def setUp(self):
iv = IntervalNode(Feature(50, 59))
for i in range(0, 110, 10):
if i == 50: continue
f = Feature(i, i + 9)
iv = iv.insert(f)
self.intervals = iv
def processTarget(oId, submatches, primaries, tfotargets, annotations, parameters): ''' process one target region for all targets and their offtargets''' output = "" query = submatches[oId] # list of submatches of the same id # quit if this primary target cannot be related to a primary target region if (not primaries.has_key(oId)): print >> sys.stderr, "[ERROR] submatch does not have a corresponding primary target : %s" % (oId) exit(1) # create an empty root intersect_tree = IntervalNode( -1, -1, -1 ) # accumulate off-target data offtarget_data = [] if (tfotargets.has_key(oId)): # put all off-targets into an interval tree: offtarget_data = tfotargets[oId] # build an interval tree from the rest of the data for tfo_region in offtarget_data.keys(): start, end = tfo_region intersect_tree = intersect_tree.insert(start, end) # query all submatches output = "" for q in query: output += processSubTarget(q, oId, intersect_tree, offtarget_data) return output
def processTarget(oId, submatches, primaries, tfotargets, annotations,
parameters):
''' process one target region for all targets and their offtargets'''
output = ""
query = submatches[oId] # list of submatches of the same id
# quit if this primary target cannot be related to a primary target region
if (not primaries.has_key(oId)):
print >> sys.stderr, "[ERROR] submatch does not have a corresponding primary target : %s" % (
oId)
exit(1)
# create an empty root
intersect_tree = IntervalNode(-1, -1, -1)
# accumulate off-target data
offtarget_data = []
if (tfotargets.has_key(oId)):
# put all off-targets into an interval tree:
offtarget_data = tfotargets[oId]
# build an interval tree from the rest of the data
for tfo_region in offtarget_data.keys():
start, end = tfo_region
intersect_tree = intersect_tree.insert(start, end)
# query all submatches
output = ""
for q in query:
output += processSubTarget(q, oId, intersect_tree, offtarget_data)
return output
def setUp(self):
iv = IntervalNode(Interval(50, 59))
for i in range(0, 110, 10):
if i == 50: continue
f = Interval(i, i + 9)
iv = iv.insert(f)
self.intervals = iv
def setUp(self):
iv = IntervalNode(Feature(1, 2))
self.max = 1000000
for i in range(0, self.max, 10):
f = Feature(i, i)
iv = iv.insert(f)
for i in range(600):
iv = iv.insert(Feature(0, 1))
self.intervals = iv
def setUp(self):
iv = IntervalNode(Interval(1, 2))
self.max = 1000000
for i in range(0, self.max, 10):
f = Interval(i, i)
iv = iv.insert(f)
for i in range(6000):
iv = iv.insert(Interval(0, 1))
self.intervals = iv
def testSimpleIntervals():
#yesterday, yesteray-year ago
td_yesterday = timedelta(days=1)
td_week = timedelta(days=7)
td_month = timedelta(days=30)
td_year = timedelta(days=365)
time_now = datetime.utcnow()
time_start = time_now - td_year
time_end = time_now
time_yesterday = time_now - td_yesterday
time_week = time_now - td_week
time_month = time_now - td_month
time_year = time_now - td_year
tree = IntervalNode(get_seconds(time_start)-1, get_seconds(time_end)+1)
tree.insert(get_seconds(time_yesterday), get_seconds(time_now), other="Today")
tree.insert(get_seconds(time_week), get_seconds(time_yesterday), other ="week-yesterday")
tree.insert(get_seconds(time_month), get_seconds(time_week), other ="month-week")
tree.insert(get_seconds(time_year), get_seconds(time_month), other ="year-month")
res = []
time_check = time_week
tree.intersect(get_seconds(time_check)-1, get_seconds(time_check), lambda x: res.append(x.other))
