def test_ARC2016_16(with_small, overwrite):
import pydream2015
from os.path import split, join
pkgdir = split(pydream2015.__file__)[0]
input_dir = join(pkgdir, 'test_input')
output_dir = join(pkgdir, 'test_output')
pydream2015.initdatapath(input_dir, output_dir)
ppiscore = 950
alpha = 0.95
ppifile = pydream2015.MYDATA_PPI_STRING_TRANSLATED
gexfile = pydream2015.MYDATA_GENEEXPR_FILLED
signalfile = join(input_dir, 'team/mschoi/arc2016_16_table(a)cnv.csv')
resfile = 'smooth_cnv_str.csv'
run(gexfile, ppifile, signalfile, resfile, ppiscore, alpha)
signalfile = join(input_dir, 'team/mschoi/arc2016_16_table(b)methyl.csv')
resfile = 'smooth_methyl_str.csv'
run(gexfile, ppifile, signalfile, resfile, ppiscore, alpha)
def test_divide_train_leaderboard():
sys.path.append('..')
import pydream2015
indir = join(pathsplit(pydream2015.__file__)[0], 'test_input')
outdir = join(pathsplit(pydream2015.__file__)[0], 'test_output')
indir = os.path.abspath(indir)
outdir = os.path.abspath(outdir)
pydream2015.initdatapath(indir, outdir)
trainids, testids = pydream2015.util.divide_train_leaderboard()
def test_calc_dd3d():
sys.path.append('../..')
import pydream2015
from pydream2015 import feature
pydream2015.initdatapath('../test_input', '../test_output')
in_dd3d = pydream2015.DATA_DRUG_DESC_3D
out_dd3d_conv = pydream2015.MYDATA_DRUG_DESC_3D_CONV
if not exists(out_dd3d_conv):
feature.calc_dd3d(in_dd3d, out_dd3d_conv)
else:
print '>>', out_dd3d_conv, '- skipped'
def test_fill_missinggex():
sys.path.append('../..')
import pydream2015
pydream2015.initdatapath('../test_input', '../test_output')
in_therapy = pydream2015.DATA_COMBITHERAPY
in_gex = pydream2015.DATA_GENEEXPR
in_cellinfo = pydream2015.DATA_CELLINFO
out_gexfilled = pydream2015.MYDATA_GENEEXPR_FILLED
fill_missinggex(in_therapy, in_gex, in_cellinfo, out_gexfilled)
assert exists(out_gexfilled)
def test_divide_combi():
import pydream2015
indir = join(pathsplit(pydream2015.__file__)[0], 'test_input')
outdir = join(pathsplit(pydream2015.__file__)[0], 'test_output')
indir = os.path.abspath(indir)
outdir = os.path.abspath(outdir)
pydream2015.initdatapath(indir, outdir)
therapy_traindata = pd.read_csv(pydream2015.DATA_COMBITHERAPY)
trainids, testids = pydream2015.util.divide_combi(therapy_traindata,
ratio=0.3)
print(len(trainids))
print(len(testids))
pass
def test_make_biomartsdict():
sys.path.append('../..')
import pydream2015
pydream2015.initdatapath('../test_input', '../test_output')
in_ensemblid = pydream2015.DATA_ENSEMBLID
out_dict = pydream2015.MYDATA_DICT
print('>>', out_dict)
if not exists(out_dict):
make_biomartsdict(in_ensemblid, out_dict, test=False)
else:
print('- skipped')
assert exists(out_dict)
def test_translate_STRING():
sys.path.append('../..')
