviewed changes

abellotti force-pushed the api_slugs branch from ced1a96 to 65d9d97 Compare March 20, 2017 21:55

abellotti added 4 commits March 21, 2017 16:06

moving to lib/extensions/ar_slug.rb, renaming slug -> href_slug, impr…

2d9fbd9

…ovements on nil id case.

Moved building href_slug logic to Api::Utils, name_for_subklass -> na…

297a690

…me_for_subclass

Renaming module to ArHrefSlug

f0de33b

abellotti force-pushed the api_slugs branch from 65d9d97 to f0de33b Compare March 21, 2017 20:16

Fryguy merged commit 5059d08 into ManageIQ:master Mar 21, 2017

Fryguy added this to the Sprint 57 Ending Mar 27, 2017 milestone Mar 21, 2017

mkanoor mentioned this pull request Mar 22, 2017

The Service model href_slug #14457

Merged

abellotti deleted the api_slugs branch April 11, 2017 14:24

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Added support for API slugs #14344

Added support for API slugs #14344

abellotti commented Mar 15, 2017 •

edited

abellotti commented Mar 15, 2017

gmcculloug left a comment

gmcculloug Mar 15, 2017

gtanzillo Mar 15, 2017

gmcculloug Mar 15, 2017

gmcculloug commented Mar 15, 2017

mkanoor commented Mar 17, 2017

abellotti commented Mar 17, 2017

abellotti commented Mar 17, 2017

gmcculloug commented Mar 17, 2017

Fryguy Mar 17, 2017 •

edited

Fryguy Mar 17, 2017 •

edited

abellotti Mar 17, 2017

bzwei Mar 17, 2017

abellotti Mar 20, 2017

gmcculloug Mar 20, 2017

abellotti Mar 21, 2017 •

edited

Fryguy Mar 21, 2017

Fryguy Mar 17, 2017 •

edited

Fryguy Mar 17, 2017

imtayadeway Mar 17, 2017

miq-bot commented Mar 21, 2017

Added support for API slugs #14344

Added support for API slugs #14344

Conversation

abellotti commented Mar 15, 2017 • edited

abellotti commented Mar 15, 2017

gmcculloug left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

gmcculloug commented Mar 15, 2017

mkanoor commented Mar 17, 2017

abellotti commented Mar 17, 2017

abellotti commented Mar 17, 2017

gmcculloug commented Mar 17, 2017

Fryguy Mar 17, 2017 • edited

Choose a reason for hiding this comment

Fryguy Mar 17, 2017 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

abellotti Mar 21, 2017 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Fryguy Mar 17, 2017 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

miq-bot commented Mar 21, 2017

abellotti commented Mar 15, 2017 •

edited

Fryguy Mar 17, 2017 •

edited

Fryguy Mar 17, 2017 •

edited

abellotti Mar 21, 2017 •

edited

Fryguy Mar 17, 2017 •

edited