"""Provide a examples that can be loaded mostly for the testing code """ import cPickle import math import os from compClust.mlx import datasets from compClust.mlx import labelings from compClust.mlx import views from compClust.config import config def LoadCho(): # Either grab the environment variable CHO_DATA (which is set in # setup.py or hope that the relative path is ok) dataroot = config.cho_data_dir # load data ds = datasets.Dataset(os.path.join(dataroot,"ChoCycling.dat")) # load some labelings cho = labelings.GlobalWrapper(ds, "cho clusters") cho.labelRows(os.path.join(dataroot, "ChoClassification.rlab")) common = labelings.GlobalWrapper(ds,"common name") common.labelRows(os.path.join(dataroot, "CommonNames.rlab")) orfs = labelings.GlobalWrapper(ds, "orfs") orfs.labelRows(os.path.join(dataroot, "ORFs.rlab")) em = labelings.GlobalWrapper(ds, "diagem clusters", ) em.labelRows(os.path.join(dataroot, "EM.rlab")) times = labelings.GlobalLabeling(ds,"time points") times.labelCols(ds, os.path.join(dataroot, "times.clab")) ds.setPrimaryRowLabeling(orfs) return ds def LoadChoSource(): """This returns a DataSource instead of a dataset. CompClustWeb needed additional code that the MLX didn't need and so I created a wrapper class. Hopefully at somepoint the two classes will be merged to make it easier to browse MLX datasets from within the Web GUI. """ # find where our datafile is stored dataroot = config.cho_data_dir datafile = os.path.join(dataroot, 'ChoCycling.dat') # make our datasource object from compClust.gui.DataSource import DataSource ds = DataSource(datafile) # specify where all of the labelings we want to attach are located # (and what kinds of labelings they are) labels = [ ('cho', os.path.join(dataroot,'ChoClassification.rlab'), True, False, None), ('em', os.path.join(dataroot, 'EM.rlab'), True, False, None), ('common', os.path.join(dataroot, 'CommonNames.rlab'), True, True, None), ('orfs', os.path.join(dataroot, 'ORFs.rlab'), True, True, "http://somesite.com/orf=%s"), ('times', os.path.join(dataroot, 'times.clab'), False, True, None)] # attach everything for name, path, isrow, isannotation, url in labels: ds.add_labeling(name, path, isrow, isannotation, url) # specify the "primary" and "secondary" annotation labelings for a datasource. # the primary MUST BE unique and it MUST BE specified for many detail plots to work. # (The primary name is used to identify which vector should be shown.) ds.primary = 'orfs' ds.secondary = 'common' return ds def LoadPGC(pickle_name=None, rowFiltered=True, colFiltered=True): """ Load a copy of the PGC dataset """ if pickle_name is None: pickle_name = os.path.expanduser("~/compclust-pgc.pkl") if os.path.exists(pickle_name): data = cPickle.load(open(pickle_name,'r')) else: base_url = "http://woldlab.caltech.edu/compclust/examples/pgc_diabetes" data = datasets.Dataset(base_url+"/compclust-reannotate_select_cal.log2.dat.gz") labeling_files = ["compclust-labeling-Age.clab.gz", "compclust-labeling-BMI.clab.gz", "compclust-labeling-Cap_(mm2).clab.gz", "compclust-labeling-Cap__Type1_(mean_n).clab.gz", "compclust-labeling-Cap__Type2a_(mean_n).clab.gz", "compclust-labeling-Cap__Type2b_(mean_n).clab.gz", "compclust-labeling-Cap__Type2c_(mean_n).clab.gz", "compclust-labeling-Cap_fiber_(mean_n).clab.gz", "compclust-labeling-Centroid_(Using_34_OXPHOS-CR_Genes).clab.gz", "compclust-labeling-Chol.clab.gz", "compclust-labeling-Glucose_0_capillary_blood_who_85.clab.gz", "compclust-labeling-Glucose_120_capillary_blood_who_85.clab.gz", "compclust-labeling-Glycogen_(mmol_kg).clab.gz", "compclust-labeling-Insulin_0.clab.gz", "compclust-labeling-Insulin_120.clab.gz", "compclust-labeling-M_value.clab.gz", "compclust-labeling-Patient_No.clab.gz", "compclust-labeling-Samplename_at_WICGR.clab.gz", "compclust-labeling-Total_Area_(um2).clab.gz", "compclust-labeling-Total_Cap_(n).clab.gz", "compclust-labeling-Total_fibers_(n).clab.gz", "compclust-labeling-Trigs.clab.gz", "compclust-labeling-Type1_(Percent).clab.gz", "compclust-labeling-Type1_(n).clab.gz", "compclust-labeling-Type1