print res
def testSimpleIntervals(self):
#yesterday, yesteray-year ago
td_yesterday = timedelta(days=1)
td_year = timedelta(days=365)
time_now = datetime.utcnow()
time_start = datetime.utcnow() - td_year
time_end = datetime.utcnow()
time_yesterday = datetime.utcnow() - td_yesterday
tree = IntervalNode(time_start.toordinal(), time_end.toordinal())
tree.insert(time_yesterday.toordinal(), time_now.toordinal(), other="From Yesterday")
tree.insert((time_yesterday - timedelta(days=7)).toordinal(), time_yesterday.toordinal(), other ="Yesterday, week")
tree.insert(time_start.toordinal(), (time_yesterday - timedelta(days=7)).toordinal(), other ="Yesterday, week")
tree.intersect(time_now.utcnow().toordinal(), time_now.utcnow().toordinal(), report_schedule)
#yesterday, day before, day before that, till forever
#yesterday, day before, nothing
pass
def setUp(self):
intervals = []
for i in range(11, 20000, 15):
for zz in range(random.randint(2, 5)):
m = random.randint(1, 10)
p = random.randint(1, 10)
intervals.append(Feature(i - m, i + p))
iv = IntervalNode(intervals[0])
for f in intervals[1:]:
iv = iv.insert(f)
self.intervals = intervals
self.tree = iv
def setUp(self):
intervals = []
for i in range(11, 20000, 15):
for zz in range(random.randint(2, 5)):
m = random.randint(1, 10)
p = random.randint(1, 10)
intervals.append(Interval(i - m, i + p))
iv = IntervalNode(intervals[0])
for f in intervals[1:]:
iv = iv.insert(f)
self.intervals = intervals
self.tree = iv
def getWindowTree(selectRes):
if len(selectRes[0]) == 3:
start, end, score = selectRes[0]
tree = IntervalNode(start, end, other=score)
# build an interval tree from the rest of the data
for start, end, score in selectRes[1:]:
tree = tree.insert(start, end, other=score)
else:
start, end = selectRes[0]
tree = IntervalNode(start, end, other=(end - start + 1))
# build an interval tree from the rest of the data
for start, end in selectRes[1:]:
# use size as the 3rd column
tree = tree.insert(start, end, other=(end - start + 1))
return tree
def __init__(self, fileName):
self.N = 0
self.headline = []
self.tree = {}
if not os.path.exists(fileName):
return
fh = open(fileName, 'rt')
a = -1
for lines in fh:
line = lines.strip()
if line[0] in ['#', '@']:
self.headline.append(line)
continue
IL1 = line.split('\t')
self.N = self.N + 1
chrom = IL1[0]
pstart = IL1[1]
pend = IL1[2]
p1 = int(pstart)
p2 = int(pend)
#chrom=chrom.replace("MT","M")
#self.IL.append([chrom,p1,p2])
start, end = p1 - 1, p2 + 1
if not (chrom == a):
if not (a == -1):
self.tree[a] = tree1
if (verbose):
print('scan %s line %d chrom %s' %
(fileName, self.N, a))
a = chrom
tree1 = IntervalNode(start, end)
else: # build an interval tree from the rest of the data
tree1 = tree1.insert(start, end)
self.tree[a] = tree1
fh.close()
print "Loaded IntervalList:\t" + fileName
print "events:\t" + str(self.N)
def get_schedule(chw_username, override_date = None):
#print "doing schedule lookup for %s" % (chw_username)
#if cached_schedules.has_key(chw_username):
#return cached_schedules[chw_username]
if override_date == None:
nowdate = datetime.now()
else:
nowdate = override_date
db = PactPatient.get_db()
chw_schedules = db.view('pactcarehq/chw_dot_schedule_condensed', key=chw_username).all()
day_intervaltree = {}
for item in chw_schedules:
single_sched = item['value']
day_of_week = int(single_sched['day_of_week'])
if day_intervaltree.has_key(day_of_week):
daytree = day_intervaltree[day_of_week]
else:
#if there's no day of week indication for this, then it's just a null interval node. to ensure that it's not checked, we make it REALLY old.
