feat: Node Graph module system - #1995
Conversation
|
Is there a reason we aren't using the directory |
|
@RobBuchananCompPhys Ironically, I originally did have everything in the I'm not opposed to moving the code into the other directory if people find that clearer (as clarity was the entire point of the move in the first place). |
That actual nodes are hard coded, but at least it's an array.
Try and actually encapsulate everything
and correct some documentation on GraphEdgeModel
40d8ceb to
18612d0
Compare
commit 42b43ffdadf4f484ef19793f45dbe922d5f8d915
Author: Adam Washington <adam.washington@stfc.ac.uk>
Date: Mon Jan 13 14:01:26 2025 +0000
Connect edges by name
commit 9e9ee4bff829c8cb9478a71ffbfab6ff2ec8c44d
Author: Adam Washington <adam.washington@stfc.ac.uk>
Date: Mon Jan 13 11:17:45 2025 +0000
Remove temperature from Generator node
commit bd2cd2c30db829b0c052c908a65cac2766b5dfe0
Author: Adam Washington <adam.washington@stfc.ac.uk>
Date: Mon Jan 13 11:08:41 2025 +0000
Create lookup table of nodes by name
This is stored in `nodeType`, which is very confusing and should be fixed
commit 3825a14532c54f9c54c29b8317383ee9b2236efb
Author: Adam Washington <adam.washington@stfc.ac.uk>
Date: Mon Jan 13 11:07:38 2025 +0000
Update nixpkgs
|
A screenshot of the current status of the GraphNodes for the generator. We now just show the graph nodes in order and display the data for the four types I've implemented thus far (with the goal of loading examples/argon/input.txt). Some immediate improvements that can be made
|
|
This has been superseded by the Node Graph system that is now in develop2 |

This contains the intial prototype for the Node Graph viewer. Full user support isn't available, but it does provide a very basic infrastructure.
Above is a basic screenshot of an example of the code. The nodes of the generator are currently linked linearly to indicate the order of the events, but it's perfectly posible to have them each connect back to the generator individually.
This most interesting files to examine are
The following changes should be made eventually, but are out of scope of the current PR:
GraphNodeModelshould return an insertion object used to insert new nodes. This could use RAII for better safety, instead of needing to remember to callbeginInsertandendInsert