def __init__(self, study, z):
'''selects study and creates MicrobPLSA instance for that study'''
self.study = study
self.z = z
self.m = microbplsa.MicrobPLSA()
self.plsa = self.m.open_model(study=self.study, z=self.z)
return None
def __init__(self, study = None, biomFile = None):
''' Opens the probs of a model previously computed and saved in a json file '''
m = microbplsa.MicrobPLSA()
m.open_data(study = study, dataFile = biomFile) #get data matrix from the results file
self.rawData = m.datamatrix
self.N,self.S = self.rawData.shape
return None
def convert_2_R(results_file):
'''converts the result matrix probabilities into tab delimited files readable by R'''
dir = os.path.dirname(os.path.realpath(__file__)) + "/Results/"
m = microbplsa.MicrobPLSA()
plsa = m.open_model(modelFile=os.path.oin(
dir, results_file)) #get model from the results file
p_z = plsa.p_z
p_w_z = plsa.p_w_z
#return document's distribution
p_z_d = plsa.document_topics()
R_file_p_z = open(dir + 'Results_for_R/p_z_' + results_file, 'w')
for row in p_z:
R_file_p_z.writelines('\n')
R_file_p_z.writelines('\t' + str(row))
R_file_p_z.writelines('\n')
R_file_p_z.close()
R_file_p_w_z = open(dir + 'Results_for_R/p_w_z_' + results_file, 'w')
for row in p_w_z:
R_file_p_w_z.writelines('\n')
R_file_p_w_z.writelines('\t' + str(i) for i in row)
R_file_p_w_z.writelines('\n')
R_file_p_w_z.close()
R_file_p_z_d = open(dir + 'Results_for_R/p_z_d_' + results_file, 'w')
#R_file_p_z_d.writelines('\t topic' + (str(i)) for i in range(1,p_z_d.shape[1]+1))
for row in p_z_d:
R_file_p_z_d.writelines('\n')
R_file_p_z_d.writelines('\t' + str(i) for i in row)
R_file_p_z_d.writelines('\n')
R_file_p_z_d.close()
return None
def train(k,
kFolds,
data,
study,
name,
z,
numRuns=1,
seed='None',
override=False,
useC=True,
folder=FOLDER):
i = 1
for trainSamples, testSamples in kFolds:
trainData, testData = data[:, trainSamples], data[:, testSamples]
print '\nTraining dataset fold {0} of {1}'.format(i, k)
m = microbplsa.MicrobPLSA()
m.study = study
m.name = name
m.datamatrix = trainData
add = '_cross_seed' + str(seed) + '_k' + str(k) + '(' + str(i) + ')'
m.generate_runs(z_i=z,
z_f=z,
numRuns=numRuns,
useC=useC,
override=override,
folder=folder,
add_to_file=add)
i += 1
return None
def main(*argv):
'''handles user input and runs plsa'''
parser = argparse.ArgumentParser(description='This scripts runs plsa for a range of topic numbers.')
