def create_new_apple(self, position=None):
if position:
if not self._is_captured_by_snake(position):
self._apple_pos = position
return
# else: a new apple has been created because a snake bytes the old one
while True:
apple = (random_range(0, BoardState.get_cols() - 1), random_range(0, BoardState.get_rows() - 1))
if not self._is_captured_by_snake(apple):
self._apple_pos = apple
break
self.leave_tail = True # for the next generations
def init_random_snake(self):
apple = (random_range(0, BoardState.get_cols() - 1), random_range(0, BoardState.get_rows() - 1))
self._apple_pos = apple
snake_head = None
while True:
snake_head = (
random_range(0, BoardState.get_cols() - 1),
random_range(0, BoardState.get_rows() - 1))
if apple != snake_head:
self._snake_positions_list.append(snake_head)
break
snake_cur_pos = snake_head
for snake_body_depth in range(BoardState.TILE_SNAKE_BODY_DEFAULT, SNAKE_INIT_LENGTH+1):
snake_next_pos = random_from_seq(list(self._get_closest_blank_points(snake_cur_pos)))
self._snake_positions_list.append(snake_next_pos)
snake_cur_pos = snake_next_pos
def random(image, clength, **kwargs):
intervals = np.array([])
for y in range(image.shape[1]):
intervals.append([])
x = 0
while True:
x += int(clength * random_range())
if x > image.shape[0]:
break
else:
intervals[y].append(x)
return intervals
def build_fwtrack(self, opt):
"""Build FWTrackII from all lines, return a FWTrackII object.
Handle multi-reads here by building a probability and enrichment index
or select only one alignment from each multi-read.
Initial alignment probabilities are set from read/mismatch qualities
or from a uniform distribution.
"""
fwtrack = FWTrackII()
i = 0
m = 0
read_total = 0
recent_tags = []
random_select_one_multi = opt.random_select_one_multi
no_multi_reads = opt.no_multi_reads
min_score = opt.min_score
prior_prob_snp = opt.prior_prob_snp
no_prior_prob_map = opt.no_prior_prob_map
if opt.qual_scale == 'auto':
opt.qual_scale = self._guess_qual_scale()
if opt.qual_scale == 'sanger+33':
qual_offset = 33
elif opt.qual_scale == 'illumina+64':
qual_offset = 64
group_starts_append = fwtrack.group_starts.append
fwtrack_add_loc = fwtrack.add_loc
match_probs = {
} # {(1,30):p(match|phred=30), (0,30):p(mismatch|phred=30)}
for grouplines in self._group_by_name(self.fhd):
read_total += 1 # in ratios, only count reads, not total alignments
if len(grouplines) == 1:
# uniquely mapping reads
i += 1
if i == 1000000:
m += 1
logging.info(" %d alignments read." % (m * 1000000))
i = 0
chromosome, fpos, strand, qualstr, mismatches = grouplines[0]
fwtrack_add_loc(chromosome, fpos, strand,
0) # 0'th index => unique
else:
if no_multi_reads: # throw away multi-reads
fwtrack.total -= 1
continue
elif random_select_one_multi: # choose one alignment at random
i += 1
if i == 1000000:
m += 1
logging.info(" %d alignments read." % (m * 1000000))
i = 0
randline = grouplines[random_range(len(grouplines))]
chromosome, fpos, strand, qualstr, mismatches = randline
fwtrack_add_loc(chromosome, fpos, strand, 0)
else: # use all alignments probabilistically
group_starts_append(
fwtrack.total_multi +
1) # starts at 1 (0 reserved for unique reads)
if no_prior_prob_map:
# don't use map quality; just assume uniform priors
for (chromosome, fpos, strand, qualstr,
mismatches) in grouplines:
i += 1
if i == 1000000:
m += 1
logging.info(" %d alignments read." %
(m * 1000000))
i = 0
fwtrack.total_multi += 1
fwtrack_add_loc(chromosome, fpos, strand,
fwtrack.total_multi)
normed_probs = [1. / len(grouplines)] * len(grouplines)
else:
# TODO: might want to be working in log space-- if many mismatches, we'll lose precision
qualstr = grouplines[0][
3] # all quality strings are shared across the group
group_total_prob = 0.
group_probs = []
group_probs_append = group_probs.append
for (chromosome, fpos, strand, qualstr,
mismatches) in grouplines:
i += 1
if i == 1000000:
m += 1
logging.info(" %d alignments read." %
(m * 1000000))
i = 0
fwtrack.total_multi += 1
fwtrack_add_loc(chromosome, fpos, strand,
fwtrack.total_multi)
mismatches = set(mismatches)
read_prob = 1.
