class DataFeed(object):
"""
API for using database quotes into application
"""
def __init__(self, level='debug'):
self.stock_db = Client()
def quotes(self, tickers, start_date=None, end_date=None, download=False):
""" Get a series of quotes
Return a list of quotes for the given security from start_date to
end-date
:param ticker: Ticker symbol of the security to quote.
:param start_date: Starting date of quote list as a string or
``Datetime.date`` object.
:param end_date: Ending date of quote list as a string or a
``Datetime.date`` object.
:returns: List of quotes for the given security and date range
"""
#NOTE Always retrieve adjusted close ?
#TODO Frequency management
#TODO No available symbols should be downloaded (with a parameter)
if isinstance(tickers, str):
tickers = [tickers]
df_tmp = dict()
for ticker in tickers:
symbol = self.guess_name(ticker)
log.info('Retrieving {} quotes from database'.format(ticker))
data = self.stock_db.get_quotes(symbol, start_date=start_date, end_date=end_date, dl=download)
if data is None:
continue
index = pd.DatetimeIndex([data[i].Date for i in range(len(data))])
index = index.tz_localize(pytz.utc)
#NOTE Symbol or name as columns ?
df_tmp[ticker] = pd.Series([data[i].AdjClose for i in range(len(data))], index=index)
return pd.DataFrame(df_tmp)
def infos(self, ticker):
symbol = self.guess_name(ticker)
return self.stock_db.get_infos(symbol=symbol)
def random_stocks(self, n=5, exchange=None):
''' Return n random stock names from database '''
log.info('Generating {} random stocks'.format(n))
stocks = []
if isinstance(exchange, str):
exchange = [exchange]
for ex in exchange:
stocks.extend(self.stock_db.available_stocks(exchange=ex))
random.shuffle(stocks)
if n > len(stocks):
log.warning('{} asked symbols but only {} availables'.format(n, len(stocks)))
n = len(stocks)
return stocks[:n]
def get_universe(self, exchange=None, limit=10):
'''
Get a set of stocks, as the universe on which we will trade
___________________________________________________________
Parameters
exhange: str or list(...)
Market(s) to search for stocks
limit: int
Limit of stocks to provide (big amount can make the simulation a mess)
___________________________________________________________
Return
stocks: list(limit | max found)
Provide stock list according to specified criteria
'''
log.info('Generating stock universe')
return self.random_stocks(n=limit, exhange=exchange)
def guess_name(self, partial_input):
''' Find the closest math of partial_input in stocks database, and return its symbol '''
match = [name for name in self.stock_db.available_stocks(key='name') if re.match(partial_input, name, re.IGNORECASE) is not None]
if not match:
log.debug('No matching name, trying symbol list...')
match = [name for name in self.stock_db.available_stocks(key='symbol') if re.match(partial_input, name, re.IGNORECASE) is not None]
if match:
infos = self.stock_db.get_infos(symbol=match[0])
else:
return None
else:
infos = self.stock_db.get_infos(name=match[0])
return infos.Ticker
class DataFeed(object):
"""
API for using database quotes into application
"""
def __init__(self, level='debug'):
self.stock_db = Client()
def quotes(self, tickers, start_date=None, end_date=None, download=False):
""" Get a series of quotes
Return a list of quotes for the given security from start_date to
end-date
:param ticker: Ticker symbol of the security to quote.
:param start_date: Starting date of quote list as a string or
``Datetime.date`` object.
:param end_date: Ending date of quote list as a string or a
``Datetime.date`` object.
:returns: List of quotes for the given security and date range
"""
#NOTE Always retrieve adjusted close ?
#TODO Frequency management
#TODO No available symbols should be downloaded (with a parameter)
if isinstance(tickers, str):
tickers = [tickers]
df_tmp = dict()
for ticker in tickers:
symbol = self.guess_name(ticker)
log.info('Retrieving {} quotes from database'.format(ticker))
data = self.stock_db.get_quotes(symbol, start_date=start_date, end_date=end_date, dl=download)
if data is None:
continue
index = pd.DatetimeIndex([data[i].Date for i in range(len(data))])
index = index.tz_localize(pytz.utc)
#NOTE Symbol or name as columns ?
df_tmp[ticker] = pd.Series([data[i].AdjClose for i in range(len(data))], index=index)
return pd.DataFrame(df_tmp)
def infos(self, ticker):
symbol = self.guess_name(ticker)
return self.stock_db.get_infos(symbol=symbol)
def random_stocks(self, n=5):
''' Return n random stock names from database '''
log.info('Generating {} random stocks'.format(n))
stocks = self.stock_db.available_stocks()
random.shuffle(stocks)
if n > len(stocks):
log.warning('{} asked symbols but only {} availables'.format(n, len(stocks)))
n = len(stocks)
return stocks[:n]
def guess_name(self, partial_input):
#NOTE Everything could work with symbols, under this interface
''' Find the closest math of partial_input in stocks database, and return its symbol '''
match = [name for name in self.stock_db.available_stocks(key='name') if re.match(partial_input, name, re.IGNORECASE) is not None]
if not match:
log.debug('No matching name, trying symbol list...')
match = [name for name in self.stock_db.available_stocks(key='symbol') if re.match(partial_input, name, re.IGNORECASE) is not None]
if match:
infos = self.stock_db.get_infos(symbol=match[0])
else:
return partial_input
else:
infos = self.stock_db.get_infos(name=match[0])
return infos.Ticker
