def input(self,filename):
# Parameter file
self.parameters = dict()
paramfile = open(filename, 'r')
for line in paramfile:
contents = line.split('\t')
self.parameters[contents[0]] = contents[1].strip()
# Abundances
self.abundances = dict()
abundfile = open(PyPluMA.prefix()+"/"+self.parameters['abundances'], 'r')
header = abundfile.readline() # Header
for line in abundfile:
contents = line.split('\t')
self.abundances["\""+contents[0]+"\""] = float(contents[1])
# Clusters
self.clusters = []
clusterfile = open(PyPluMA.prefix()+"/"+self.parameters['clusters'], 'r')
clustercount = 0
for line in clusterfile:
if (line.strip() == "\"\",\"x\""):
self.clusters.append([])
clustercount += 1
else:
contents = line.split(',')
self.clusters[clustercount-1].append(contents[1].strip())
# Centroids
self.centroids = []
centroidfile = open(PyPluMA.prefix()+"/"+self.parameters['centroids'], 'r')
header = centroidfile.readline()
for line in centroidfile:
contents = line.split(',')
self.centroids.append(contents[1].strip())
def input(self, infile):
inputfile = open(infile, 'r')
self.parameters = dict()
for line in inputfile:
contents = line.strip().split('\t')
self.parameters[contents[0]] = contents[1]
self.infl1 = PyPluMA.prefix() + "/" + self.parameters["infile1"]
self.infl2 = PyPluMA.prefix() + "/" + self.parameters["infile2"]
def input(self, filename):
self.txtfile = open(filename, 'r')
self.parameters = dict()
for line in self.txtfile:
contents = line.split('\t')
self.parameters[contents[0]] = contents[1].strip()
if len(PyPluMA.prefix()) != 0:
self.parameters['csvfile'] = PyPluMA.prefix(
) + "/" + self.parameters['csvfile']
def run(self):
self.newlines = []
column = self.parameters['column']
csvfile = open(self.parameters['csvfile'], 'r')
criteria = self.parameters['criteria']
self.header = csvfile.readline().strip(
) # First line has the target column
headercontents = self.header.split(',')
tocheck = []
if (column[0] == '\"'): # Take out quotes
column = column[1:len(column) - 1]
for i in range(len(headercontents)):
if (headercontents[i] != '\"\"' and headercontents[i][0] == '\"'):
headercontents[i] = headercontents[i][1:len(headercontents[i]
) - 1]
#if (headercontents[i][0] == 'V'):
# print(headercontents[i].startswith(column))
# print(headercontents[len(column)])
# print(is_number(headercontents[len(column)+1:]))
if (headercontents[i] == column):
print(headercontents[i] + "\n")
tocheck.append(i)
elif (headercontents[i].startswith(column)
and # For cases where you have more than one strain
headercontents[i][len(column)] == '-' and is_number(
headercontents[i][len(column) + 1:])):
print(headercontents[i] + "\n")
tocheck.append(i)
#if (headercontents.count(column) == 0):
if (len(tocheck) == 0):
PyPluMA.log("[CSVScreen] WARNING: TARGET COLUMN NOT FOUND")
else:
self.newlines = []
for line in csvfile:
line = line.strip()
contents = line.split(',')
if (criteria == "nonzero"
): # Only one has to be nonzero to keep it
for targetindex in tocheck:
if (not is_number(contents[targetindex])):
self.newlines.append(line)
elif (float(contents[targetindex]) != 0):
self.newlines.append(line)
break
else: # All have to be zero to keep it
isZero = True
for targetindex in tocheck:
if (not is_number(contents[targetindex])):
isZero = False
break
elif (is_number(contents[targetindex])
and float(contents[targetindex]) != 0):
isZero = False
break
if (isZero):
self.newlines.append(line)
def input(self, filename):
# Format expected:
# correlationfile <somefile.csv>
# pathwayfile <somefile.txt>
thefile = open(filename, 'r')
for line in thefile:
myline = line.strip()
entries = myline.split('\t')
if (entries[0] == 'correlationfile'):
self.myfile = PyPluMA.prefix() + "/" + entries[1]
