[🐛BUG] KGAT model dont work, even in quickstart

[kranov582]

When trying to run the quickstart function with the KGAT model and the 'ml-100k' dataset, an unexpected keyword argument appears.

`
from recbole.quick_start import run_recbole

run_recbole(model='KGAT', dataset='ml-100k')
`

Traceback (most recent call last):
File "C:\Users\e-pedrob\PycharmProjects\recboles_tests\recbole_run.py", line 3, in
run_recbole(model='KGAT', dataset='ml-100k')
File "C:\Users\e-pedrob\PycharmProjects\recboles_tests.venv\Lib\site-packages\recbole\quick_start\quick_start.py", line 137, in run_recbole
model = get_model(config["model"])(config, train_data._dataset).to(config["device"])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\e-pedrob\PycharmProjects\recboles_tests.venv\Lib\site-packages\recbole\model\knowledge_aware_recommender\kgat.py", line 111, in init
self.init_graph()
File "C:\Users\e-pedrob\PycharmProjects\recboles_tests.venv\Lib\site-packages\recbole\model\knowledge_aware_recommender\kgat.py", line 154, in init_graph
dgl.edge_subgraph(self.ckg, edge_idxs, preserve_nodes=True)
TypeError: edge_subgraph() got an unexpected keyword argument 'preserve_nodes'

Desktop

• OS: Windows
• RecBole Version 1.2.0
• Python Version 3.12
• PyTorch Version 2.2.1
• cudatoolkit Version none