Skip to content

Commit

Permalink
Merge branch 'master' of https://github.com/isb-cgc/ISB-CGC-Common in…
Browse files Browse the repository at this point in the history 
…to isb-cgc-test
  • Loading branch information
@phyllers
phyllers committed Oct 31, 2016
2 parents 41fc8a8 + f901f3b commit e93e724
Show file tree
Hide file tree
Showing 2 changed files with 95 additions and 27 deletions.
104 changes: 77 additions & 27 deletions cohorts/views.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -467,10 +467,12 @@ def get_sample_participant_list(user, inc_filters=None, cohort_id=None):
# Query the resulting 'filter_table' (which might just be our original base_table) for the samples
# and participants

cursor.execute("SELECT DISTINCT %s FROM %s;" % ('SampleBarcode', filter_table,))
cursor.execute("SELECT DISTINCT ms.SampleBarcode, ps.id FROM %s ms JOIN (SELECT id,name FROM projects_study WHERE owner_id = 1) ps ON ps.name = ms.Study;" % (filter_table,))

for row in cursor.fetchall():
samples_and_participants['items'].append({'sample_barcode': row[0], 'study_id': None})
samples_and_participants['items'].append({'sample_barcode': row[0], 'study_id': row[1]})

# Fetch the study IDs for these samples

samples_and_participants['count'] = len(samples_and_participants['items'])

Expand Down Expand Up @@ -1678,40 +1680,81 @@ def set_operation(request):
cohorts = Cohort.objects.filter(id__in=cohort_ids, active=True, cohort_perms__in=request.user.cohort_perms_set.all())
request.user.cohort_perms_set.all()
if len(cohorts):
sample_study_map = {}

cohort_samples = Samples.objects.filter(cohort=cohorts[0])
cohort_samples_ids = set(Samples.objects.filter(cohort=cohorts[0]).values_list('sample_id',flat=True))
cohort_patients = set(Patients.objects.filter(cohort=cohorts[0]).values_list('patient_id', flat=True))
cohort_samples = set(Samples.objects.filter(cohort=cohorts[0]).values_list('sample_id', 'study_id'))

# Samples from older cohorts made from ISB-CGC data may have null for their study IDs; we should treat
# these as 'matching' studies

for sample in cohort_samples:
if sample.sample_id not in sample_study_map:
sample_study_map[sample.sample_id] = []
if sample.study is None:
if -1 not in sample_study_map[sample.sample_id]:
sample_study_map[sample.sample_id].append(-1);
elif sample.study.id not in sample_study_map[sample.sample_id]:
sample_study_map[sample.sample_id].append(sample.study.id)

notes = 'Intersection of ' + cohorts[0].name

# print "Start of intersection with %s has %d" % (cohorts[0].name, len(cohort_samples))
for i in range(1, len(cohorts)):
cohort = cohorts[i]
notes += ', ' + cohort.name

cohort_samples = Samples.objects.filter(cohort=cohort)

for sample in cohort_samples:
if sample.sample_id in sample_study_map:
if sample.study is None:
if -1 not in sample_study_map[sample.sample_id]:
sample_study_map[sample.sample_id].append(-1);
elif sample.study.id not in sample_study_map[sample.sample_id]:
sample_study_map[sample.sample_id].append(sample.study.id)

cohort_samples_ids = cohort_samples_ids.intersection(Samples.objects.filter(cohort=cohort).values_list('sample_id',flat=True))
cohort_patients = cohort_patients.intersection(Patients.objects.filter(cohort=cohort).values_list('patient_id', flat=True))
cohort_samples = cohort_samples.intersection(Samples.objects.filter(cohort=cohort).values_list('sample_id', 'study_id'))