elif (entries[0] == 'pathwayfile'):
self.mypathways = PyPluMA.prefix() + "/" + entries[1]
def output(self, filename):
#gmlfilename = self.myfile[0:len(self.myfile)-3] + "gml"
#gmlfile = open(gmlfilename, 'w')
noafile = open(filename, 'w')
PyPluMA.log("Writing NOA file ")
noafile.write("name\tMM\n")
for bac in self.bacteria:
bac = bac.strip()
if (bac[0] == '\"'):
bac = bac[1:len(bac) - 1]
if (bac[0] == 'x'):
noafile.write(bac + "\t1\n")
else:
noafile.write(bac + "\t0\n")
def run(self):
self.parameters = dict()
parameterfile = open(self.myfile, 'r')
for line in parameterfile:
contents = line.strip().split('\t')
self.parameters[contents[0]] = contents[1]
filestuff = open(PyPluMA.prefix() + "/" + self.parameters["csvfile"],
'r')
self.firstline = filestuff.readline().strip()
lines = []
for line in filestuff:
lines.append(line)
self.m = len(lines)
self.samples = []
self.bacteria = self.firstline.split(',')
if (self.bacteria.count('\"\"') != 0):
self.bacteria.remove('\"\"')
self.n = len(self.bacteria)
self.ADJ = []
i = 0
for i in range(self.m):
self.ADJ.append([])
tmpsample = []
contents = lines[i].split(',')
self.samples.append(contents[0])
for j in range(self.n):
value = float(contents[j + 1].strip())
self.ADJ[i].append(value)
i += 1
j = 0
while (j < len(self.bacteria)):
#zeroflag = True
nonzerocount = 0
for i in range(self.m):
if (self.ADJ[i][j] > float(self.parameters["minval"])):
#print(self.ADJ[i][j])
#zeroflag = False
nonzerocount += 1
#break
#if (zeroflag):
if ((float(nonzerocount) / float(self.m)) < float(
self.parameters["threshold"])):
print("DELETING " + self.bacteria[j] + " " +
str((float(nonzerocount) / float(self.m))))
self.firstline = self.firstline.replace(self.bacteria[j], "")
if (self.firstline.find(",,") != -1):
self.firstline = self.firstline.replace(",,", ",")
if (self.firstline.endswith(',')):
self.firstline = self.firstline[:len(self.firstline) - 1]
del self.bacteria[j]
for i in range(self.m):
del self.ADJ[i][j]
else:
j += 1
self.n = len(self.bacteria)
def input(self, filename):
filestuff = open(filename, 'r')
networkdata = dict()
for line in filestuff:
keyval = line.split('\t')
networkdata[keyval[0]] = keyval[1].strip()
if (not ('correlations' in networkdata)):
PyPluMA.log("Error in CSV2DotPlugin, no correlations file defined")
sys.exit(1)
else:
self.correlations = open(networkdata['correlations'], 'r')
if ('abundances' in networkdata):
self.abundances = open(networkdata['abundances'], 'r')
else:
self.abundances = None
if ('clusters' in networkdata):
self.clusters = open(networkdata['clusters'], 'r')
else:
self.clusters = None
def output(self, outputfile):
samples = self.firstline
samples.remove('taxlevel')
samples.remove('rankID')
samples.remove('taxon')
samples.remove('daughterlevels')
samples.remove('total')
#samples.sort()
newfirstline = "\"\","
for i in range(len(samples)):
newfirstline += "\"" + samples[i] + "\""
if (i != len(samples) - 1):
newfirstline += ','
else:
newfirstline += '\n'
directoryflag = False
if (os.path.exists(PyPluMA.prefix() + "/kingdom")):
directoryflag = True
outputfile = outputfile.replace(PyPluMA.prefix(), '')
for i in range(len(self.classify)):
if (not directoryflag):
filestuff = open(outputfile + "." + self.classify[i] + ".csv",
'w')
else:
filestuff = open(
PyPluMA.prefix() + "/" + self.classify[i] + "/" +
outputfile + ".csv", 'w')
filestuff.write(newfirstline)
for j in range(len(self.taxa[i])):
mylines = self.lines[i]
filestuff.write("\"" + self.taxa[i][j] + "\"" + ",")
contents = mylines[j].split('\t')
contents = contents[5:]
for k in range(len(contents)):
filestuff.write(contents[k])
if (k != len(contents) - 1):
filestuff.write(',')
def output(self, filename):
