Getting to work on the exporting the results

[pablopareja]

Hey @andre-nunes

Did you already get to work on how to export the results to either XML or JSON? This part of the project is independent from the steps/DSL definition or CLI implementation Do you have a preliminary idea on how you're going to do it? Let me know if you have any question about this :wink:

[andr3nun3s]

Can I use this library? The other way is to implement it myself.

Do I need to use this model?

[pablopareja]

Sure you can use that library, just keep in mind that you will have to use the GraphSONUtility and implement yourself at least part of the logic to export a query to JSON. (we're obviously not exporting the whole DB :smiley: )

The model you mention represents the data types together with their properties, not the data itself of specific instances from specific types that would be the result of a given Gremlin query.

[laughedelic]

oh.. I was writing almost the same reply! ")

[pablopareja]

:smile: :stuck_out_tongue:

[andr3nun3s]

What's the difference between a Jettison and a Jackson? Which should I use?

[pablopareja]

They are two different APIs/libraries with similar characteristics, use the one you feel more comfortable with.

[laughedelic]

Don't choose it randomly, take a look at the first 3 links from the google search "Jettison vs Jackson", it may give you some clues how to chose it for the purposes of the project:

• http://stackoverflow.com/questions/11956274/difference-between-jettison-and-jackson-in-jboss
• http://stackoverflow.com/questions/7248350/jackson-jettison-usage-in-jersey
• http://cxf.547215.n5.nabble.com/JSON-provider-Jettison-vs-Jackson-td5736930.html

Why not to make a list of pros and cons for the two? And then maybe the choice will be obvious. If not, we'll decide it together :wink:

[andr3nun3s]

Jettison

• "Old"
• "Jettison is for for mapping JSON to XML"
• "Jettison is a much simpler provider. I think that CXF Provider and Jettison are quite flexible though. In some scenarios it is more realistic to set various Jettison properties than adding JSON annotations, etc."

Jackson

• "Jackson is arguably the standard Java Library for processing JSON. Jackson is included in Glassfish"
• "Jackson is for JSON generation/parsing"
• "Jackson is a streaming writer, so if a service has medium to large JSON sequences, then it will probably start making a difference. Generally speaking, Jackson is a more sophisticated/advanced provider. "
• "schema validation doesn't work with Jackson provider"

From what I've gathered, Jackson seems the obvious choice, do you agree?

[laughedelic]

absolutely agree! well done :+1:

[andr3nun3s]

Ok, imagine that a query iterates a bunch of vertices, using GraphSONUtility I'll just print each one of those vertices into the output file?

[laughedelic]

well, if the query result is multiple vertices, I guess, the output should be a JSON array of the serialized vertices.

[andr3nun3s]

Ok, so I've tested exporting to GraphSON and it worked. I used a simple query using our DSL on the classic TinkerGraph to test it, here's the commit: https://github.com/bio4j/exporter/commit/06f613726d4ab86e9ba738d49db51f7581f85ce5

As GraphML is a specification for the serialization of an entire graph and not the individual elements of a graph, methods that support input and output of single vertices and edges are not supported. I'll have to add the vertices/edges to a graph and export the whole graph. Source