Reporting provenance using JSON - ashish-gehani/SPADE GitHub Wiki

This reporter allows provenance to be described in JSON using W3C PROV concepts. The information must be stored in a file that contains valid JSON and conforms to the data model described below.

---

Data model

The file must contain a JSON array of JSON objects. Each provenance vertex and edge must be described in a separate JSON object. Vertices and edges have different key-value pairs, as described below. In both cases, the JSON object can optionally contain an annotations key. If this key is present, the value associated with it must a JSON object of key-value pairs (that describe domain-specific provenance details).

Provenance element	Key	Value	Required
Vertex
	type	can be one of: Agent Activity Entity	Yes
	id	unique identifier	Yes
	annotations	JSON object with key-value pairs	No
Edge
	type	can be one of: ActedOnBehalfOf WasInformedBy WasDerivedFrom WasAssociatedWith WasAttributedTo Used WasGeneratedBy	Yes
	from	unique identifier of source vertex	Yes
	to	unique identifier of destination vertex	Yes
	annotations	JSON object with key-value pairs	No

---

Sample valid JSON provenance

In the example below, the first JSON object represents an Activity vertex with unique identifier 1. The second JSON object represents an Entity vertex with unique identifier 2. Finally, the third JSON object represents a Used edge, indicating how the activity in vertex 1 is connected to the entity in vertex 2.

[
    {
        "type": "Activity",
        "id": 1,
        "annotations": {
            "program": "firefox",
            "pid": "1234"
        },
    },
    {
        "type": "Entity",
        "id": 2,
        "annotations": {
            "filename": "index.html",
            "owner": "user"
        },
    },
    {
        "type": "Used",
        "from": 1,
        "to": 2,
        "annotations": {
            "time": "0420"
        }
    }
]

---

Configuring JSON reporting

The JSON reporter takes a single argument, which is the location in the filesystem where the JSON file is located. Note that this must be done in the SPADE controller (after the SPADE server has been started):

-> add reporter JSON input=/tmp/provenance.json
Adding reporter JSON... done

---

The reporter can be deactivated using the following command in the SPADE controller:

-> remove reporter JSON
Shutting down reporter JSON... done