diff options
author | BonfaceKilz | 2021-04-30 12:16:51 +0300 |
---|---|---|
committer | BonfaceKilz | 2021-04-30 13:45:15 +0300 |
commit | c7e661b8ff9f70955418fbc4527378904beb0cf4 (patch) | |
tree | 7a164b42d46e15b6f2775a50137b412c8713e2f3 /wqflask/base/data_set.py | |
parent | 385da724b63f57d0fb1bbe3476cea31ef837c081 (diff) | |
download | genenetwork2-c7e661b8ff9f70955418fbc4527378904beb0cf4.tar.gz |
autopep8: Fix E20-E27
Run:
python -m autopep8 --in-place --recrusive ./ --select\
E20,E211,E22,E224,E224,E225,E226,E227,E228,E231,E241,\
E242,E251,E252,E26,E265,E266,E27 -p 3
Diffstat (limited to 'wqflask/base/data_set.py')
-rw-r--r-- | wqflask/base/data_set.py | 18 |
1 files changed, 9 insertions, 9 deletions
diff --git a/wqflask/base/data_set.py b/wqflask/base/data_set.py index 3183363b..55ab45f5 100644 --- a/wqflask/base/data_set.py +++ b/wqflask/base/data_set.py @@ -150,7 +150,7 @@ class DatasetType: "FROM PublishFreeze, InbredSet " "WHERE InbredSet.Name = '%s' AND " "PublishFreeze.InbredSetId = InbredSet.Id"), - 'geno': ("SELECT GenoFreeze.Id FROM GenoFreeze WHERE " + 'geno': ("SELECT GenoFreeze.Id FROM GenoFreeze WHERE " "GenoFreeze.Name = \"%s\" ") } @@ -215,7 +215,7 @@ def create_datasets_list(): if USE_REDIS: r.set(key, pickle.dumps(datasets, pickle.HIGHEST_PROTOCOL)) - r.expire(key, 60*60) + r.expire(key, 60 * 60) return datasets @@ -239,7 +239,7 @@ class Markers: for line in bimbam_fh: marker = {} marker['name'] = line.split(delimiter)[0].rstrip() - marker['Mb'] = float(line.split(delimiter)[1].rstrip())/1000000 + marker['Mb'] = float(line.split(delimiter)[1].rstrip()) / 1000000 marker['chr'] = line.split(delimiter)[2].rstrip() markers.append(marker) @@ -369,8 +369,8 @@ class DatasetGroup: def get_markers(self): def check_plink_gemma(): if flat_file_exists("mapping"): - MAPPING_PATH = flat_files("mapping")+"/" - if os.path.isfile(MAPPING_PATH+self.name+".bed"): + MAPPING_PATH = flat_files("mapping") + "/" + if os.path.isfile(MAPPING_PATH + self.name + ".bed"): return True return False @@ -416,7 +416,7 @@ class DatasetGroup: else: logger.debug("Cache not hit") - genotype_fn = locate_ignore_error(self.name+".geno", 'genotype') + genotype_fn = locate_ignore_error(self.name + ".geno", 'genotype') if genotype_fn: self.samplelist = get_group_samplelists.get_samplelist( "geno", genotype_fn) @@ -425,7 +425,7 @@ class DatasetGroup: if USE_REDIS: r.set(key, json.dumps(self.samplelist)) - r.expire(key, 60*5) + r.expire(key, 60 * 5) def all_samples_ordered(self): result = [] @@ -531,7 +531,7 @@ def datasets(group_name, this_group=None): if USE_REDIS: r.set(key, pickle.dumps(dataset_menu, pickle.HIGHEST_PROTOCOL)) - r.expire(key, 60*5) + r.expire(key, 60 * 5) if this_group != None: this_group._datasets = dataset_menu @@ -622,7 +622,7 @@ class DataSet: WHERE ProbeSetFreeze.ProbeFreezeId = ProbeFreeze.Id AND ProbeFreeze.TissueId = Tissue.Id AND (ProbeSetFreeze.Name = '%s' OR ProbeSetFreeze.FullName = '%s' OR ProbeSetFreeze.ShortName = '%s') - """ % (query_args), "/dataset/"+self.name+".json", + """ % (query_args), "/dataset/" + self.name + ".json", lambda r: (r["id"], r["name"], r["full_name"], r["short_name"], r["data_scale"], r["tissue"]) ) |