def lay_arb_info(sel1, sel2):
"""Get lay arb info"""
# get the market name and print it out:
mark1 = database.DBMaster().ReturnMarkets('SELECT * FROM markets WHERE market_id = ?',
(s1.mid,))[0]
mark2 = database.DBMaster().ReturnMarkets('SELECT * FROM markets WHERE market_id = ?',
(s2.mid,))[0]
print mark1.name
print mark2.name
print olay, s1.name, s1.exid, oback
def __init__(self):
# key to match_cache is event name, value is list of tuples
# (m1, m2) where m1 and m2 are the matching markets (m1 is the
# BDAQ market, m2 is the BF market. The ordering of the
# tuples is the same as the ordering that appears in the GUI.
self._match_cache = {}
# store a dict that maps BDAQ mids to BF mids
self._BDAQmap_cache = {}
# and a dict that maps BF mids to BDAQ mids
self._BFmap_cache = {}
# and a dict that maps BDAQ mids to their market objects
self._BDAQ_cache = {}
# and a dict that maps BF mids to their market objects
self._BF_cache = {}
# singleton that controls DB access
self._dbman = database.DBMaster()
if self.INITDB:
self.init_caches_from_DB()
else:
self.init_caches_empty()
def __init__(self):
# matching selections keys are bdaqmid, values are
# [(bdaq_sels, bf_sels)] where bdaq_sels and bf_sels are lists
# of bdaq and bf selections, ordered as displayed on the BDAQ
# website. Instead of dumping everything from the DB into the
# cache immediately, we add to the cache 'as needed'.
self._match_cache = {}
# map a BF sid to a BDAQ sid; note this has an additional
# layer compared to the BDAQ cache since we must store the
# market id.
self._BFmap_cache = {}
# map a BDAQ sid to selection object. Note: we do not have an
# equivalent cache for BF sid's since these do not map
# uniquely onto selections (an annoying feature of the BF
# API). That is, we need both the BF mid and sid to map to a
# BDAQ sid.
self._BDAQ_cache = {}
# we need to access the market store since when we add
# matching selections, we need to find the BF mid that matches
# the given BDAQ mid.
self._mstore = MatchMarketStore.Instance()
# singleton that controls DB access
self._dbman = database.DBMaster()
if self.INITDB:
self.init_caches_from_DB()
def save_strategies(bdaqids=[], instantonly=True):
"""
Using info in the DB, get strategies and save these to pickle
file.
bdaqids - list of BDAQ market ids allowed. If empty, all market
ids are allowed.
"""
sgroup = strategy.StrategyGroup()
# add strategies here
msels = database.DBMaster().ReturnSelectionMatches(bdaqids)
if bdaqids:
# filter out any strategies not involving one of the markets
# listed in bdaqids
for ms in list(msels):
if ms[0].mid not in bdaqids:
msels.remove(ms)
# alter the prices so that we get instant opp!!
#msels[0][0].backprices[0] = (1.50, 10)
#msels[0][1].layprices[0] = (1.01, 10)
for m in msels:
sgroup.add(cxstrategy.CXStrategy(m[0], m[1], instantonly))
betlog.betlog.debug("Created {0} strategies from DB"\
.format(len(sgroup)))
# pickle the strategy group - the main application will read this
# pickle file.
pickle_stratgroup(sgroup)
return
def __init__(self):
"""Class for storing latest prices."""
# current state of prices
self._prices = {const.BDAQID: {}, const.BFID: {}}
# note newprices is not 'private' since it is accessed by the
# engine; it will be fed to the strategies.
self.newprices = {const.BDAQID: {}, const.BFID: {}}
# for writing to database
self._dbman = database.DBMaster()
def unpickle_stratgroup():
"""Return StrategyGroup object from pickle file."""