parser.add_argument('-s','--study', help='The study number', default = None)
parser.add_argument('-n','--name', help='The name of the dataset')
parser.add_argument('-z','--topics', help='The range of topics to be run [z_start, z_end]', nargs = '+', type = int, required = True)
parser.add_argument('-z_inc','--increment', help='Increment of topic numbers', type = int, default = 1)
parser.add_argument('-useC', help='use C code to run plsa', action = 'store_true')
parser.add_argument('-runs','-numruns', help='Specify the number of runs', type = int, default = 1)
parser.add_argument('-override', help='Overrides a result file', action = 'store_true')
parser.add_argument('-add', help='Add text to result file', default = '')
args = parser.parse_args()
if args.study is None and args.name is None:
print "***Study number or a data name must be specified.***\n"
parser.print_help()
sys.exit()
if args.study:
study = str(args.study)
else: study = None
name = args.name
if len(args.topics) == 1:
z_i = args.topics[0]
z_f = args.topics[0]
elif len(args.topics) == 2:
z_i = min(args.topics)
if z_i < 2:
z_i = 2
z_f = max(args.topics)
else:
print "\n***Too many arguments specified for number of topics***\n"
parser.print_help()
sys.exit()
z_inc = args.increment
numRuns = args.runs
useC = args.useC
override = args.override
add = args.add
print (" Study: %s" % study)
print (" Name: %s" % name)
if z_i != z_f:
print (" Topics will be run from {0} to {1} in increments of {2}".format(z_i, z_f, z_inc))
else:
print (" PLSA will be run with {0} topics".format(z_i))
print (" Using C: %s" % args.useC)
print (" Number of runs: %s" % args.runs)
print (" Override result files: %s" % args.override)
print (" Text to add to file: %s" % args.add)
m = microbplsa.MicrobPLSA()
m.open_data(study = study, name = name)
m.generate_runs(z_i = z_i, z_f = z_f, z_inc = z_inc, numRuns = numRuns, useC = useC, override = override, add_to_file = add)
def makedendrogram(study=None, filename=None, showme=True):
m = microbplsa.MicrobPLSA()
m.open_data(study=study, dataFile=filename)
data = m.datamatrix
X = data.T
Y = pdist(X, 'euclidean')
Z = linkage(Y)
t = 0.7 * max(Z[:, 2])
D = dendrogram(Z, color_threshold=t)
leaves_order = D['ivl']
if showme:
show()
else:
plt.clf()
return leaves_order
sys.path.insert(0, _root_dir)
import microbplsa
analysis_dir = _root_dir + '/Analysis'
sys.path.insert(0, analysis_dir)
from labelling import Labelling
from string import replace
FOLDER = 'Models'
name = 'bac_final0.03.otutable_GOODSAMPLES'
z = 20
run = 1
pcoordfile = os.path.join(_root_dir, 'D3', 'pcplots',
'paracoords_LTSP_topics' + '.js')
m = microbplsa.MicrobPLSA(name=name)
m.open_model(z=z, run=run, useC=True,
folder=FOLDER) #get model from the results file
plsa = m.model
p_z_d = plsa.document_topics() #return document's distribution
Z, N = p_z_d.shape #number of samples
print Z, N
samples = ['Sample' + str(i) for i in range(1, N + 1)]
labels = ['Topic ' + str(i) for i in range(1, Z + 1)]
f = open(pcoordfile, 'w')
f.write('var topics = [\n')
for s, distribution in enumerate(p_z_d.T):
import microbplsa
analysis_dir = _root_dir + '/Analysis'
sys.path.insert(0, analysis_dir)
from labelling import Labelling
from string import replace
study = '1526'
z = 8
CORRELATION_THRESHOLD = 0.0
pcoordfile = _root_dir + '/D3/pcplots/topics.js'
f = '/Users/sperez/git/microbPLSA/MicrobProcessor/Results/study_' + study + '_' + str(
z) + '_topics_.txt'
datafile = '/Users/sperez/Documents/PLSAfun/EMPL data/study_' + study + '_split_library_seqs_and_mapping/study_' + study + '_closed_reference_otu_table.biom'
m = microbplsa.MicrobPLSA()
plsa = m.open_model(f) #get model from the results file
p_z_d = plsa.document_topics() #return document's distribution
Z, N = p_z_d.shape #number of samples
Lab = Labelling(study, Z, ignore_continuous=False,
adjusted_metadata=True) #get labels!
