Skip to main content


Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Eclipse Projects » Eclipse Sprotty » Sprotty for Mining and Analysis Workflow Visualization
Sprotty for Mining and Analysis Workflow Visualization [message #1802949] Tue, 19 February 2019 13:37 Go to next message
Patrick Neubauer is currently offline Patrick NeubauerFriend
Messages: 4
Registered: May 2017
Junior Member
Dear Sprotty developers,

We are members of the CROSSMINER EU project, which involves mining of open-source software from various providers, and developed Crossflow, i.e. a language to specify mining and analysis workflows [1].

As part of our web application, we would like to graphically visualise the workflow execution during runtime and found Sprotty to be a fitting candidate.

As a result, we generated an Xtext-based language of Crossflow, i.e. our workflow language, replaced the exemplary "multicore" language with it, and drafted the diagram generator accordingly.

Although we were able to produce a graph with nodes as well as compute its layout [2], we were unable to successfully visualise the result in the browser [3].

The experiment is available on Github [4] and is fairly easy to reproduce.

Can you provide any hint on what may go wrong?

Cheers, Patrick

[1] Crossflow metamodel: https://github.com/crossminer/scava/blob/crossflow/crossflow/org.eclipse.scava.crossflow.language/model/crossflow.pdf
Exemplary Crossflow workflow graph: https://github.com/patrickneubauer/sprotty-experiment/raw/master/img/img5.png

[2] Console output: https://github.com/patrickneubauer/sprotty-experiment/raw/master/img/img3.png

[3] Browser output: https://github.com/patrickneubauer/sprotty-experiment/raw/master/img/img2.png

[4] Experimental repo: https://github.com/patrickneubauer/sprotty-experiment
Re: Sprotty for Mining and Analysis Workflow Visualization [message #1802984 is a reply to message #1802949] Wed, 20 February 2019 08:21 Go to previous messageGo to next message
Miro Spönemann is currently offline Miro SpönemannFriend
Messages: 74
Registered: March 2015
Location: Kiel, Germany
Member

The "multicore" example is quite outdated. If you want to use Sprotty with Xtext, you should rather take these other examples as reference:

https://github.com/OpenKieler/elkgraph-web
https://github.com/TypeFox/theia-xtext-sprotty-example
Re: Sprotty for Mining and Analysis Workflow Visualization [message #1803436 is a reply to message #1802984] Thu, 28 February 2019 16:48 Go to previous messageGo to next message
Patrick Neubauer is currently offline Patrick NeubauerFriend
Messages: 4
Registered: May 2017
Junior Member
Hi,

thank you very much for your reply!

The repos you mentioned helped us greatly in getting the example running in our web app (after minor changes like adapting the web socket port to 9090 from 8080, i.e. already used by our web app) [1].

Now we are wondering what we can *exclude* for our use case, which intends only the visualisation of a model (at least for the foreseeable future). In other words, no change in the structure of the model is required at runtime.

In general, we intend to specify either an ETL model transformation or a EGL code generation, which we are familiar with, to yield elk graph models from our Ecore models and feed the result to the elkgraph-web language server in order to get the layout for the client.

Two questions in this regard:
1. what part of the client code can be excluded?
2. the server code is still required to generate/refresh the model layout during runtime in case the user changes the browser window size and the like, correct?

If you could point out the right places where to look at, that would already be of great help!

Thanks and all the best,

Patrick

[1] our web app code is in the project org.eclipse.scava.crossflow.web in branch 'crossflow' ( https://github.com/crossminer/scava/tree/crossflow/crossflow ); the server-side code of your example is placed in the project org.eclipse.scava.crossflow.elkgraph

[Updated on: Thu, 28 February 2019 16:55]

Report message to a moderator

Re: Sprotty for Mining and Analysis Workflow Visualization [message #1803469 is a reply to message #1803436] Fri, 01 March 2019 09:45 Go to previous message
Miro Spönemann is currently offline Miro SpönemannFriend
Messages: 74
Registered: March 2015
Location: Kiel, Germany
Member

Patrick Neubauer wrote on Thu, 28 February 2019 17:48
Two questions in this regard:
1. what part of the client code can be excluded?
2. the server code is still required to generate/refresh the model layout during runtime in case the user changes the browser window size and the like, correct?


1. I don't understand the question. What would you like to exclude?

2. In our Xtext examples the server is responsible for providing the model ("model source") and to compute layouts. In case the model is provided elsewhere, you can use elkjs to compute the layout in the client, so the server is not really necessary (see also elkgraph-web example with json). But if the source of your models is Xtext / EMF you'll need a server component anyway.
Previous Topic:Multiple Diagram Servers
Next Topic:Stacking complex actions - add only for selected nodes
Goto Forum:
  


Current Time: Fri Sep 24 23:35:02 GMT 2021

Powered by FUDForum. Page generated in 0.01780 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.2.
Copyright ©2001-2010 FUDforum Bulletin Board Software

Back to the top