def pulldata(ticker, interval, depth):
Cdata = client.get_historical_klines(ticker, interval, depth)
print(Cdata)
df = pd.DataFrame(Cdata)
if not df.empty:
df[0] = pd.to_datetime(df[0], unit='ms')
df.columns = [
'Date', 'Open', 'High', 'Low', 'Close', 'Volume', 'IGNORE',
'Quote_Volume', 'Trades_Count', 'BUY_VOL', 'BUY_VOL_VAL', 'x'
]
df = df.set_index('Date')
del df['IGNORE']
del df['BUY_VOL']
del df['BUY_VOL_VAL']
del df['x']
df["Open"] = pd.to_numeric(df["Open"])
df["Open"] = pd.to_numeric(df["Open"])
df["High"] = pd.to_numeric(df["High"])
df["Low"] = pd.to_numeric(df["Low"])
df["Close"] = pd.to_numeric(df["Close"])
df["Volume"] = round(pd.to_numeric(df["Volume"]))
df["Quote_Volume"] = round(pd.to_numeric(df["Quote_Volume"]))
df["Trades_Count"] = pd.to_numeric(df["Trades_Count"])
df['div'] = df['Open'] / df['Close']
df['Log_VolumeGain'] = (
np.log(df["Quote_Volume"] / df.Quote_Volume.shift(1)) *
100).fillna(0)
df['pricegain'] = (df.Open.pct_change() * 100).fillna(0)
df.to_csv(f'files/{ticker}.csv')
print(df)
def pullData(ticker, interval, depth):
raw = client.get_historical_klines(ticker, interval, depth)
raw = pd.DataFrame(raw)
print(raw)
if not raw.empty:
raw[0] = pd.to_datetime(raw[0], unit='ms')
print(raw)
raw.columns = [
'timestamp', 'open', 'high', 'low', 'close', 'volume', 'IGNORE',
'quoteVolume', 'SELLVolume', 'BUY_VOL', 'BUY_VOL_VAL', 'x'
]
del raw['IGNORE']
del raw['BUY_VOL']
del raw['BUY_VOL_VAL']
del raw['x']
del raw['SELLVolume']
# convert to numbers
raw["open"] = pd.to_numeric(raw["open"])
raw["high"] = pd.to_numeric(raw["high"])
raw["low"] = pd.to_numeric(raw["low"])
raw["close"] = pd.to_numeric(raw["close"])
raw["volume"] = round(pd.to_numeric(raw["volume"]))
raw["quoteVolume"] = round(pd.to_numeric(raw["quoteVolume"]))
raw.loc[raw.quoteVolume < 100, 'quoteVolume'] = 100
raw['pchange1h'] = raw.close.diff(1).fillna(
0) # diff can has if for different timeperiods
raw['pchange1hpct'] = round((raw['pchange1h'] / raw["close"]) * 100, 2)
raw['pchange24h'] = raw.close.diff(23).fillna(
0) # diff can has if for different timeperiods
raw['pchange24hpct'] = round((raw['pchange24h'] / raw["close"]) * 100,
2)
raw['v1h'] = raw.quoteVolume.rolling(window=1).sum().fillna(
0) #.shift()
raw['vchange1h'] = raw.v1h.diff(1).fillna(
0) # diff can has if for different timeperiods
raw['vchange1hpct'] = round(
(raw['vchange1h'] / raw["quoteVolume"]) * 100, 2)
raw['v4h'] = raw.quoteVolume.rolling(window=4).sum().fillna(
0) #.shift()
raw['vchange4h'] = raw.v4h.diff(4).fillna(
0) # diff can has if for different timeperiods
raw['vchange4hpct'] = round(
(raw['vchange4h'] / raw["quoteVolume"]) * 100, 2)
raw['v24'] = raw.quoteVolume.rolling(window=23).sum().fillna(
0) #.shift()
raw['vchange24h'] = raw.v24.diff(23).fillna(
0) # diff can has if for different timeperiods
raw['vchange24hpct'] = round(
(raw['vchange24h'] / raw["quoteVolume"]) * 100, 2)
print(raw)
return raw
def get_coin_data(get_flag="",
path="",
name="",
pair="",
start_date="",
end_date="",
convert_date=convert_date):
"""Collects coin price data from binance API saves a .csv file.
Arguments:
get_flag {str}: If left blank the function defaults to the whats in the config file.
path {str}: where .csv will be saved. The function defaults to to the path data/coins/ if path is blank.
pair {str}: cryptocurrency pair like: "BTCUSDT" = Bitcoin and Tether.
start_date {str}: Date in the ISO 8601 format *with hour* YYYY-MM-DDT01 - Can be left blank the default is "2015-05-01T00" for collect.
end_date {str}:Date in the ISO 8601 format *with hour* YYYY-MM-DDT01
Returns:
Nothing.
