def adjust_for_register(self, weights):
new_map = {}
for k in weights.keys():
distance = abs(self.register_center-k)+1
new_map[k] = pow(distance, - self.register)
weights = w.apply(new_map, weights)
return w.ensure_safe(weights)
def adjust_for_uniformity(self, weights):
new_map = {}
for k in weights.keys():
occurrences = self.prev_picks.count(k)
new_map[k] = pow(2, - self.uniformity * occurrences)
weights = w.apply(new_map, weights)
return w.ensure_safe(weights)
def adjust_for_directionality(self, weights):
_PAST_NOTES_TO_COUNT = 6
new_map = {}
direction_count = {'down':0, 'same':0, 'up':0}
lpp = len(self.prev_picks)
if lpp == 0:
return weights
for i in range(0,min(_PAST_NOTES_TO_COUNT, lpp-1)):
picks = self.prev_picks[lpp-i-2:lpp-i]
if picks[1] < picks[0]:
direction_count['down'] += 1
elif picks[1] > picks[0]:
direction_count['up'] += 1
else:
direction_count['same'] += 1
direction_difference = direction_count['up'] - direction_count['down']
score = _PAST_NOTES_TO_COUNT - abs(direction_difference) + 1
sign = lambda x,y: x < y
if direction_difference < 0:
sign = lambda x,y: x > y
for k in weights.keys():
prev_pick = self.prev_picks[-1:][0]
if sign(prev_pick, k):
new_map[k] = pow(score, self.directionality)
weights = w.apply(new_map, weights)
return w.ensure_safe(weights)
def adjust_for_directionality(self, weights):
_PAST_NOTES_TO_COUNT = 6
new_map = {}
direction_count = {'down': 0, 'same': 0, 'up': 0}
lpp = len(self.prev_picks)
if lpp == 0:
return weights
for i in range(0, min(_PAST_NOTES_TO_COUNT, lpp - 1)):
picks = self.prev_picks[lpp - i - 2:lpp - i]
if picks[1] < picks[0]:
direction_count['down'] += 1
elif picks[1] > picks[0]:
direction_count['up'] += 1
else:
direction_count['same'] += 1
direction_difference = direction_count['up'] - direction_count['down']
score = _PAST_NOTES_TO_COUNT - abs(direction_difference) + 1
sign = lambda x, y: x < y
if direction_difference < 0:
sign = lambda x, y: x > y
for k in weights.keys():
prev_pick = self.prev_picks[-1:][0]
if sign(prev_pick, k):
new_map[k] = pow(score, self.directionality)
weights = w.apply(new_map, weights)
return w.ensure_safe(weights)
def adjust_for_variance(self, weights):
map = {}
prev_picks = self.prev_picks[-self.max_length:]
prev_picks_length = len(self.prev_picks)
for k in weights.keys():
occurrence_score = sum([1/pow(prev_picks_length - index, 0.5)
for index,val in enumerate(self.prev_picks) if self.identity(k, val)])
map[k] = pow(2, (- self.variance * occurrence_score))
return w.ensure_safe(w.apply(map, weights))
def adjust_for_linearity(self, weights):
new_map = {}
for k in weights.keys():
if len(self.prev_picks) == 0:
distance = 0
else:
distance = abs(k - self.prev_picks[-1])
new_map[k] = pow(2, (- self.linearity * distance))
weights = w.apply(new_map, weights)
return w.ensure_safe(weights)
def adjust_for_linearity(self, weights):
new_map = {}
for k in weights.keys():
if len(self.prev_picks) == 0:
distance = 0
else:
distance = abs(k - self.prev_picks[-1])
new_map[k] = pow(2, (-self.linearity * distance))
weights = w.apply(new_map, weights)
return w.ensure_safe(weights)
def adjust_for_leads(self, weights):
map = {}
if len(self.prev_picks) == 0: return weights
prev = self.prev_picks[-1]
if (prev not in self.leads.keys()): return weights
leadChoices = self.leads[prev]
if len(set(leadChoices.keys()).union(weights.keys())) == 0:
return weights # lead choices and weights have no options in common
for k,v in weights.items():
if k in leadChoices:
map[k] = v
else:
map[k] = 0
return w.ensure_safe(w.apply(map, weights))
def adjust_for_spacing(self, weights):
new_map = {k:1 for k in weights.keys()}
for prev in self.prev_picks:
octave = max(0, prev - 36) / 12
too_close = 0
if octave == 0:
too_close = 11
elif octave == 1:
too_close = 4
elif octave == 2:
too_close = 2
for k in weights.keys():
if abs(prev - k) <= too_close:
new_map[k] *= pow(0.5, self.spacing)
weights = w.apply(new_map, weights)
return w.ensure_safe(weights)
def adjust_for_balance(self,weights):
location = 0.
ons = 0
offs = 0
new_map = {}
for pick in self.prev_picks:
location += pick
if int(location) == location:
ons += 1
else:
offs += 1
#print "offs",offs,"ons",ons
for k in weights:
if int(k) != k and int(location) != location or int(k) == k and int(location) == location:
new_map[k] = pow(2, offs-ons)
else:
new_map[k] = pow(2, ons-offs)
weights = w.apply(new_map, weights)
#print weights
return w.ensure_safe(weights)
def adjust_for_uniqueness(self, weights):
for k, v in weights.items():
if k in self.prev_picks:
weights[k] = v * pow(2., - self.uniqueness)
return w.ensure_safe(weights)