daytree = IntervalNode(get_seconds(datetime.min), get_seconds(nowdate + timedelta(days=10)))
if single_sched['ended_date'] == None:
enddate = nowdate+timedelta(days=9)
else:
enddate = datetime.strptime(single_sched['ended_date'], "%Y-%m-%dT%H:%M:%SZ")
#enddate = single_sched['ended_date']
startdate = datetime.strptime(single_sched['active_date'], "%Y-%m-%dT%H:%M:%SZ")
#startdate = single_sched['active_date']
pact_id = single_sched['pact_id']
daytree.insert(get_seconds(startdate), get_seconds(enddate), other=pact_id)
day_intervaltree[day_of_week] = daytree
#cached_schedules[chw_username] = CHWPatientSchedule(chw_username, day_intervaltree, chw_schedules)
#return cached_schedules[chw_username]
return CHWPatientSchedule(chw_username, day_intervaltree, chw_schedules)
def getWindowTree(selectRes):
if len(selectRes[0]) == 3:
start, end, score = selectRes[0]
tree = IntervalNode(start, end, other=score)
# build an interval tree from the rest of the data
for start, end, score in selectRes[1:]:
tree = tree.insert(start, end, other=score)
else:
start, end = selectRes[0]
tree = IntervalNode(start, end, other=(end - start + 1))
# build an interval tree from the rest of the data
for start, end in selectRes[1:]:
# use size as the 3rd column
tree = tree.insert(start, end, other=(end - start + 1))
return tree
f_file = dirOutPerCategory + prefix + "." + chr + ".genomicFeature"
outfile = open(f_file, 'w')
outFile[chr] = open(f_file, 'w')
outFile[chr].write(
'readName,chr,category, geneID, geneName, flag_multiMapped\n')
#DATA STRUCTURE - per chr
tree_utr3 = {}
tree_utr5 = {}
tree_cds = {}
tree_geneCoordinates = {}
tree_rRNA = {}
tree_intergenic = {} # +10,000
for chr in chr_list:
tree_utr3[chr] = IntervalNode(0, 0)
tree_utr5[chr] = IntervalNode(0, 0)
tree_cds[chr] = IntervalNode(0, 0)
tree_geneCoordinates[chr] = IntervalNode(0, 0)
tree_rRNA[chr] = IntervalNode(0, 0)
tree_intergenic[chr] = IntervalNode(0, 0)
print("Load gene annotations ...")
geneUTR3 = {}
#UTR3
print("Load", utr3_file)
with open(utr3_file, 'r') as f:
reader = csv.reader(f)
"Generates random interval over a size and span"
lo = randint(10000, SIZE)
hi = lo + randint(1, SPAN)
return (lo, hi)
def find(start, end, tree):
"Returns a list with the overlapping intervals"
out = []
tree.intersect( start, end, lambda x: out.append(x) )
return [ (x.start, x.end) for x in out ]
# use this to force both examples to generate the same data
seed(10)
# generate 10 thousand random intervals
data = map(generate, xrange(N))
# generate the intervals to query over
query = map(generate, xrange(10))
# start the root at the first element
start, end = data[0]
tree = IntervalNode( start, end )
# build an interval tree from the rest of the data
for start, end in data[1:]:
tree = tree.insert( start, end )
for start, end in query:
overlap = find(start, end , tree)
print '(%s, %s) -> %s' % (start, end, overlap)
'''Returns a list with the overlapping intervals'''
out = []
tree.contains( start, end, lambda x: out.append(x) )
return [ (x.start, x.end, x.other) for x in out ]
#Salon 1
sch = {1:[(930,1030),(1130,1230),(1345,1600)],2:[800,900]}
s1 = Classroom("S1", 50,sch)
#Salon 2
sch = {1:[(730,1000),(1350,1500),(1600,1800)],2:[1000,1500]}
s2 = Classroom("S2", 30,sch)
weekTrees = {}
mondayTree = IntervalNode( s1.schedule[1][0][0],s1.schedule[1][0][1],other=s1)
for interval in s1.schedule[1][1:]:
#print interval
mondayTree = mondayTree.insert( interval[0],interval[1],other=s1 )
for interval in s2.schedule[1]:
#print interval
mondayTree = mondayTree.insert( interval[0],interval[1],other=s2 )
tuesdayTree = deepcopy(mondayTree)
ilist = IntervalList(5)
#interval list
for interval in s1.schedule[1]:
ilist.insert(interval[0],interval[1],s1,1)
def build_interval_tree(intervals):
# root = IntervalNode(intervals[0].start, intervals[0].end,
# other=intervals[0])
root = IntervalNode(intervals[0])
return reduce(lambda tree, x: tree.insert(x), intervals[1:], root)
hi = lo + randint(1, SPAN)
return (lo, hi)
def find(start, end, tree):
"Returns a list with the overlapping intervals"
out = []
tree.intersect(start, end, lambda x: out.append(x))
return [(x.start, x.end) for x in out]
# use this to force both examples to generate the same data
seed(10)
# generate 10 thousand random intervals
data = map(generate, xrange(N))
# generate the intervals to query over
query = map(generate, xrange(10))
# start the root at the first element
start, end = data[0]
tree = IntervalNode(start, end)
# build an interval tree from the rest of the data
for start, end in data[1:]:
tree = tree.insert(start, end)
for start, end in query:
overlap = find(start, end, tree)
print '(%s, %s) -> %s' % (start, end, overlap)