#gmlfilename = self.myfile[0:len(self.myfile)-3] + "gml"
#gmlfile = open(gmlfilename, 'w')
dotfile = open(filename, 'w')
PyPluMA.log("Writing Dot file ")
dotfile.write("graph {\n")
dotfile.write("forcelabels=true;\n")
dotfile.write("penwidth=10;\n")
if (self.abundances != None):
for i in range(self.n):
dotfile.write(self.bacteria[i]+"[label="+self.bacteria[i]+",fontsize="+str(int(200+self.ABUND[i]*100))+",penwidth=10,width="+str('{:f}'.format(20+self.ABUND[i]*10))+",height="+str('{:f}'.format(20+self.ABUND[i]*10))+"];\n")
else:
for i in range(self.n):
dotfile.write(self.bacteria[i]+"[label="+self.bacteria[i]+",fontsize=100,penwidth=10,width=10,height=10];\n")
for i in range(self.n):
for j in range(i+1, self.n):
if (self.ADJ[i][j] != 0):
dotfile.write(self.bacteria[i]+" -- "+self.bacteria[j]+"[w="+str(self.ADJ[i][j])+",color=\"")
if (self.ADJ[i][j] < 0):
dotfile.write("red")
else:
dotfile.write("green")
dotfile.write("\";penwidth="+str(self.ADJ[i][j]*10)+"];\n")
if (self.clusters != None):
count = 0
for cluster in self.CLUST:
dotfile.write("subgraph clusterC"+str(count)+" { ")
for node in cluster:
dotfile.write(node+"; ")
dotfile.write("}\n")
count += 1
dotfile.write("}\n")
def run(self):
filestuff = open(self.myfile, 'r')
file1 = open(PyPluMA.prefix() + "/" + filestuff.readline().strip(),
'r')
file2 = open(PyPluMA.prefix() + "/" + filestuff.readline().strip(),
'r')
self.firstline1 = file1.readline().strip()
self.bacteria1 = self.firstline1.split(',')
if (self.bacteria1.count('\"\"') != 0):
self.bacteria1.remove('\"\"')
self.firstline2 = file2.readline().strip()
self.bacteria2 = self.firstline2.split(',')
if (self.bacteria2.count('\"\"') != 0):
self.bacteria2.remove('\"\"')
self.p = len(self.bacteria1)
self.q = len(self.bacteria2)
self.ADJ = []
i = 0
for line in file1:
contents = line.split(',')
self.ADJ.append([])
for j in range(self.p):
value = float(contents[j + 1])
self.ADJ[i].append(value)
line2 = file2.readline()
contents2 = line2.split(',')
for j in range(self.q):
value = float(contents2[j + 1])
self.ADJ[i].append(value)
i += 1
self.m = len(self.ADJ)
self.n = self.p + self.q
def run(self):
filestuff = open(self.myfile, 'r')
firstline = filestuff.readline().strip()
if (len(PyPluMA.prefix()) != 0):
firstline = PyPluMA.prefix() + "/" + firstline
firstfile = open(firstline, 'r')
# Use first file to get indices
self.firstline = firstfile.readline().strip()
lines = []
for line in firstfile:
lines.append(line.strip())
self.m = len(lines)
self.samples = []
self.bacteria = self.firstline.split(',')
if (self.bacteria.count('\"\"') != 0):
self.bacteria.remove('\"\"')
self.n = len(self.bacteria)
self.ADJ = [] #numpy.zeros([self.m, self.n])
i = 0
for i in range(self.m):
self.ADJ.append([])
contents = lines[i].split(',')
self.samples.append(contents[0])
for j in range(self.n):
#print contents[j+1]
value = contents[j + 1].strip() #float(contents[j+1].strip())
#print self.ADJ[i][j]
self.ADJ[i].append(value) #[j] = value
i += 1
for line in filestuff:
myline = line.strip()
if (len(PyPluMA.prefix()) != 0):
myline = PyPluMA.prefix() + "/" + myline
newfile = open(myline, 'r')
firstline = newfile.readline().strip()
bac = firstline.split(',')
if (bac.count('\"\"') != 0):
bac.remove('\"\"')
lines = []
for line2 in newfile:
lines.append(line2.strip())
for line2 in lines:
contents = line2.split(',')
bac2 = contents[0]
if (bac2 in self.samples):
x = self.samples.index(bac2)
else:
self.ADJ.append([])
for i in range(0, self.n):
self.ADJ[len(self.ADJ) - 1].append(0)
x = len(self.ADJ) - 1
self.m += 1
self.samples.append(bac2)
for j in range(1, len(contents)):
if (bac[j - 1] in self.bacteria):
#print("FOUND "+bac[j-1]+", NOT APPENDING")
#xxx = input()
y = self.bacteria.index(bac[j - 1])