x, y, z = Lab.metadata(non_labels=[])
print y
R = Lab.correlate()
labels_r = Lab.assignlabels(R, num_labels=1)
print labels_r
oldlabels, r = zip(*labels_r)
goodlabels = []
for lab, r in labels_r:
if r > CORRELATION_THRESHOLD or r < -CORRELATION_THRESHOLD:
def makePCA(datafile, num_components):
m = microbplsa.MicrobPLSA()
m.open_data(datafile=dataFile) #get data of OTU abundances per sample
X = m.datamatrix.T
plsa = m.open_model(modelFile=resultfile) #get model from the results file
#return document's distribution
p_d_z = plsa.p_d_z
N, Z = p_d_z.shape
#get topic labels
if MANUAL_LABELS:
labels = MANUAL_LABELS
else:
Lab = Labelling(study, Z, ignore_continuous=False)
Lab.metadata(non_labels=['BarcodeSequence'])
R = Lab.correlate()
labels_r = Lab.assignlabels(R, num_labels=1)
labels, r = zip(*labels_r)
labels = [l.replace('(', '\n(') for l in labels]
#get primary topic per sample
topics = []
for i, row in enumerate(p_d_z):
max_topic_index = np.argmax(row)
topics.append(max_topic_index)
topics = np.array(topics)
pca = PCA(n_components=num_components, whiten=True)
pca.fit(X)
X_r = pca.fit(X).transform(X)
# Percentage of variance explained for each components
print('Explained variance ratio (first two components): %s' %
str(pca.explained_variance_ratio_))
#initiate plot and colors
colors = [float(c) / float(Z) for c in range(0, Z)]
colors = plt.cm.rainbow(np.linspace(0, 1, Z))
fig = plt.figure(1, figsize=(4, 3))
plt.clf()
ax = plt.subplot(111, projection='3d')
if num_components == 2:
for c, i, l in zip(colors, range(0, Z), labels):
ax.plot(X_r[topics == i, 0],
X_r[topics == i, 1],
'o',
color=c,
label=l)
box = ax.get_position()
ax.set_position([box.x0, box.y0, 0.5, box.height])
elif num_components == 3:
for c, i, l in zip(colors, range(0, Z), labels):
ax.plot(X_r[topics == i, 0],
X_r[topics == i, 1],
X_r[topics == i, 2],
'o',
color=c,
label=l)
fontP = FontProperties()
if Z > 12:
columns = 2
else:
columns = 1
box = ax.get_position()
ax.set_position([box.x0, box.y0, 0.5, box.height])
plt.legend(prop=fontP,
loc='center left',
bbox_to_anchor=(1, 0.5),
ncol=columns)
plt.title('PCA of Study %s with Z=%s' % (study, str(z)))
plt.show()
return None
def main(*argv):
'''handles user input and runs different analysis functions using plsa model'''
parser = argparse.ArgumentParser(
description='This scripts runs plsa for a range of topic numbers.')
parser.add_argument('-s', '--study', help='The study number', default=None)
parser.add_argument('-n', '--name', help='The name of the dataset')
parser.add_argument('-z',
'--topics',
help='The range of topics to be run [z_start, z_end]',
nargs='+',
type=int,
required=True)
parser.add_argument('-z_inc',
'--increment',
help='Increment of topic numbers',
type=int,
default=1)
parser.add_argument('-useC',
help='use C code to run plsa',
action='store_true')
parser.add_argument('-run',
help='Specify the run number',
type=int,
default=1)
parser.add_argument('-topotus',
help='Specify to calculate the top otus',
action='store_true')
parser.add_argument(
'-n_otus',
help='Specify the number of top otus to return per topic',
type=int,
default=5)
parser.add_argument('-calculateX',
help='Specify to calculate X',
action='store_true')
args = parser.parse_args()
if args.study is None and args.name is None:
print "***Study number or a data name must be specified.***\n"
parser.print_help()
sys.exit()
if not args.calculateX and not args.topotus:
print "***Please specify an action to perform from the following: 'topotus', 'calculateX' "
parser.print_help()
sys.exit()
if args.study:
study = str(args.study)
else:
study = None
name = args.name
if len(args.topics) == 1:
z_i = args.topics[0]
z_f = args.topics[0]
elif len(args.topics) == 2:
z_i = min(args.topics)
if z_i < 2:
z_i = 2
z_f = max(args.topics)
else:
print "\n***Too many arguments specified for number of topics***\n"
parser.print_help()
sys.exit()
z_inc = args.increment
run = args.run
useC = args.useC
topotus = args.topotus
n_otus = args.n_otus
calculateX = args.calculateX
print(" Study: %s" % study)
print(" Name: %s" % name)
if z_i != z_f:
print(" Topics range: [{0} - {1}] in increments of {2}".format(
z_i, z_f, z_inc))
else:
print(" Topic number used: {0}".format(z_i))
print(" Using C: %s" % args.useC)
print(" Run number: %s" % args.run)
m = microbplsa.MicrobPLSA(study=study, name=name)
for z in range(z_i, z_f + 1, z_inc):
if topotus:
print "Finding top otus per each topic"
m.open_model(z=z, run=run, useC=True, folder=FOLDER)
analysis.top_otus(m, z, n_otus=n_otus)
if calculateX:
pass
def main(*argv):
'''handles user input and runs plsa'''
parser = argparse.ArgumentParser(
description=
'This scripts runs cross validations to determine the optimal number of topics.'