"""
##set up
jfile = open("config.json")
config = json.load(jfile)
fname = "{}.csv".format(name.replace(" ", "-"))
if path == "":
path = os.path.join(config["data"]["base_path"], "coins/")
if get_flag == "":
get_flag = config["data"]["coins_get"]
jfile.close()
if get_flag == "collect":
status = "w+"
if start_date == "":
start_date == "May 1, 2015 12:00 AM EST"
elif get_flag == "append":
write_status = "a+"
###get start date from file
if start_date != "":
print(
"Warning the start_date should be left blank for appending so that the last end_date is used."
)
if datetime.strptime(start_date,
"%Y-%m-%dT%H") >= datetime.strptime(
end_date, "%Y-%m-%dT%H"):
print("Data acquisition already satisfied")
return False
elif start_date == "":
coin_path = os.path.join(config["data"]["base_path"],
"coins/{}".format(fname))
coindf = pd.read_csv(coin_path)
start_date = coindf["open_time_iso"].iloc[-1:].values[
0][:-6].replace(" ", "T")
print(start_date)
if datetime.strptime(start_date,
"%Y-%m-%dT%H") >= datetime.strptime(
end_date, "%Y-%m-%dT%H"):
print("Data acquisition already satisfied")
return False
print("Getting data for Coin: {}, Pair: {}, dates: {} to {}".format(
name, pair, start_date, end_date))
start_date, end_date = convert_date(start_date), convert_date(end_date)
print(start_date, end_date)
client = Client()
klines = client.get_historical_klines(pair, Client.KLINE_INTERVAL_1HOUR,
start_date, end_date)
save_path = os.path.join(path, fname)
print("Got data. Now {}ing...".format(get_flag))
with open(save_path, write_status) as f:
if write_status == "collect":
f.write(
"open_time_iso,open_time_unix,open,high,low,close,volume,close_time,number_of_trades\n"
)
for kline in klines:
f.write("{},{},{},{},{},{},{},{},{}\n".format(
datetime.fromtimestamp(kline[0] / 1000).isoformat().replace(
"T", " "), kline[0], kline[1], kline[2], kline[3],
kline[4], kline[5], kline[6], kline[8]))
print("Data saved here {}.".format(save_path))
def get_historical_data(*, client, symbol, interval):
"""
Return historical pricing data from the Binance Exchange.
Parameters:
client (binance.client.Client): An open connected client to the Binance Exchange API.
symbol (str): A Crypto-Pair symbol.
interval (str): An OHLC candlestick width.
Returns:
pandas.core.frame.DataFrame: Historical pricing data for the given symbol and interval.
"""
raw = client.get_historical_klines(
symbol=symbol,
interval=interval,
start_str='2019-01-01 00:00:00'
)
df = pandas.DataFrame(
data=raw[:-1],
columns=[
'OPEN_TIME', 'OPEN', 'HIGH', 'LOW', 'CLOSE', 'VOLUME', 'CLOSE_TIME',
'QUOTE_ASSET_VOLUME', 'NUMBER_TRADES', 'TAKER_BASE_ASSET_VOLUME', 'TAKER_QUOTE_ASSET_VOLUME', 'IGNORE'
]
).drop(
columns=['QUOTE_ASSET_VOLUME', 'TAKER_BASE_ASSET_VOLUME', 'TAKER_QUOTE_ASSET_VOLUME', 'IGNORE']
).sort_values(
by=['OPEN_TIME']
).reset_index(
drop=True
).rename(
{
'OPEN_TIME': 'open_time',
'OPEN': 'open',
'HIGH': 'high',
'LOW': 'low',
'CLOSE': 'close',
'NUMBER_TRADES': 'number_trades',
'VOLUME': 'volume',
'CLOSE_TIME': 'close_time'
},
axis='columns'
)
df['symbol'] = symbol
df['width'] = interval
df['open_time'] = pandas.to_datetime(df['open_time'], unit='ms')
df['close_time'] = pandas.to_datetime(df['close_time'], unit='ms')
def format_hour(timestamp):
timestamp = timestamp.replace(second=0, microsecond=0)
if timestamp.minute == 59:
timestamp += numpy.timedelta64(1, 'm')
timestamp.replace(minute=0)
return timestamp
df['open_time'] = df['open_time'].map(lambda x: format_hour(x))
df['close_time'] = df['close_time'].map(lambda x: format_hour(x))
df['open'] = df['open'].astype(float)
df['high'] = df['high'].astype(float)
df['low'] = df['low'].astype(float)
df['close'] = df['close'].astype(float)
df['volume'] = df['volume'].astype(float)
df['number_trades'] = df['number_trades'].astype(float)
return df
def getHistorical():
#earliestPossible = client._get_earliest_valid_timestamp("ETHUSDT", timeframe)
historical = client.get_historical_klines("ETHUSDT", timeframe,
"2 weeks ago UTC")
return (historical)