# se1 = set(x[0] for x in s1)
# se2 = set(x[0] for x in s2)
# TODO: work this out with user data when activated
# cohort_samples = cohort_samples.extra(
# tables=[Samples._meta.db_table+"` AS `t"+str(1)], # TODO This is ugly :(
# where=[
# 't'+str(i)+'.sample_id = ' + Samples._meta.db_table + '.sample_id',
# 't'+str(i)+'.study_id = ' + Samples._meta.db_table + '.study_id',
# 't'+str(i)+'.cohort_id = ' + Samples._meta.db_table + '.cohort_id',
# ]
# )
# cohort_patients = cohort_patients.extra(
# tables=[Patients._meta.db_table+"` AS `t"+str(1)], # TODO This is ugly :(
# where=[
# 't'+str(i)+'.patient_id = ' + Patients._meta.db_table + '.patient_id',
# 't'+str(i)+'.cohort_id = ' + Patients._meta.db_table + '.cohort_id',
# ]
# )

cohort_sample_list = []

print >> sys.stdout, 'unique ID set: '+cohort_samples_ids.__str__()

for sample in cohort_samples_ids:
if len(sample_study_map[sample]) > 1:
print >> sys.stdout, "Studies: "+sample_study_map[sample].__str__()
studies = Study.objects.filter(id__in=sample_study_map[sample])
print >> sys.stdout, "Study objs: "+studies.__str__()
no_match = False
root = -1
max_depth = -1
deepest_study = -1
for study in studies:
study_rd = study.get_my_root_and_depth()

if root < 0:
root = study_rd['root']
max_depth = study_rd['depth']
deepest_study = study.id
else:
if root != study_rd['root']:
no_match = True
else:
if max_depth < 0 or study_rd['depth'] > max_depth:
max_depth = study_rd['depth']
deepest_study = study.id

if not no_match:
cohort_sample_list.append({'id':sample, 'study':deepest_study, })

else:
# If a study's ID is <= 0 it's a null study ID, so just record None
study = (None if sample_study_map[sample][0] <=0 else sample_study_map[sample][0])
cohort_sample_list.append({'id': sample, 'study':study})

patients = list(cohort_patients)
samples = list(cohort_samples)
samples = cohort_sample_list

elif op == 'complement':
base_id = request.POST.get('base-id')
Expand Down Expand Up @@ -1748,7 +1791,10 @@ def set_operation(request):
project_id = settings.BQ_PROJECT_ID
cohort_settings = settings.GET_BQ_COHORT_SETTINGS()
bcs = BigQueryCohortSupport(project_id, cohort_settings.dataset_id, cohort_settings.table_id)
bcs.add_cohort_with_sample_barcodes(new_cohort.id, samples)
if op == 'intersect':
bcs.add_cohort_with_sample_barcodes(new_cohort.id, [v['id'] for v in samples])
else:
bcs.add_cohort_with_sample_barcodes(new_cohort.id, samples)

# Store cohort to CloudSQL
patient_list = []
Expand All @@ -1757,8 +1803,12 @@ def set_operation(request):
Patients.objects.bulk_create(patient_list)

sample_list = []
print >> sys.stdout, "Samples in intersection: "+samples.__str__()
for sample in samples:
sample_list.append(Samples(cohort=new_cohort, sample_id=sample[0], study_id=sample[1]))
if op == 'intersect':
sample_list.append(Samples(cohort=new_cohort, sample_id=sample['id'], study_id=sample['study']))
else:
sample_list.append(Samples(cohort=new_cohort, sample_id=sample[0], study_id=sample[1]))
Samples.objects.bulk_create(sample_list)

# Create Sources
Expand Down
18 changes: 18 additions & 0 deletions projects/models.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -79,6 +79,7 @@ class Project_Last_View(models.Model):


class Study(models.Model):
id = models.AutoField(primary_key=True, null=False, blank=False)
name = models.CharField(max_length=255)
description = models.TextField(null=True, blank=True)
active = models.BooleanField(default=True)
Expand Down Expand Up @@ -114,6 +115,23 @@ def mark_viewed (self, request, user=None):

return last_view

'''
Get the root/parent study of this study's extension hierarchy, and its depth
'''
def get_my_root_and_depth(self):
root = self.id
depth = 1
ancestor = self.extends

while ancestor is not None:
ancStudy = Study.objects.filter(id=ancestor)
ancestor = ancStudy.extends
depth += 1
root = ancStudy.id

return {'root': root, 'depth': depth}


def get_status (self):
status = 'Complete'
for datatable in self.user_data_tables_set.all():
Expand Down

0 comments on commit e93e724

Please sign in to comment.