/
utils.py
155 lines (108 loc) · 5.55 KB
/
utils.py
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
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
import pandas as pd
import pickle
import torch
import numpy as np
class MADataset:
def __init__(self):
self.s_year = 1997
self.e_year = 2020
self.read_path = '../MA_data/data'
# read frequent acuqirer info tuples
with open(self.read_path+f"/freq_a_info_{self.s_year}_{self.e_year}.pickle", "rb") as f:
a_freq, a_freq_lst, a_freq_idx_to_gvkey_mapping, a_freq_gvkey_to_idx_mapping = pickle.load(f)
self.a_freq, self.a_freq_lst, self.a_freq_idx_to_gvkey_mapping, self.a_freq_gvkey_to_idx_mapping = a_freq, a_freq_lst, a_freq_idx_to_gvkey_mapping, a_freq_gvkey_to_idx_mapping
#read dataset
with open(self.read_path+f"/ma_dataset_N01.pickle", 'rb') as f: ###
ma_dataset = pickle.load(f)
self.ma_dataset = ma_dataset # a list
assert len(self.ma_dataset) == len(self.a_freq_lst), "a_freq and ma_dataset length dismatch"
# def load_data(self):
# with open(self.read_path+f"/ma_dataset1.pickle", 'rb') as f:
# ma_dataset = pickle.load(f)
# modified = []
# for ele in ma_dataset:
# arr_b, arr_c, arr_delta_time, event_data, non_event_data, estimate_length, choice_data_dict = ele
# # lst to array
# arr_b_idx, arr_c_idx, arr_delta_time = event_data
# event_data = np.array(arr_b_idx), np.array(arr_c_idx), arr_delta_time
# # lst to array
# arr_b_idx, arr_c_idx, arr_delta_time = non_event_data
# non_event_data = np.array(arr_b_idx), np.array(arr_c_idx), arr_delta_time
# aa, true_tar_idxs, bb, cc = choice_data_dict
# new = {}
# for k,v in true_tar_idxs.items():
# new[k] = v.to(torch.float).numpy()
# print(type(v))
# choice_data_dict = aa, new, bb, cc
# modified.append((arr_b, arr_c, arr_delta_time, event_data, non_event_data, estimate_length, choice_data_dict))
# return modified
def __getitem__(self, idx):
return self.ma_dataset[idx]
def __len__(self):
return len(self.a_freq_lst)
class MADataset_test:
def __init__(self):
self.s_year = 1997
self.e_year = 2020
self.read_path = '../MA_data/data'
# read frequent acuqirer info tuples
with open(self.read_path+f"/freq_a_info_{self.s_year}_{self.e_year}.pickle", "rb") as f:
a_freq, a_freq_lst, a_freq_idx_to_gvkey_mapping, a_freq_gvkey_to_idx_mapping = pickle.load(f)
self.a_freq, self.a_freq_lst, self.a_freq_idx_to_gvkey_mapping, self.a_freq_gvkey_to_idx_mapping = a_freq, a_freq_lst, a_freq_idx_to_gvkey_mapping, a_freq_gvkey_to_idx_mapping
#read dataset
with open(self.read_path+f"/ma_dataset_test.pickle", 'rb') as f: ###
ma_dataset = pickle.load(f)
self.ma_dataset = ma_dataset # a list
#assert len(self.ma_dataset) == len(self.a_freq_lst), "a_freq and ma_dataset length dismatch"
# def load_data(self):
# with open(self.read_path+f"/ma_dataset1.pickle", 'rb') as f:
# ma_dataset = pickle.load(f)
# modified = []
# for ele in ma_dataset:
# arr_b, arr_c, arr_delta_time, event_data, non_event_data, estimate_length, choice_data_dict = ele
# # lst to array
# arr_b_idx, arr_c_idx, arr_delta_time = event_data
# event_data = np.array(arr_b_idx), np.array(arr_c_idx), arr_delta_time
# # lst to array
# arr_b_idx, arr_c_idx, arr_delta_time = non_event_data
# non_event_data = np.array(arr_b_idx), np.array(arr_c_idx), arr_delta_time
# aa, true_tar_idxs, bb, cc = choice_data_dict
# new = {}
# for k,v in true_tar_idxs.items():
# new[k] = v.to(torch.float).numpy()
# print(type(v))
# choice_data_dict = aa, new, bb, cc
# modified.append((arr_b, arr_c, arr_delta_time, event_data, non_event_data, estimate_length, choice_data_dict))
# return modified
def __getitem__(self, idx):
return self.ma_dataset[idx]
def __len__(self):
return 1
# class MADataset:
# '''
# '''
# def __init__(self):
# ###### basic params no need to tuning
# self.tmp_data_path = '../MA_data/data/tmp'
# self.data_path = '../MA_data/data'
# # WARINING: those are self.event year, not TNIC year
# # if the focal event is in ith year, use i-1th year TNIC and fv data
# self.s_year = 1997
# self.e_year = 2020
# # basic data set prepared
# self.sdc_tnic = self.read_tnic()
# self.
# self.idx_to_gvkey_mapping =
# @staticmethod
# def read_tnic(self):
# sdc_tnic = pd.read_pickle(self.tmp_data_path+f"/sdc_tnic_{self.s_year}_{self.e_year}")
# return sdc_tnic
# def get_freq_acquirer(self):
# A_freq = pd.DataFrame(sdc_tnic.AGVKEY.value_counts()).reset_index(drop=False)
# A_freq = A_freq[A_freq.AGVKEY >= 5]
# A_freq.columns = ["GVKEY", "freq"]
# print(f"totally {A_freq.shape[0]} numbers of frequent Acquirers")
# def get_data_d(self, focal_gvkey):
# c_d, b_d, timeline_d = self.dataloader_preproceser(focal_gvkey)
# def __getitem__(self, idx):
# gvkey = self.idx_to_gvkey_mapping[idx]