#print(x)
#print(len(self.ADJ))
#print(y)
#print("ROW LENGTH "+str(x))
#print(len(self.ADJ[x]))
#print(j)
#print(len(contents))
#print(len(self.bacteria))
#print(self.n)
self.ADJ[x][y] = contents[j].strip()
else:
self.bacteria.append(bac[j - 1])
#print("APPENDING: "+str(len(self.bacteria)))
#xxx = input()
self.n += 1
#print(self.n)
for row in range(len(self.ADJ)):
self.ADJ[row].append(0.0)
self.ADJ[x][len(self.ADJ[x]) - 1] = contents[j].strip()
def input(self, filename):
self.parameters = dict()
paramfile = open(filename, 'r')
for line in paramfile:
contents = line.split('\t')
self.parameters[contents[0]] = contents[1].strip()
classnames_file = open(
PyPluMA.prefix() + "/" + self.parameters['classnames'], 'r')
self.class_names = []
for line in classnames_file:
self.class_names.append(line.strip())
train_file = open(PyPluMA.prefix() + "/" + self.parameters['trainset'],
'r')
pos = 0
train_image_list = []
train_label_list = []
for line in train_file:
line = line.strip()
contents = line.split(',')
data = numpy.asarray(im.open(PyPluMA.prefix() + "/" + contents[0]))
print("READING FILE " + contents[0])
train_image_list.append([data])
train_label_list.append(
numpy.asarray([[self.class_names.index(contents[1])]]))
pos += 1
train_images = numpy.vstack(tuple(train_image_list))
train_labels = numpy.vstack(tuple(train_label_list))
self.inputfilenames = []
test_file = open(PyPluMA.prefix() + "/" + self.parameters['testset'],
'r')
pos = 0
test_image_list = []
test_label_list = []
for line in test_file:
line = line.strip()
contents = line.split(',')
data = numpy.asarray(im.open(PyPluMA.prefix() + "/" + contents[0]))
self.inputfilenames.append(contents[0])
print("READING FILE " + contents[0])
test_image_list.append([data])
test_label_list.append(
numpy.asarray([[self.class_names.index(contents[1])]]))
pos += 1
self.test_images = numpy.vstack(tuple(test_image_list))
self.test_labels = numpy.vstack(tuple(test_label_list))
# Normalize pixel values to be between 0 and 1
train_images, self.test_images = train_images / 255.0, self.test_images / 255.0
self.model = models.Sequential()
tensorfile = open(PyPluMA.prefix() + "/" + self.parameters['tensor'],
'r')
iter = 0
for line in tensorfile:
contents = line.strip().split('\t')
if (iter == 0):
print(len(train_images[0]))
print(len(train_images[0][0]))
self.model.add(
layers.Conv2D(int(contents[0]),
(int(contents[1]), int(contents[2])),
activation=contents[3],
input_shape=(len(train_images[0]),
len(train_images[0][0]),
3))) # Assuming RGB (3)
else:
self.model.add(layers.MaxPooling2D((2, 2)))
self.model.add(
layers.Conv2D(int(contents[0]),
(int(contents[1]), int(contents[2])),
activation=contents[3]))
iter += 1
#self.model.add(layers.MaxPooling2D((2, 2)))
#self.model.add(layers.Conv2D(64, (3, 3), activation='relu'))
self.model.summary()
self.model.add(layers.Flatten())
densefile = open(PyPluMA.prefix() + "/" + self.parameters['dense'])
for line in densefile:
contents = line.strip().split('\t')
self.model.add(
layers.Dense(int(contents[0]), activation=contents[1]))
self.model.add(layers.Dense(len(self.class_names)))
self.model.summary()
self.model.compile(optimizer=self.parameters['optimize'],
loss=tf.keras.losses.SparseCategoricalCrossentropy(
from_logits=True),
metrics=[self.parameters['metric']])
#print(self.model.output)
#print(dir(self.model))
#exit()
history = self.model.fit(train_images,
train_labels,
epochs=int(self.parameters['epochs']),
validation_data=(self.test_images,
self.test_labels))
def input(self, filename):
tempstuff = open(filename, 'r')
self.myfile = PyPluMA.prefix() + "/" + tempstuff.readline().strip()
self.mycountfile = PyPluMA.prefix() + "/" + tempstuff.readline().strip(
)