_Area_(um2).clab.gz", "compclust-labeling-Type1_Max__Area_(um2).clab.gz", "compclust-labeling-Type1_Mean_Area_(um2).clab.gz", "compclust-labeling-Type1_Min__Area_(um2).clab.gz", "compclust-labeling-Type1__Area__(Percent).clab.gz", "compclust-labeling-Type2a_(Percent).clab.gz", "compclust-labeling-Type2a_(n).clab.gz", "compclust-labeling-Type2a_Area_(Percent).clab.gz", "compclust-labeling-Type2a_Area_(um2).clab.gz", "compclust-labeling-Type2a_Max__Area_(um2).clab.gz", "compclust-labeling-Type2a_Mean_Area_(um2).clab.gz", "compclust-labeling-Type2a_Min__Area_(um2).clab.gz", "compclust-labeling-Type2b_(Percent).clab.gz", "compclust-labeling-Type2b_(n).clab.gz", "compclust-labeling-Type2b_Area_(Percent).clab.gz", "compclust-labeling-Type2b_Area_(um2).clab.gz", "compclust-labeling-Type2b_Max__area_(um2).clab.gz", "compclust-labeling-Type2b_Mean_Area_(um2).clab.gz", "compclust-labeling-Type2b_Min__area_(um2).clab.gz", "compclust-labeling-Type2c_(Percent).clab.gz", "compclust-labeling-Type2c_(n).clab.gz", "compclust-labeling-Type2c_Area_(Percent).clab.gz", "compclust-labeling-Type2c_Area_(um2).clab.gz", "compclust-labeling-Type2c_Max__area_(um2).clab.gz", "compclust-labeling-Type2c_Mean_area_(um2).clab.gz", "compclust-labeling-Type2c_Min__area_(um2).clab.gz", "compclust-labeling-UQCRB_(209065_at).clab.gz", "compclust-labeling-VO2_max_(ml_kg_min_total_body_weight).clab.gz", "compclust-labeling-WHR.clab.gz", "compclust-labeling-status.clab.gz",] for l in labeling_files: name = l[len("compclust-labeling-"):-len('.clab.gz')] new_labeling = labelings.GlobalWrapper(data,name) new_labeling.labelCols(base_url+"/"+l) labeling_files = ["compclust-labeling-Name.rlab.gz", "compclust-labeling-uid.rlab.gz"] for l in labeling_files: name = l[len("compclust-labeling-"):-len('.rlab.gz')] new_labeling = labelings.GlobalWrapper(data,name) new_labeling.labelRows(base_url+"/"+l) cPickle.dump(data, open(pickle_name,'w')) if rowFiltered: log2_100 = math.log(100,2) row_indices = [ row_index for row_index in xrange(data.getNumRows()) if [ value for value in data.getRowData(row_index) if value > log2_100 ]] data = views.RowSubsetView(data, row_indices) if colFiltered: status = data.getLabeling('status') diabetic_vs_healthy_cols = status.getColsByLabel('NGT') + status.getColsByLabel('DM2') data = views.ColumnSubsetView(data, diabetic_vs_healthy_cols) return data def LoadGNF(pickle_name=None): """ Load GNF dataset pickle_name specifies where you want the cached copy of the dataset to live """ if pickle_name is None: pickle_name = os.path.expanduser("~/compclust-gnf.pkl") if os.path.exists(pickle_name): return cPickle.load(open(pickle_name,'r')) base_url = "http://woldlab.caltech.edu/compclust/examples/gnf_human" data = datasets.Dataset(base_url + "/U133A+GNF1B_101402.AD.filtered.log2.data.tab.gz") tissue = labelings.GlobalWrapper(data, 'Tissue') tissue.labelCols(base_url + '/U133A+GNF1B_101402.AD.column_labeling.tab.gz') probe_id = labelings.GlobalWrapper(data, 'ProbeId') probe_id.labelRows(base_url + '/U133A+GNF1B_101402.AD.row_labeling.tab.gz') data.setPrimaryRowLabeling(probe_id) labeling_files = ["U133A+GNF1B_101402.AD.row_labeling-Aliases.tab.gz", "U133A+GNF1B_101402.AD.row_labeling-Description.tab.gz", "U133A+GNF1B_101402.AD.row_labeling-Ensembl.tab.gz", "U133A+GNF1B_101402.AD.row_labeling-Function.tab.gz", "U133A+GNF1B_101402.AD.row_labeling-Genome_Location.tab.gz", "U133A+GNF1B_101402.AD.row_labeling-LocusLink.tab.gz", "U133A+GNF1B_101402.AD.row_labeling-Name.tab.gz", "U133A+GNF1B_101402.AD.row_labeling-Protein_Families.tab.gz", "U133A+GNF1B_101402.AD.row_labeling-RefSeq.tab.gz", "U133A+GNF1B_101402.AD.row_labeling-Reporters.tab.gz", "U133A+GNF1B_101402.AD.row_labeling-Taxon.tab.gz", "U133A+GNF1B_101402.AD.row_labeling-UniGene.tab.gz", "U133A+GNF1B_101402.AD.row_labeling-UniProt.tab.gz",] for l in labeling_files: name = l[len("U133A+GNF1B_101402.AD.row_labeling-"):-len('.tab.gz')] new_labeling = labelings.GlobalWrapper(data,name) new_labeling.labelRows(base_url+"/"+l) cPickle.dump(data, open(pickle_name,'w')) return data