FEA: add LightGCL

[WanliYoung]

Add GeneralRecommender LightGCL to RecBole-GNN.

[WanliYoung]

作者们好~我主要在ml-1m数据集上验证LightGCL的性能。
超参数调整范围：
temp choice [0.2,0.5,0.8,2]
lambda1 choice [0.01,0.005,0.0001]
lambda2 choice [1e-5,1e-7]
dropout choice [0.1,0.0]
learning_rate choice [0.001,0.0005]

最优超参数设置：
dropout:0.1, lambda1:0.01, lambda2:1e-05, learning_rate:0.001, temp:2

最优结果：
Valid result:
recall@10 : 0.163 mrr@10 : 0.3629 ndcg@10 : 0.201 hit@10 : 0.7128 precision@10 : 0.149
Test result:
recall@10 : 0.1865 mrr@10 : 0.4273 ndcg@10 : 0.2473 hit@10 : 0.7377 precision@10 : 0.182

[hyp1231]

Thanks for the great contribution! 🎉