During the past years, the data deluge that prevails in the World Wide Web has been accompanied by a number of APIs that expose business logic. In this paper, we discuss a novel approach to enrich existing API standards definitions with business rules. Taking advantage of the REST principles, we aim at enabling the creation of generic clients that can dynamically navigate through semantically enriched web affordances with the help of Hydrabased Hypermedia API descriptions, which encapsulate the finite state machine of possible actions into SWRL rules.

1. Adding Rules on
Existing Hypermedia
APIs
Michael Petychakis, Fenareti Lampathaki, Dimitrios
Askounis
National Technical University of Athens
WS-REST, 18th May 2015, Florence Italy

2. Who I am
Michael Petychakis
Researcher in DSS lab, National Technical University of
Athens (NTUA)
PhD Candidate in NTUA
Electrical and Computer Engineer, Dipl.-M.Eng., NTUA
Research Interests
APIs/Linked Data/Web 3.0
@mpetyx

3. Web, APIs, IoT and .. Hypermedia

4. REST (as was supposed to be)
Level 3:
Hypermedia
Level 2: HTTP
Verbs
Level 1: Resources
Level 0: POX

5. REST (today)
Level 3:
Hypermedia
Level 2: HTTP
Verbs
Level 1: Resources
Level 0: POX

6. Hypermedia as FSM
"hydra:member" : [
{
"@type": "schema:Basket",
"schema:contains": [
{
"@id": "/books/1",
"@type": "schema:Book",
"schema:author": "James Joyce",
"schema:title": "Ulysses"
},

7. Rules are Good for FSM
If This
then That
If
Basket_Contains( ?Book )
loggen_in_validated_user( ?Michael )
paypal_account_integrated( ?Michael )
Then
Checkout( ?Michael, ?Book )

8. Rules on the Web
dc:description
Rules
SWRL
Ruleml
RIF
Rule
Inference
Producing valid statements
within rule based system

9. Rules are there for a Loong Time
Not Web Web

10. Blending Rules and APIs
Hydra/JSON_LD
SWRL
DeepGraphs

11. DeepGraphs Approach
Goal: Modelling Hypermedia Responses
1 Vocabulary
 Hydra Documents
 SWRL Rules
 JSON-LD Responses
1 Vocabulary

12. Vocabulary
• Affordances
• Duration(maxAllowedDuration)
• According to latest updates on Http 1.1
• Parallel Processes
• Synchronizing multiple FSM Clients
• Constraints
• The Actual Rules
Schema:activity Hydra:Resource dg:affordances

13. Similar Approaches/ Media Types
DeepGraphs
Mason
Uber
JSON-LD
JSON API
Siren
HAL
Collection
JSON
JSON
Schema
JSON
Hyper
Schema
XML
Schema

14. Example: Bookstore
"dg:affordances": [
{
"@type": ["schema:Action","hydra:Operation"],
"@id": "/RemoveBook",
"hydra:method": "DELETE",
"hydra:expects": "schema:Book",
"hydra:title": "Deletes a Book resource.",
"hydra:returns": "owl:Nothing",
"rdfs:label": "Deletes the Book resource."
},

15. Example in Diagrams

16. Complexity grows Exponentially
Clients
Bookstore
Paypal
Sms
Agent
Email
Facebook
Twitter
Servers
Bookstore
Paypal
Sms
-
Email
Facebook
Twitter

17. Complexity grows Exponentially
Clients
Bookstore
Paypal
Sms
Agent
Email
Facebook
Twitter
Servers
Bookstore
Paypal
Sms
-
Email
Facebook
Twitter
Can We Avoid Building one Client per Server and Vice
Versa?

18. DeepGraphs Generic Server Implementations
• Developer Designs the FSM
• A designer’s console for:
 Designing the API Atlas
 Verification of API FSM
 Validation (Running Example Workflows)
 Adding Integrity Constraints
Defining the Business Logic through Rules

19. API Designer
http://api-builder.tools.epu.ntua.gr/web/

20. DeepGraphs Generic AI Clients
SW
Interoperabilty
Reasoning
REST
Multiple
Servers
Multiple
Transactions

21. In APIs We Trust

22. Tools Around Spec
● API Builder
● PyAPI
o from [Swagger, API Blueprints, Hydra, WADL, RAML] to
[Swagger, API Blueprints, Hydra, RAML]
● PyRIF
o from [RuleML, RIF, SWRL] to [RuleML, RIF, SWRL]
● DeepGraphs Parser ( Soon )
● DeepGraphs Reasoner ( Soon )
o Datalog Approach
o FSM resolver implemented now
o Fluent Editor has been used for Evaluation

23. Dereferencable Rules on the Web
SWRL and Rif
Solution:
Declare 1 Rule within a Graph, and
Reference this. Status: Working, But..
dg:constraint
http://deepgraphs.org/repository/rule1

24. Advances of the Methodology
• DeepGraphs is not only REST
XMPP
MQTT
CoAP, DDS, AMQP (and more!)
• DeepGraphs aims into reusing Existing APIs that
speak Swagger, RAML, etc
• DeepGraphs aims to facilitate:
• Distributed Reasoning over the web with Affordances

25. Next Steps
1. First Release of the Specification on June
2. Create a public Community Group
3. DeepGraphs Tools
• Parser
• Client
• Reasoner
• Designer
4. More Events
5. Address Rule Dereferencability by open discussion

26. Thank you
Michael Petychakis
<a
href="mailto:mpetyx@epu.ntua.gr?Subject=
Hello" target="_top">Drop me an e-mail</a>
@mpetyx
This work has been created in the context of the EU-funded
project OPENi (Open-Source, Web-Based, Framework for
Integrating Applications with Social Media Services and Personal
Cloudlets), Contract No: FP7-ICT-317883.