Add Models for TRIDENT Detector in Graphnet

[040601]

This PR extends the Graphnet project to include support for the TRIDENT detector by adding specific code files for training and testing. The changes include four new files which are described in detail below.

Changes Made:

• TRIDENTNodeDefinition.py: Defines the node structure TRIDENTGraphDefinition used for training with TRIDENT data.
• TRIDENTGraphDefinition.py: Contains both the TRIDENT detector information and the graph definition class TRIDENTGraphDefinition to represent the TRIDENT detector structure and graph type.
• MiddleReconModel.py: Defines the model MiddleReconModel used in TRIDENT training. Key fuctions include: compute_loss, forward, shared_step, construct_trainer, fit, _print_callbacks, _contains_callback, configure_optimizers, training_step, validation_step, predict_step, inference, train, predict, predict_as_dataframe, _create_default_callbacks and _add_early_stopping.
• TridentNet.py: Contains three network classes used in TRIDENT training: StaticEdgeConv, DynamicEdgeConv, and TridentTrackNet.

Additional Notes:

• We would also like to seek some advice on where it would make more sense to place the added code files.

[040601]

tagging @wlhwl @cmo-ft

[Aske-Rosted]

PR is closed due to inactivity.