f = open(PICKPATH, 'rb')
sgroup = pickle.load(f)
f.close()
# this initialises cursor object
for s in sgroup.strategies:
if hasattr(s, 'dbman'):
s.dbman = database.DBMaster()
return sgroup
def __init__(self):
# the current state of all orders placed since the start of
# the application. The keys to each sub-dictionary are the
# order ids, with values that are the order objects. Note each
# order object will typically be updated multiple times.
self._orders = {const.BDAQID: {}, const.BFID: {}}
# the state of all orders immediately after they were placed.
# Among other things, these order objects are useful since
# they have a 'tplaced' attribute.
self._neworders = {const.BDAQID: {}, const.BFID: {}}
# store mapping of order ref to time placed.
self._tplaced = {const.BDAQID: {}, const.BFID: {}}
# the state of all orders immediately after they were placed.
self._cancelorders = {const.BDAQID: {}, const.BFID: {}}
# for writing to database
self._dbman = database.DBMaster()
# store dicts of cancelled, updated and new orders placed by
# the APIs from the latest tick.
self.latest = [{
const.BDAQID: {},
const.BFID: {}
}, {
const.BDAQID: {},
const.BFID: {}
}, {
const.BDAQID: {},
const.BFID: {}
}]
# store dict of updated orders from last tick (from querying
# order status via the API).
self.latest_updates = {const.BDAQID: {}, const.BFID: {}}
# store dict used by strategies to seach for order references
# immediately after order placed. Since we get the order
# reference when placing bets on BF, we write to this when
# placing bets. But when placing bets on BDAQ, we don't get
# the order reference, so instead this is written when we call
# ListOrdersChangedSince.
self.orders_tosearch = {const.BDAQID: {}, const.BFID: {}}
def __init__(self, ex1sel=None, ex2sel=None, instantonly=True):
"""
ex1sel - BDAQ selection
ex2sel - BF selection
instantonly - if True, only look for 'instant' opportunities.
This means only 'taking' and not 'making'.
"""
# this gives us self.brain, a StateMachine, and self.toplace,
# a dictionary of orders waiting to be placed (see
# strategy.py).
super(CXStrategy, self).__init__()
self.sel1 = ex1sel
self.sel2 = ex2sel
# if this is true, we will only look for 'instant'
# opportunities, when we can back and lay simultaneously.
self.ionly = instantonly
# database interface
self.dbman = database.DBMaster()
# add states
noopp_state = CXStateNoOpp(self)
instantopp_state = CXStateInstantOpp(self)
opp_state = CXStateOpp(self)
lay_placed_state = CXStateLayPlaced(self)
back_placed_state = CXStateLayPlaced(self)
both_placed_state = CXStateBothPlaced(self)
lay_matched_state = CXStateLayMatched(self)
back_matched_state = CXStateBackMatched(self)
both_matched_state = CXStateBothMatched(self)
self.brain.add_state(noopp_state)
self.brain.add_state(instantopp_state)
self.brain.add_state(opp_state)
self.brain.add_state(lay_placed_state)
self.brain.add_state(back_placed_state)
self.brain.add_state(both_placed_state)
self.brain.add_state(lay_matched_state)
self.brain.add_state(back_matched_state)
self.brain.add_state(both_matched_state)
# initialise into noopp state
self.brain.set_state(noopp_state.name)
def __init__(self, deltat):
"""deltat is time between price refreshes in seconds."""
self.clock = betman.all.clock.Clock(deltat)
self.dbman = database.DBMaster()
self.load_strategies()
# market ids for all strategies (for updating prices)
self.marketids = self.stratgroup.get_marketids()
# we store selection objects as a dictionary of dictionaries.
# This contains the selection objects (e.g. a particular
# horse), and the selection objects contain the current price,
# hence the name.
self.prices = {const.BDAQID: {}, const.BFID: {}}
# orders for both exchanges
self.orders = {const.BDAQID: {}, const.BFID: {}}
# call the API functions to refresh prices etc.
self.on_startup()
import matplotlib.pyplot as plt
from betman import database
dbman = database.DBMaster()
mms = dbman.return_market_matches()
bdaqmids = [m[0].id for m in mms]
# load all time series data for all runners in both markets (last pair only).
mymid = 4028996
# get all selection matches for this mid
smatches = dbman.return_selection_matches([mymid])
fig1 = plt.figure()
fig2 = plt.figure()
ax1 = fig1.add_subplot(111)
ax2 = fig2.add_subplot(111)
for (s1, s2) in zip(smatches[0], smatches[1]):
# time series data from betdaq
pdata = dbman.return_selections(
'select * from histselections where exchange_id = ? '
'and market_id = ? and selection_id = ?', (s2.exid, s2.mid, s2.id))
# back price
line, = ax1.plot_date([p.tstamp for p in pdata],
[p.padback[0][0] for p in pdata],
fmt='-')
col = line.get_color()
# lay price
ax1.plot_date([p.tstamp for p in pdata], [p.padlay[0][0] for p in pdata],