1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
|
_pandas = None
_WITH_PANDAS = None
def _try_import_pandas() -> bool:
try:
import pandas # type: ignore[import]
global _pandas
_pandas = pandas
return True
except ImportError:
return False
# pandas used only for prototyping, will be shortly replaced with TorchArrow
def _with_pandas() -> bool:
global _WITH_PANDAS
if _WITH_PANDAS is None:
_WITH_PANDAS = _try_import_pandas()
return _WITH_PANDAS
class PandasWrapper:
@classmethod
def create_dataframe(cls, data, columns):
if not _with_pandas():
raise Exception("DataFrames prototype requires pandas to function")
return _pandas.DataFrame(data, columns=columns) # type: ignore[union-attr]
@classmethod
def is_dataframe(cls, data):
if not _with_pandas():
return False
return isinstance(data, _pandas.core.frame.DataFrame) # type: ignore[union-attr]
@classmethod
def is_column(cls, data):
if not _with_pandas():
return False
return isinstance(data, _pandas.core.series.Series) # type: ignore[union-attr]
@classmethod
def iterate(cls, data):
if not _with_pandas():
raise Exception("DataFrames prototype requires pandas to function")
for d in data.itertuples(index=False):
yield d
@classmethod
def concat(cls, buffer):
if not _with_pandas():
raise Exception("DataFrames prototype requires pandas to function")
return _pandas.concat(buffer) # type: ignore[union-attr]
@classmethod
def get_item(cls, data, idx):
if not _with_pandas():
raise Exception("DataFrames prototype requires pandas to function")
return data[idx: idx + 1]
@classmethod
def get_len(cls, df):
if not _with_pandas():
raise Exception("DataFrames prototype requires pandas to function")
return len(df.index)
@classmethod
def get_columns(cls, df):
if not _with_pandas():
raise Exception("DataFrames prototype requires pandas to function")
return list(df.columns.values.tolist())
# When you build own implementation just override it with dataframe_wrapper.set_df_wrapper(new_wrapper_class)
default_wrapper = PandasWrapper
def get_df_wrapper():
return default_wrapper
def set_df_wrapper(wrapper):
global default_wrapper
default_wrapper = wrapper
def create_dataframe(data, columns=None):
wrapper = get_df_wrapper()
return wrapper.create_dataframe(data, columns)
def is_dataframe(data):
wrapper = get_df_wrapper()
return wrapper.is_dataframe(data)
def get_columns(data):
wrapper = get_df_wrapper()
return wrapper.get_columns(data)
def is_column(data):
wrapper = get_df_wrapper()
return wrapper.is_column(data)
def concat(buffer):
wrapper = get_df_wrapper()
return wrapper.concat(buffer)
def iterate(data):
wrapper = get_df_wrapper()
return wrapper.iterate(data)
def get_item(data, idx):
wrapper = get_df_wrapper()
return wrapper.get_item(data, idx)
def get_len(df):
wrapper = get_df_wrapper()
return wrapper.get_len(df)
|
