From 7c4218b975372c9dbe90d1a11d46039699f63b1f Mon Sep 17 00:00:00 2001
From: jiangchangting1 <jiangchangting@huawei.com>
Date: Thu, 27 Jul 2023 08:42:07 +0000
Subject: [PATCH 1/2] =?UTF-8?q?=E6=B7=BB=E5=8A=A0=E6=8A=BD=E5=8F=96api?=
 =?UTF-8?q?=E4=BF=A1=E6=81=AF=E5=9F=BA=E7=B1=BB?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: jiangchangting1 <jiangchangting@huawei.com>
---
 precision/api_ut_tools/dump/api_info.py | 89 +++++++++++++++++++++++++
 1 file changed, 89 insertions(+)

diff --git a/precision/api_ut_tools/dump/api_info.py b/precision/api_ut_tools/dump/api_info.py
index 5c6138f6e4..8e8adf1a05 100644
--- a/precision/api_ut_tools/dump/api_info.py
+++ b/precision/api_ut_tools/dump/api_info.py
@@ -1,6 +1,95 @@
 # 定义API INFO，保存基本信息，用于后续结构体的落盘，注意考虑random场景及真实数据场景
 
+import inspect
+import torch
+from .utils import DumpUtil, DumpConst, write_npy
+from ..common.utils import print_error_log
 
 class APIInfo:
     def __init__(self, api_name):
         self.api_name = api_name
+        self.save_real_data = DumpUtil.save_real_data
+        self.buildin_class = []
+
+    def analyze_element(self, element):
+        if isinstance(element, (list, tuple)):
+            out = []
+            for item in element:
+                out.append(self.analyze_element(item))
+        elif isinstance(element, dict):
+            out = {}
+            for key, value in element.items():
+                out[key] = self.analyze_element(value)
+        elif isinstance(element, torch.Tensor):
+            out = self.analyze_tensor(element, self.save_real_data)
+
+        elif self.is_builtin_class(element):
+            out = self.analyze_builtin(element)
+        else:
+            msg = f"Type {type(element)} is unsupported at analyze_element"
+            print_error_log(msg)
+            return None
+            # raise NotImplementedError(msg)
+        return out
+
+    def analyze_tensor(self, arg, save_real_data):
+        single_arg = {}
+        if not save_real_data:
+            
+            single_arg.update({'type' : 'torch.Tensor'})
+            single_arg.update({'dtype' : str(arg.dtype)})
+            single_arg.update({'shape' : arg.shape})
+            single_arg.update({'Max' : self.transfer_types(self.get_tensor_extremum(arg,'max'), str(arg.dtype))})
+            single_arg.update({'Min' : self.transfer_types(self.get_tensor_extremum(arg,'min'), str(arg.dtype))})
+            single_arg.update({'requires_grad': arg.requires_grad})
+            
+        else:
+            npy_path = write_npy(self.api_name, arg.contiguous().cpu().detach().numpy())
+            single_arg.update({'type' : 'torch.Tensor'})
+            single_arg.update({'datapath' : npy_path})
+            single_arg.update({'requires_grad': arg.requires_grad})
+        return single_arg
+
+    def analyze_builtin(self, arg):
+        single_arg = {}
+        if isinstance(arg, slice):
+            single_arg.update({'type' : "slice"})
+            single_arg.update({'value' : [arg.start, arg.stop, arg.step]})
+        else:
+            single_arg.update({'type' : self.get_type_name(str(type(arg)))})
+            single_arg.update({'value' : arg})
+        return single_arg
+
+    def transfer_types(self, data, dtype):
+        if 'bool' in dtype:
+            breakpoint()
+        if 'int' in dtype or 'bool' in dtype:
+            return int(data)
+        else:
+            return float(data)
+
+    def is_builtin_class(self, element):
+        if element is None or isinstance(element, (bool,int,float,str,slice)):
+            return True
+        return False
+
+    
+    def get_tensor_extremum(self, data, operator):
+        if data.dtype is torch.bool:
+            if operator == 'max':
+                return True in data
+            elif operator == 'min':
+                if False in data:
+                    return False
+                else:
+                    return True
+        if operator == 'max':
+            return torch._C._VariableFunctionsClass.max(data).item()
+        else:
+            return torch._C._VariableFunctionsClass.min(data).item()
+    
+    def get_type_name(self, name):
+
+        left = name.index("'")
+        right = name.rindex("'")
+        return name[left + 1 : right]
-- 
Gitee


From 50d15d7308c3db89ac68f2e0969c3236d732af83 Mon Sep 17 00:00:00 2001
From: jiangchangting1 <jiangchangting@huawei.com>
Date: Thu, 27 Jul 2023 11:19:01 +0000
Subject: [PATCH 2/2] update precision/api_ut_tools/dump/api_info.py.

Signed-off-by: jiangchangting1 <jiangchangting@huawei.com>
---
 precision/api_ut_tools/dump/api_info.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/precision/api_ut_tools/dump/api_info.py b/precision/api_ut_tools/dump/api_info.py
index 8e8adf1a05..911bba77ce 100644
--- a/precision/api_ut_tools/dump/api_info.py
+++ b/precision/api_ut_tools/dump/api_info.py
@@ -28,8 +28,8 @@ class APIInfo:
         else:
             msg = f"Type {type(element)} is unsupported at analyze_element"
             print_error_log(msg)
-            return None
-            # raise NotImplementedError(msg)
+
+            raise NotImplementedError(msg)
         return out
 
     def analyze_tensor(self, arg, save_real_data):
-- 
Gitee