)
parser.add_argument(
'-action',
help='Action to perform from "all", "train", "test", "mse"',
required=True)
parser.add_argument('-k',
help='Number of folds in kFold cross validation',
type=int,
default=5)
parser.add_argument('-s', '--study', help='The study number', default=None)
parser.add_argument('-n', '--name', help='The name of the dataset')
parser.add_argument('-z',
'--topics',
help='The range of topics to be run [z_start, z_end]',
nargs='+',
type=int,
required=True)
parser.add_argument('-z_inc',
'--increment',
help='Increment of topic numbers',
type=int,
default=1)
parser.add_argument('-useC',
help='use C code to run plsa',
action='store_true')
parser.add_argument('-run',
help='Specify the number of the run',
type=int,
default=1)
parser.add_argument('-seed',
help='Random seed for kFold generator',
type=int,
default=2)
args = parser.parse_args()
if args.action not in ['all', 'train', 'test', 'mse']:
print "***The specified action is not recognized.***\n"
parser.print_help()
sys.exit()
else:
action = str(args.action)
k = args.k
if args.study is None and args.name is None:
print "***Study number or a data name must be specified.***\n"
parser.print_help()
sys.exit()
if args.study:
study = str(args.study)
else:
study = None
name = args.name
if len(args.topics) == 1:
z_i = args.topics[0]
z_f = args.topics[0]
elif len(args.topics) == 2:
z_i = min(args.topics)
if z_i < 2:
z_i = 2
z_f = max(args.topics)
else:
print "\n***Too many arguments specified for number of topics***\n"
parser.print_help()
sys.exit()
z_inc = args.increment
run = args.run
useC = args.useC
seed = args.seed
print(" Study: %s" % study)
print(" Name: %s" % name)
if z_i != z_f:
print(" Topics tested will be from {0} to {1} in increments of {2}".
format(z_i, z_f, z_inc))
else:
print(" Number of topics: {0}".format(z_i))
print(" Using C: %s" % args.useC)
print(" Number of run used: %s" % args.run)
print(" Action performed: %s" % action)
print(" Number of folds, k= : %s" % k)
m = microbplsa.MicrobPLSA(study=study, name=name)
mseAll = []
if action == 'train' or action == 'test' or action == 'all':
m.open_data()
print 'Data loaded.'
for z in range(z_i, z_f + 1, z_inc):
if action == 'train' or action == 'all':
kFolds = kf.create_folds(m, k, z, shuffle=True, seed=seed)
data = m.datamatrix
kf.train(k,
kFolds,
data,
study,
name,
z,
numRuns=run,
seed=seed,
useC=useC,
override=False)
if action == 'test' or action == 'all':
kFolds = kf.open_kFold(study, name, k, z)
kf.test(m, kFolds, k, z, useC=useC, seed=seed)
if action == 'mse' or action == 'all':
kFolds = kf.open_kFold(study, name, k, z)
mse = kf.measure_error(m, kFolds, k, z)
print "\n The cross validation error for study {0} with {1} topics and {2} folds is: {3} +/-{4}\n".format(
study, z, k, round(np.mean(mse), 5), round(np.std(mse), 5))
mse.insert(0, z)
mseAll.append(mse)
if mseAll:
kf.save_mse(mseAll, k, z, study, name, seed, run)
