gkontos / gasket Goto Github PK

A gasket is a seal which fills the space between two or more mating surfaces, generally to prevent leakage.

The gasket project is an example API written in golang featuring :

• 'Normalized' JSON object model mapped to an rdf data store (see schema below for details)
• Integration of Cayley as a go library
• Example of capturing quad meta-data using cayley v0.6
• Examples of miscellaneous plumbing required for an API in golang (external configuration, logging, unit tests for handlers)

This project will build, compile, and run as a functional API; however, there are certain crucial elements which are not implemented. Those include :

• Validation
• Security
• Documentation (ala Swagger)

• mongoDb download
• glide Glide - Github

• The project will look for a configuration path from an Environment variable at ACES_CFG. There are two separate files that can be set. log.toml and config.toml. Example files for each can be found within the project within the config folder.
• If a database server name is not found in config.toml, the project will not run

• run 'go build' from the project directory

• run ./gasket to start the server

The project utilizes three JSON objects. Node, Relation, and Metadata.

Node : Is a basic object with an extensible list of properties. The JSON schema for a Node is defined as :

"node": {
	"type":"object",
	"properties": {
		"id": {"type":"string"},
		"name":{"type":"string"},
		"label":{"type":"string"},
	},
	"patternProperties": {
        	"^(/[^/]+)+$": { "type": "object" }
		/**
		 * Accept any property without a '/' character
		 */
   		}
	"required":["name"]
}

Each node will be stored as several quads within the graph store. The basic quad representation is

<id> <schema:name> "nameParam" "label"
<id> <patternPropertyA> "propertyAParam" "label"
<id> <patternPropertyB> "propertyBParam" "label"
... etc 

Relation : will relate two nodes within a typical rdf quad. The schema is :

"relation": { 
	"type":"object",
	"properties": {
		"id": {"type":"string"},
		"sourceId":{"type":"string"},
		"type":{"type":"string"},
		"targetId":{"type":"string"},
		"label":{"type":"string"},
	}
	"required":["sourceId", "type", "targetId", "label"]
}
	

The relation will be stored as a quad with

quad.Quad{
	Subject : sourceId,
	Predicate : type,
	Object : targetId,
	Label : label,
}

A JSON representation of this quad will be stored and assigned an id. This id will be associated with the JSON object.

Metadata : Provides a way of storing additional details about the relation. The schema is similar to a node :

"metadata": {
	"type":"object",
	"properties": {
		"id": {"type":"string"},
		"relationId":{"type":"string"},
	},
	"patternProperties": {
        	"^(/[^/]+)+$": { "type": "object" }
		/**
		 * Accept any property without a '/' character
		 */
   		}
	"required":["relationId"]

}

And the method of mapping metadata to quads is similar to that used for nodes.

Released under the MIT license. See LICENSE file for more details.