# P(SNP) = prior probability a SNP occurs at any base
# P(SE) = probability there was a sequencing error (from PHRED)
# _P(Map|SNP,SE)__MATCH__SNP__SE_
# 0 0 0 0 # can't map here without explanation
# 1 0 0 1
# 1 0 1 0
# 1 0 1 1
# 1 1 0 0
# 1 1 0 1
# 0 1 1 0 # wouldn't map here if SNP, but sequencer read reference
# 1 1 1 1
# we are interested in P(Mapping | Match), which is equivalent to:
# \Sum_{SNP \in {0,1}, SE \in {0,1}} p(SNP) * p(SE) * p(Map|SE,SNP), or:
# p(Map|match = 0):
# p(SE) + p(SNP) + p(SE)*p(SNP)
# p(Map|match = 1):
# 1 - (p(SE) + p(SE)*p(SNP))
for b in xrange(len(qualstr)):
tup = (b in mismatches, qualstr[b])
if tup in match_probs:
prob = match_probs[tup]
elif tup[0]: # mismatch
p_seq_error = 10.**(
(qualstr[b] - qual_offset) / -10.)
prob = p_seq_error + prior_prob_snp + p_seq_error * prior_prob_snp
match_probs[tup] = prob
else: # match
p_seq_error = 10.**(
(qualstr[b] - qual_offset) / -10.)
prob = 1. - (p_seq_error +
p_seq_error * prior_prob_snp)
match_probs[tup] = prob
read_prob *= prob
# quick & dirty check-- only looking at last base
assert qualstr[
b] >= qual_offset # Specified quality scale yielded a negative phred score! You probably have the wrong PHRED scale!
assert 0. <= read_prob <= 1. # error with map qualities
#raise BaseQualityError("Specified quality scale yielded a negative phred score! You probably have the wrong PHRED scale!")
group_probs_append(read_prob)
group_total_prob += read_prob
normed_probs = [
p / group_total_prob for p in group_probs
]
fwtrack.prob_aligns.extend(normed_probs)
fwtrack.prior_aligns.extend(normed_probs)
fwtrack.enrich_scores.extend([min_score] * len(grouplines))
fwtrack.total = read_total # overwrite the running total, counting each read once
return fwtrack
def generate_pseudo_random_dates_and_times(
self, num_pseudo_random_dates_and_times_to_generate: int,
weeks_max_delta: int, days_max_delta: int, hours_max_delta: int,
minutes_max_delta: int, seconds_max_delta: int,
milliseconds_max_delta: int, microseconds_max_delta: int) -> list:
"""
Generate and return a given number of Pseudo-Random Dates and Times,
starting from a given delta Date and Time, for a generator/list.
:param num_pseudo_random_dates_and_times_to_generate: the number of Pseudo-Random
Dates and Times to be generated.
:param weeks_max_delta: the delta number for weeks.
:param days_max_delta: the delta number for days.
:param hours_max_delta: the delta number for hours.
:param minutes_max_delta: the delta number for minutes.
:param seconds_max_delta: the delta number for seconds.
:param milliseconds_max_delta: the delta number for milliseconds.
:param microseconds_max_delta: the delta number for microseconds.
:return date_and_time_initialisation_plus_pseudo_random_delta: the Pseudo-Random Timestamps,
starting from a given delta Date and Time,
for a generator/list.
"""
date_and_time_initialisation = \
self.get_date_and_time_initialisation()
"""
Retrieve the date and time of initialisation of the Qiskrypt's Timestamp Generator.
"""
while num_pseudo_random_dates_and_times_to_generate > 0:
"""
While there are still Pseudo-Random Dates and Times to be generated.
"""
if weeks_max_delta == 0:
"""
If the maximum delta number for weeks is not set up,
will not be defined any pseudo-random delta for weeks.
"""
weeks_random_delta = 0
"""
Set the pseudo-random delta for weeks as zero (0).
"""
else:
"""
If the maximum delta number for weeks is set up,
will be defined a pseudo-random delta for the number of weeks.
"""
weeks_random_delta = random_range(weeks_max_delta)
"""
Set the pseudo-random delta for the number of weeks,
between zero (0) and the maximum delta number for weeks.
"""
if days_max_delta == 0:
"""
If the maximum delta number for days is not set up,
will not be defined any pseudo-random delta for days.
"""
days_random_delta = 0
"""
Set the pseudo-random delta for days as zero (0).
"""
else:
"""
If the maximum delta number for days is set up,
will be defined a pseudo-random delta for the number of days.