import pydream2015
pydream2015.initdatapath('../test_input', '../test_output')
in_ppi = pydream2015.DATA_PPI_STRING
in_dict = pydream2015.MYDATA_DICT
out_processed = pydream2015.MYDATA_PPI_STRING_PROCESSED
out_translated = pydream2015.MYDATA_PPI_STRING_TRANSLATED
out_hist = pydream2015.MYDATA_PPI_STRING_HIST
translate_STRING(in_ppi,
in_dict,
out_processed,
out_translated,
out_hist,
test=False)
assert exists(out_processed)
assert exists(out_translated)
assert exists(out_hist)
# This file is part of Pydream2015. #************************************************************************* import os, sys, cPickle as pickle, re, json, pandas as pd, numpy as np, pytest import pydream2015 from sets import Set from multiprocessing import Pool from os.path import exists, split as pathsplit, join from pdb import set_trace from os.path import split,join indir = join(split(pydream2015.__file__)[0], 'test_input') outdir = join(split(pydream2015.__file__)[0], 'test_output') pydream2015.initdatapath(indir, outdir) therapy = pydream2015.DATA_COMBITHERAPY therapy_test = pydream2015.DATA_COMBITHERAPY_TEST therapy_leader = pydream2015.DATA_COMBITHERAPY_LEADER gexdata = pydream2015.MYDATA_GENEEXPR_FILLED mutdata = pydream2015.DATA_MUTATION dd3dc = pydream2015.MYDATA_DRUG_DESC_3D_CONV smoothed_CNV_STRING = pydream2015.MYDATA_SMOOTHED_CNV_STRING
def test_reshape_therapy(with_small,
overwrite,
outputfile='therapy_expanded.csv'):
overwrite = True
if exists(outputfile) and (overwrite == False):
return
import sys
sys.path.append('..')
import pydream2015
from pydream2015.util import update_progress
indir = join(pathsplit(pydream2015.__file__)[0], 'test_input')
outdir = join(pathsplit(pydream2015.__file__)[0], 'test_output')
indir = os.path.abspath(indir)
outdir = os.path.abspath(outdir)
pydream2015.initdatapath(indir, outdir)
output_df = pd.DataFrame([], columns=['CELL_LINE', 'COMBINATION_ID', \
'COMPOUND_A', 'COMPOUND_B', 'DOSE_A', 'DOSE_B', 'UNIT_A', \
'UNIT_B', 'RESPONSE', 'METHOD', 'FILE_SOURCE'])
df1 = pd.read_csv(pydream2015.DATA_COMBITHERAPY)
# 디렉토리 내용가져오기
raw_data_dir = pydream2015.DIR_TRAINING_COMBI_THERAPY
files = glob.glob(join(raw_data_dir, '*.csv'))
idx = 0
for kk, thisfile in enumerate(files):
update_progress(kk, len(files))
filedata = pd.read_csv(thisfile, index_col='Unnamed: 0')
dim_agent_1 = filedata.index.values.tolist().index('(=Agent 1)')
dim_agent_2 = filedata.columns.values.tolist().index('(=Agent 2)')
agent2_doses = filedata.columns.values.tolist()[0:dim_agent_2]
agent1_doses = filedata.index.values.tolist()[0:dim_agent_1]
lbl_agent_1 = filedata.loc['Agent1', agent2_doses[0]]
lbl_agent_2 = filedata.loc['Agent2', agent2_doses[0]]
unit_1 = filedata.loc['Unit1', agent2_doses[0]]
unit_2 = filedata.loc['Unit2', agent2_doses[0]]
title = filedata.loc['Title', agent2_doses[0]]
for agent1 in agent1_doses:
for agent2 in agent2_doses:
response = filedata.loc[agent1, agent2]
output_df.loc[idx, 'CELL_LINE'] = title
output_df.loc[
idx, 'COMBINATION_ID'] = lbl_agent_1 + '.' + lbl_agent_2
output_df.loc[idx, 'COMPOUND_A'] = lbl_agent_1
output_df.loc[idx, 'COMPOUND_B'] = lbl_agent_2
output_df.loc[idx, 'DOSE_A'] = agent1
output_df.loc[idx, 'DOSE_B'] = agent2
output_df.loc[idx, 'UNIT_A'] = unit_1
output_df.loc[idx, 'UNIT_B'] = unit_2
output_df.loc[idx, 'RESPONSE'] = response
if (agent1 == "0") and (agent2 == "0"):
output_df.loc[idx, 'METHOD'] = 'CONTROL'
elif (agent1 == "0") or (agent2 == "0"):
output_df.loc[idx, 'METHOD'] = 'MONO'
else:
output_df.loc[idx, 'METHOD'] = 'COMBI'
output_df.loc[idx, 'FILE_SOURCE'] = os.path.split(thisfile)[1]
idx += 1
output_df.to_csv('therapy_expanded.csv', index=False)