def test_fill_nan_data():
f = 'stock_db/test/SPY.csv'
df = fu.load_data_frame(f, datetime.date(2018, 1, 1), datetime.date(2018, 4, 1))
# Test by adding some NaN
df.iloc[0:10, 0] = np.nan # beginning
df.iloc[11:20, 1] = np.nan # middle
df.iloc[-10:, 2] = np.nan # end
assert df.isna().any().any()
fu.fill_nan_data(df)
assert not df.isna().any().any()
def test_load_data_frame():
f = 'stock_db/test/SPY.csv'
df = fu.load_data_frame(f, datetime.date(2018, 1, 1), datetime.date(2018, 4, 1))
assert len(fu.get_date(df)) > 0
assert len(fu.get_open(df)) > 0
assert len(fu.get_high(df)) > 0
assert len(fu.get_low(df)) > 0
assert len(fu.get_close(df)) > 0
assert len(fu.get_volume(df)) > 0
def test_load_data_frame_no_adj():
f = 'stock_db/test/SPY.csv'
df = fu.load_data_frame(f, datetime.date(2018, 1, 1),
datetime.date(2018, 4, 1), False)
col = list(df.columns)
test_col = ['Open', 'High', 'Low', 'Close', 'Adj Close', 'Volume']
assert len(col) == len(test_col)
for c in test_col:
assert c in col
def _main():
sf = '../stock_db/dj.txt'
print("symbol file {} contains the following stocks: {}".format(sf, fu.get_symbols_from_file(sf)))
d = '../stock_db/test'
s = 'SPY'
f = fu.symbol_to_filename(s, d)
print("symbol {} with directory {} gives filename {}".format(s, d, f))
print("filename {} gives symbol {}".format(f, fu.filename_to_symbol(f)))
print("filename {} gives symbol {}".format(f.upper(), fu.filename_to_symbol(f.upper())))
print("validate_symbol_data {} = {}".format(f, fu.validate_symbol_data(f)))
print("directory {} contains the following stocks: {}".format(d, fu.get_all_symbols(d)))
start_date = datetime.date(1900, 1, 1)
end_date = datetime.date.today()
if False:
fu.download_data(s, d, start_date, end_date)
fu.update_all_symbols(d, start_date, end_date)
df = fu.load_data_frame(f, datetime.date(2018, 1, 1), datetime.date(2018, 4, 1))
print(df.describe())
print(df.head())
print(fu.get_date(df)[0:3])
print(fu.get_open(df)[0:3])
print(fu.get_high(df)[0:3])
print(fu.get_low(df)[0:3])
print(fu.get_close(df)[0:3])
print(fu.get_volume(df)[0:3])
# Not applicable for a single stock, but just to test...
print(fu.normalize_data_frame(df).head())
# Test by adding some NaN
df.iloc[0:10, 0] = np.nan # beginning
df.iloc[11:20, 1] = np.nan # middle
df.iloc[-10:, 2] = np.nan # end
print(df.isna().any())
fu.fill_nan_data(df)
print(df.isna().any())
url_array = [
'https://www.google.ca',
'https://www.bad234123421342134.com',
'https://www.tmall.com',
'https://tmxmoney.com/en/index.html',
]
for u in url_array:
print(fu.download_url(u)[:50])
def test_fill_nan_data_inplace():
f = 'stock_db/test/SPY.csv'
df = fu.load_data_frame(f, datetime.date(2018, 1, 1),
datetime.date(2018, 4, 1))
# Test by adding some NaN
df.iloc[0:10, 0] = np.nan # beginning
df.iloc[11:20, 1] = np.nan # middle
df.iloc[-10:, 2] = np.nan # end
assert df.isna().any().any()
df2 = fu.fill_nan_data(df, inplace=True)
# Original df modified:
assert not df.isna().any().any()
# Returned df2 None
assert df2 is None
def test_fill_nan_data_notinplace():
f = 'stock_db/test/SPY.csv'
df = fu.load_data_frame(f, datetime.date(2018, 1, 1),
datetime.date(2018, 4, 1))
# Test by adding some NaN
df.iloc[0:10, 0] = np.nan # beginning
df.iloc[11:20, 1] = np.nan # middle
df.iloc[-10:, 2] = np.nan # end
assert df.isna().any().any()
df2 = fu.fill_nan_data(df) # default = not inplace
# New df2 modified:
assert not df2.isna().any().any()
# Original df not modified:
assert df.isna().any().any()
def get_symbol_data(self, symbol):
"""Return a single symbol data as a DataFrame."""
if symbol not in self._dic:
f = fu.symbol_to_filename(symbol, self._basedir)
if not os.path.exists(f):
self.download_data(symbol)
df = fu.load_data_frame(f,
self._start_date,
self._end_date,
adjust_price=self._adjust_price)
if df is None:
print("ERROR: data for {} contains error".format(symbol))
self._dic[symbol] = df # Store it for next time
# if data is already there, assume it is up to date (to save repetitive download)
return self._dic[symbol]
def test_normalize_data_frame():
f = 'stock_db/test/SPY.csv'
df = fu.load_data_frame(f, datetime.date(2018, 1, 1), datetime.date(2018, 4, 1))
# Not applicable for a single stock, but just to test...
assert fu.normalize_data_frame(df).iloc[0].mean() == 1.0