"""
days_random_delta = random_range(days_max_delta)
"""
Set the pseudo-random delta for the number of days,
between zero (0) and the maximum delta number for days.
"""
if hours_max_delta == 0:
"""
If the maximum delta number for hours is not set up,
will not be defined any pseudo-random delta for hours.
"""
hours_random_delta = 0
"""
Set the pseudo-random delta for hours as zero (0).
"""
else:
"""
If the maximum delta number for hours is set up,
will be defined a pseudo-random delta for the number of hours.
"""
hours_random_delta = random_range(days_max_delta)
"""
Set the pseudo-random delta for the number of hours,
between zero (0) and the maximum delta number for hours.
"""
if minutes_max_delta == 0:
"""
If the maximum delta number for minutes is not set up,
will not be defined any pseudo-random delta for minutes.
"""
minutes_random_delta = 0
"""
Set the pseudo-random delta for minutes as zero (0).
"""
else:
"""
If the maximum delta number for minutes is set up,
will be defined a pseudo-random delta for the number of minutes.
"""
minutes_random_delta = random_range(minutes_max_delta)
"""
Set the pseudo-random delta for the number of minutes,
between zero (0) and the maximum delta number for minutes.
"""
if seconds_max_delta == 0:
"""
If the maximum delta number for seconds is not set up,
will not be defined any pseudo-random delta for seconds.
"""
seconds_random_delta = 0
"""
Set the pseudo-random delta for seconds as zero (0).
"""
else:
"""
If the maximum delta number for seconds is set up,
will be defined a pseudo-random delta for the number of seconds.
"""
seconds_random_delta = random_range(seconds_max_delta)
"""
Set the pseudo-random delta for the number of seconds,
between zero (0) and the maximum delta number for seconds.
"""
if milliseconds_max_delta == 0:
"""
If the maximum delta number for milliseconds is not set up,
will not be defined any pseudo-random delta for milliseconds.
"""
milliseconds_random_delta = 0
"""
Set the pseudo-random delta for milliseconds as zero (0).
"""
else:
"""
If the maximum delta number for milliseconds is set up,
will be defined a pseudo-random delta for the number of milliseconds.
"""
milliseconds_random_delta = random_range(
milliseconds_max_delta)
"""
Set the pseudo-random delta for the number of milliseconds,
between zero (0) and the maximum delta number for milliseconds.
"""
if microseconds_max_delta == 0:
"""
If the maximum delta number for microseconds is not set up,
will not be defined any pseudo-random delta for microseconds.
"""
microseconds_random_delta = 0
"""
Set the pseudo-random delta for microseconds as zero (0).
"""
else:
"""
If the maximum delta number for microseconds is set up,
will be defined a pseudo-random delta for the number of microseconds.
"""
microseconds_random_delta = random_range(
microseconds_max_delta)
"""
Set the pseudo-random delta for the number of microseconds,
between zero (0) and the maximum delta number for microseconds.
"""
date_and_time_initialisation_plus_pseudo_random_delta = date_and_time_initialisation + \
time_delta(weeks=weeks_random_delta, days=days_random_delta, hours=hours_random_delta,
minutes=minutes_random_delta, seconds=seconds_random_delta,
milliseconds=milliseconds_random_delta, microseconds=microseconds_random_delta)
"""
Compute the current date and time of initialisation of the Qiskrypt's Timestamp Generator,
with a given delta Date and Time, considering the pseudo-random
weeks, days, hours, minutes, seconds, milliseconds and microseconds computed previously.
"""
yield date_and_time_initialisation_plus_pseudo_random_delta
"""
Yield the current date and time of initialisation of the Qiskrypt's Timestamp Generator,
with a given delta Date and Time, considering the pseudo-random
weeks, days, hours, minutes, seconds, milliseconds and microseconds computed previously,
for a generator/list.
"""
num_pseudo_random_dates_and_times_to_generate -= 1
"""
def build_fwtrack (self, opt):
"""Build FWTrackII from all lines, return a FWTrackII object.
Handle multi-reads here by building a probability and enrichment index
or select only one alignment from each multi-read.
Initial alignment probabilities are set from read/mismatch qualities
or from a uniform distribution.
