diff options
-rwxr-xr-x | wqflask/base/data_set.py | 72 | ||||
-rw-r--r-- | wqflask/wqflask/correlation/show_corr_results.py | 5 |
2 files changed, 49 insertions, 28 deletions
diff --git a/wqflask/base/data_set.py b/wqflask/base/data_set.py index b2836480..edee6685 100755 --- a/wqflask/base/data_set.py +++ b/wqflask/base/data_set.py @@ -38,6 +38,7 @@ from base import species from dbFunction import webqtlDatabaseFunction from utility import webqtlUtil from utility.benchmark import Bench +from wqflask.my_pylmm.pyLMM import chunks from MySQLdb import escape_string as escape from pprint import pformat as pf @@ -68,6 +69,13 @@ def create_dataset(dataset_name): dataset_class = globals()[dataset_ob] return dataset_class(dataset_name) +def create_in_clause(items): + """Create an in clause for mysql""" + in_clause = ', '.join("'{}'".format(x) for x in mescape(*items)) + in_clause = '( {} )'.format(in_clause) + return in_clause + + def mescape(*items): """Multiple escape""" escaped = [escape(item) for item in items] @@ -626,31 +634,45 @@ class MrnaAssayDataSet(DataSet): def get_trait_data(self): self.samplelist = self.group.samplelist + self.group.parlist + self.group.f1list - sample_ids = [] - where_clause = "" - #for sample in self.samplelist: - # if len(where_clause): - # where_clause += " or " - # where_clause += "Strain.Name = '{}'".format(*mescape(sample)) - + #query_samplelist = ', '.join("'{}'".format(x) for x in mescape(*samplelist)) + #query_samplelist = '( ' + query_samplelist + ' )' + #query_samplelist = create_in(samplelist) + + print("self.samplelist is:", self.samplelist) + query = """ - SELECT Strain.Id, Strain.Name FROM Strain, Species - WHERE ({}) + SELECT Strain.Name, Strain.Id FROM Strain, Species + WHERE Strain.Name IN {} and Strain.SpeciesId=Species.Id and Species.name = '{}' - """.format(where_clause, *mescape(self.group.species)) - print("raspberry query: ", query) - result = g.db.execute(query).fetchall() + """.format(create_in_clause(self.samplelist), *mescape(self.group.species)) + results = dict(g.db.execute(query).fetchall()) + print("results are:", results) + print("type results are:", type(results)) + + #sample_ids = [] + #for item in self.samplelist: + # sample_ids.append(results[item]) + + sample_ids = [results[item] for item in self.samplelist] + print("sample_ids are:", sample_ids) + + #for sample in self.samplelist: + # pass - print("[blackberry] result is:", pf(result)) - #sample_ids.append('%d' % this_id) + #for index in range(len(results)): + # sample_ids.append(results[index][0]) # MySQL limits the number of tables that can be used in a join to 61, # so we break the sample ids into smaller chunks - chunk_count = 50 - n = len(sample_ids) / chunk_count - if len(sample_ids) % chunk_count: - n += 1 + # Postgres doesn't have that limit, so we can get rid of this after we transition + chunk_size = 50 + + number_chunks = int(math.ceil(len(sample_ids) / chunk_size)) + + trait_sample_data = [] + for sample_ids_step in chunks.divide_into_chunks(sample_ids, number_chunks): + #XZ, 09/24/2008: build one temporary table that only contains the records associated with the input GeneId #tempTable = None #if GeneId and db.type == "ProbeSet": @@ -664,12 +686,14 @@ class MrnaAssayDataSet(DataSet): # TissueProbeSetFreezeId=tissueProbeSetFreezeId, # method=method, # returnNumber=returnNumber) - trait_sample_data = [] - for step in range(int(n)): - temp = [] - sample_ids_step = sample_ids[step*chunk_count:min(len(sample_ids), (step+1)*chunk_count)] - for item in sample_ids_step: - temp.append('T%s.value' % item) + + #for step in range(int(n)): + #temp = [] + #sample_ids_step = sample_ids[step*chunk_size:min(len(sample_ids), (step+1)*chunk_size)] + #for item in sample_ids_step: + # temp.append('T%s.value' % item) + + temp = ['T%s.value' % item for item in sample_ids_step] query = "SELECT {}.Name,".format(escape(self.type)) data_start_pos = 1 query += string.join(temp, ', ') diff --git a/wqflask/wqflask/correlation/show_corr_results.py b/wqflask/wqflask/correlation/show_corr_results.py index ee732050..9b1843bd 100644 --- a/wqflask/wqflask/correlation/show_corr_results.py +++ b/wqflask/wqflask/correlation/show_corr_results.py @@ -312,9 +312,6 @@ class CorrelationResults(object): self.process_samples(start_vars, self.this_trait.data.keys(), primary_samples) self.target_dataset = data_set.create_dataset(start_vars['corr_dataset']) self.target_dataset.get_trait_data() - # Lei Yan todo - import pdb - pdb.set_trace() correlation_data = collections.defaultdict(list) for trait, values in self.target_dataset.trait_data.iteritems(): values_1 = [] @@ -327,7 +324,7 @@ class CorrelationResults(object): values_2.append(target_value) correlation = calCorrelation(values_1, values_2) correlation_data[trait] = correlation - print ('%s %s' % (trait, correlation)) + print ('correlation result: %s %s' % (trait, correlation)) #XZ, 09/18/2008: get all information about the user selected database. #target_db_name = fd.corr_dataset |