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
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
|
#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
Collect the apple dataset.
This script uses the yfinance package to download the data from Yahoo Finance
and subsequently reformats it to a JSON file that adheres to our dataset
schema. See the README file for more information on the dataset.
Author: G.J.J. van den Burg
License: This file is part of TCPD, see the top-level LICENSE file.
Copyright: 2019, The Alan Turing Institute
"""
import argparse
import clevercsv
import hashlib
import json
import os
import yfinance
import sys
import time
from functools import wraps
from urllib.error import URLError
MD5_CSV = "9021c03bb9fea3f16ecc812d77926168"
MD5_JSON = "22edb48471bd3711f7a6e15de6413643"
SAMPLE = 3
NAME_CSV = "AAPL.csv"
NAME_JSON = "apple.json"
class ValidationError(Exception):
def __init__(self, filename):
self.message = (
"Validating the file '%s' failed. \n"
"Please raise an issue on the GitHub page for this project \n"
"if the error persists." % filename
)
def check_md5sum(filename, checksum):
with open(filename, "rb") as fp:
data = fp.read()
h = hashlib.md5(data).hexdigest()
return h == checksum
def validate(checksum):
"""Decorator that validates the target file."""
def validate_decorator(func):
@wraps(func)
def wrapper(*args, **kwargs):
target = kwargs.get("target_path", None)
if os.path.exists(target) and check_md5sum(target, checksum):
return
out = func(*args, **kwargs)
if not os.path.exists(target):
raise FileNotFoundError("Target file expected at: %s" % target)
if not check_md5sum(target, checksum):
raise ValidationError(target)
return out
return wrapper
return validate_decorator
def write_csv(target_path=None):
count = 0
while count < 5:
count += 1
try:
aapl = yfinance.download(
"AAPL",
start="1996-12-12",
end="2004-05-15",
progress=False,
rounding=False,
threads=False,
)
# Somewhere between 2020-08-26 and 2020-09-07 an API change
# happened (I guess) that changed these prices and the volume by a
# factor of 4. Long term this is a fragile solution, but for now
# this condition ensures that the correct file is generated.
if 0.2131696 <= aapl["Close"][0] <= 0.2131697:
aapl["Open"] = aapl["Open"] * 4
aapl["High"] = aapl["High"] * 4
aapl["Low"] = aapl["Low"] * 4
aapl["Close"] = aapl["Close"] * 4
# Adj Close doesn't adhere to factor 4
aapl["Volume"] = aapl["Volume"] // 4
aapl.round(6).to_csv(target_path, float_format="%.6f")
return
except URLError as err:
print(
"Error occurred (%r) when trying to download csv. Retrying in 5 seconds"
% err,
sys.stderr,
)
time.sleep(5)
@validate(MD5_JSON)
def write_json(csv_path, target_path=None):
with open(csv_path, "r", newline="", encoding="ascii") as fp:
reader = clevercsv.DictReader(
fp, delimiter=",", quotechar="", escapechar=""
)
rows = list(reader)
# offset to ensure drop is visible in sampled series
rows = rows[1:]
if SAMPLE:
rows = [r for i, r in enumerate(rows) if i % SAMPLE == 0]
time = [r["Date"] for r in rows]
close = [float(r["Close"]) for r in rows]
volume = [int(r["Volume"]) for r in rows]
name = "apple"
longname = "Apple Stock"
time_fmt = "%Y-%m-%d"
series = [
{"label": "Close", "type": "float", "raw": close},
{"label": "Volume", "type": "int", "raw": volume},
]
data = {
"name": name,
"longname": longname,
"n_obs": len(time),
"n_dim": len(series),
"time": {
"type": "string",
"format": time_fmt,
"index": list(range(0, len(time))),
"raw": time,
},
"series": series,
}
with open(target_path, "w") as fp:
json.dump(data, fp, indent="\t")
def collect(output_dir="."):
csv_path = os.path.join(output_dir, NAME_CSV)
json_path = os.path.join(output_dir, NAME_JSON)
write_csv(target_path=csv_path)
write_json(csv_path, target_path=json_path)
def clean(output_dir="."):
csv_path = os.path.join(output_dir, NAME_CSV)
json_path = os.path.join(output_dir, NAME_JSON)
if os.path.exists(csv_path):
os.unlink(csv_path)
if os.path.exists(json_path):
os.unlink(json_path)
def parse_args():
parser = argparse.ArgumentParser()
parser.add_argument(
"-o", "--output-dir", help="output directory to use", default="."
)
parser.add_argument(
"action",
choices=["collect", "clean"],
help="Action to perform",
default="collect",
nargs="?",
)
return parser.parse_args()
def main(output_dir="."):
args = parse_args()
if args.action == "collect":
collect(output_dir=args.output_dir)
elif args.action == "clean":
clean(output_dir=args.output_dir)
if __name__ == "__main__":
main()
|