"""
fwtrack = FWTrackII()
i = 0
m = 0
read_total = 0
recent_tags = []
random_select_one_multi = opt.random_select_one_multi
no_multi_reads = opt.no_multi_reads
min_score = opt.min_score
prior_prob_snp = opt.prior_prob_snp
no_prior_prob_map = opt.no_prior_prob_map
if opt.qual_scale == 'auto':
opt.qual_scale = self._guess_qual_scale()
if opt.qual_scale == 'sanger+33':
qual_offset = 33
elif opt.qual_scale == 'illumina+64':
qual_offset = 64
group_starts_append = fwtrack.group_starts.append
fwtrack_add_loc = fwtrack.add_loc
match_probs = {} # {(1,30):p(match|phred=30), (0,30):p(mismatch|phred=30)}
for grouplines in self._group_by_name(self.fhd):
read_total += 1 # in ratios, only count reads, not total alignments
if len(grouplines) == 1:
# uniquely mapping reads
i+=1
if i == 1000000:
m += 1
logging.info(" %d alignments read." % (m*1000000))
i=0
chromosome, fpos, strand, qualstr, mismatches = grouplines[0]
fwtrack_add_loc(chromosome,fpos,strand,0) # 0'th index => unique
else:
if no_multi_reads: # throw away multi-reads
fwtrack.total -= 1
continue
elif random_select_one_multi: # choose one alignment at random
i+=1
if i == 1000000:
m += 1
logging.info(" %d alignments read." % (m*1000000))
i=0
randline = grouplines[random_range(len(grouplines))]
chromosome,fpos,strand,qualstr,mismatches = randline
fwtrack_add_loc(chromosome,fpos,strand,0)
else: # use all alignments probabilistically
group_starts_append(fwtrack.total_multi + 1) # starts at 1 (0 reserved for unique reads)
if no_prior_prob_map:
# don't use map quality; just assume uniform priors
for (chromosome,fpos,strand, qualstr,mismatches) in grouplines:
i+=1
if i == 1000000:
m += 1
logging.info(" %d alignments read." % (m*1000000))
i=0
fwtrack.total_multi += 1
fwtrack_add_loc(chromosome,fpos,strand,
fwtrack.total_multi)
normed_probs = [1./len(grouplines)] * len(grouplines)
else:
# TODO: might want to be working in log space-- if many mismatches, we'll lose precision
qualstr = grouplines[0][3] # all quality strings are shared across the group
group_total_prob = 0.
group_probs = []
group_probs_append = group_probs.append
for (chromosome,fpos,strand, qualstr, mismatches) in grouplines:
i+=1
if i == 1000000:
m += 1
logging.info(" %d alignments read." % (m*1000000))
i=0
fwtrack.total_multi += 1
fwtrack_add_loc(chromosome,fpos,strand,
fwtrack.total_multi)
mismatches = set(mismatches)
read_prob = 1.
# P(SNP) = prior probability a SNP occurs at any base
# P(SE) = probability there was a sequencing error (from PHRED)
# _P(Map|SNP,SE)__MATCH__SNP__SE_
# 0 0 0 0 # can't map here without explanation
# 1 0 0 1
# 1 0 1 0
# 1 0 1 1
# 1 1 0 0
# 1 1 0 1
# 0 1 1 0 # wouldn't map here if SNP, but sequencer read reference
# 1 1 1 1
# we are interested in P(Mapping | Match), which is equivalent to:
# \Sum_{SNP \in {0,1}, SE \in {0,1}} p(SNP) * p(SE) * p(Map|SE,SNP), or:
# p(Map|match = 0):
# p(SE) + p(SNP) + p(SE)*p(SNP)
# p(Map|match = 1):
# 1 - (p(SE) + p(SE)*p(SNP))
for b in xrange(len(qualstr)):
tup = (b in mismatches,qualstr[b])
if tup in match_probs:
prob = match_probs[tup]
elif tup[0]: # mismatch
p_seq_error = 10. ** ((qualstr[b]-qual_offset)/-10.)
prob = p_seq_error + prior_prob_snp + p_seq_error * prior_prob_snp
match_probs[tup] = prob
else: # match
p_seq_error = 10. ** ((qualstr[b]-qual_offset)/-10.)
prob = 1. - (p_seq_error + p_seq_error * prior_prob_snp)
match_probs[tup] = prob
read_prob *= prob
# quick & dirty check-- only looking at last base
assert qualstr[b] >= qual_offset # Specified quality scale yielded a negative phred score! You probably have the wrong PHRED scale!
assert 0.<=read_prob<=1. # error with map qualities
#raise BaseQualityError("Specified quality scale yielded a negative phred score! You probably have the wrong PHRED scale!")
group_probs_append(read_prob)
group_total_prob += read_prob
normed_probs = [p / group_total_prob for p in group_probs]
fwtrack.prob_aligns.extend(normed_probs)
fwtrack.prior_aligns.extend(normed_probs)
fwtrack.enrich_scores.extend([min_score] * len(grouplines))
fwtrack.total = read_total # overwrite the running total, counting each read once
return fwtrack
