Zgc/ditorch support overflow apply to third party interface

[zhaoguochun1995]

1. 精度溢出检测工具支持作用在非torch接口上
2. 完善readme
3. 新增配置管理(终端中打印出相关的配置，用户对可配置选项能更清楚)

>>> import op_tools
>>>        
>>> autocompare = op_tools.OpAutoCompare()
>>> autocompare.start()
>>> import torch
>>> x = torch.randn(3,4,device="cuda")
skip OpAutoCompareHook on torch.randn
>>> 
>>> y = x + x
option: OP_AUTOCOMPARE_DISABLE_LIST=torch.rand,torch.randn,torch_mlu.*,torch_npu.*
option: OP_AUTOCOMPARE_LIST=.*
option: OP_TOOLS_PRINT_STACK=0
apply OpAutoCompareHook on torch.Tensor.add
option: OP_DTYPE_CAST_DICT=torch.float16->torch.float32,torch.bfloat16->torch.float32
option: AUTOCOMPARE_ERROR_TOLERANCE_FLOAT32=1e-5,1e-5
option: OP_TOOLS_MAX_CACHE_SIZE=1000

autocompare    torch.Tensor.add forward_id: 1    
<stdin>:1 <module>: 
+---------------------------------+--------+---------+-------+--------+--------+---------------+---------------+----------------+
|               name              | device |  dtype  | numel | shape  | stride | requires_grad |     layout    |    data_ptr    |
+---------------------------------+--------+---------+-------+--------+--------+---------------+---------------+----------------+
|    torch.Tensor.add inputs[0]   | cuda:0 | float32 |   12  | (3, 4) | (4, 1) |     False     | torch.strided | 20067179823104 |
|    torch.Tensor.add inputs[1]   | cuda:0 | float32 |   12  | (3, 4) | (4, 1) |     False     | torch.strided | 20067179823104 |
|     torch.Tensor.add outputs    | cuda:0 | float32 |   12  | (3, 4) | (4, 1) |     False     | torch.strided | 20067179823616 |
| torch.Tensor.add inputs(cpu)[0] |  cpu   | float32 |   12  | (3, 4) | (4, 1) |     False     | torch.strided |   546780864    |
| torch.Tensor.add inputs(cpu)[1] |  cpu   | float32 |   12  | (3, 4) | (4, 1) |     False     | torch.strided |   546827200    |
|  torch.Tensor.add outputs(cpu)  |  cpu   | float32 |   12  | (3, 4) | (4, 1) |     False     | torch.strided |   556165184    |
+---------------------------------+--------+---------+-------+--------+--------+---------------+---------------+----------------+
+--------------------------------+----------+-------------------+--------------+-------------------+-------------+-------------+------------+
|              name              | allclose | cosine_similarity | max_abs_diff | max_relative_diff |     atol    |     rtol    | error_info |
+--------------------------------+----------+-------------------+--------------+-------------------+-------------+-------------+------------+
| torch.Tensor.add input[0]      |   True   |    1.000000000    | 0.000000000  |    0.000000000    | 0.000010000 | 0.000010000 |            |
| torch.Tensor.add input[1]      |   True   |    1.000000000    | 0.000000000  |    0.000000000    | 0.000010000 | 0.000010000 |            |
| torch.Tensor.add output        |   True   |    1.000000000    | 0.000000000  |    0.000000000    | 0.000010000 | 0.000010000 |            |
+--------------------------------+----------+-------------------+--------------+-------------------+-------------+-------------+------------+



>>>