* add bitwise_not trt converter

* run pre-commit

* modify neg_one_tensor_dims init way

* fix BOOL type support requires TensorRT 8.4

* fix int8 & uint8 type

* improve data type readability

* modify filter logic

* fix coverage CI