def setUp(self):
x = [0.5, 2, 4, 5.5, 6.5, 7, 7.25, 7.3]
y = [i * 2 for i in x]
t = [i for i in range(len(x))]
ts = TimeSeries(values=x, time=t, feature='x')
ts2 = TimeSeries(values=y, time=t, feature='y')
self.tsg = TimeSeriesGroup([ts, ts2])
def cluster(self):
data = self.tsg.as_df()
self.agg_clust.fit(data)
labels = self.agg_clust.labels_
df = self.tsg.as_df()
df['cluster'] = labels
mon = Monitor(dire=self.dire)
for label, df in df.groupby(by='cluster'):
df = df.drop('cluster', axis=1)
mon[label] = TimeSeriesGroup(df)
return mon
def do_kscan(self):
kscan = {}
for i in self.ks:
agg_clust_i = AgglomerativeClustering(n_clusters=i, **self.kwargs)
agg_clust_i.fit(self.tsg.as_df())
labels = agg_clust_i.labels_
df = self.tsg.as_df()
df['cluster'] = labels
mon = Monitor(dire=self.dire)
for label, df in df.groupby(by='cluster'):
df = df.drop('cluster', axis=1)
mon[label] = TimeSeriesGroup(df)
kscan[i] = mon
return kscan
def __init__(self, tsg, nclust=1, db_file=None,
dtw_dist=euclidean, fast=False, radius=1):
self.nclust = nclust
self.dtw_dist = dtw_dist
self.fast = fast
self.radius = radius
self.tsg = tsg
self.clusters = {i: TimeSeriesGroup(self.tsg[i]) for i in self.tsg.features}
self.evolution_of_clusters = {}
self.db_file = db_file
if self.db_file is None:
self.db_file = os.path.join(os.getcwd(), os.path.split(__file__)[1][:-3]+'.db')
if not isinstance(self.nclust, int):
raise ValueError("nclust should be of type int")
def setUp(self):
x = [0.5, 2, 4, 5.5, 6.5, 7, 7.25, 7.3]
y = [i * 2 for i in x]
t = [i for i in range(len(x))]
ts = TimeSeries(values=x, time=t, feature='x')
ts2 = TimeSeries(values=y, time=t, feature='y')
self.tsg = TimeSeriesGroup([ts, ts2])
a = Node('A')
b = Node('B')
c = Node('C')
d = Node('D')
self.first_order_deg = Edge(source=a, target=None, id=1)
self.zero_order_prod = Edge(source=None, target=b, id=2)
self.first_order_transition = Edge(source=a, target=b, id=3)
self.complex_break = Edge(source=a, target=[b, c], id=4)
self.complex_form = Edge(source=[a, b], target=c, id=5)
self.second_order_transition = Edge(source=[a, b], target=[c, d], id=6)
self.second_order_deg = Edge(source=[a, b], target=None, id=7)
def result(self):
ys = [i.y.feature for i in self.dtw]
return TimeSeriesGroup(self.tsg.loc[ys])
def get_tsg(go_id, tsg):
dups = get_dups(go_id, tsg)
return TimeSeriesGroup(data.loc[dups])
site.addsitedir(r'/home/b3053674/Documents')
from pytseries.core import TimeSeries, TimeSeriesGroup
from pytseries.clust import TimeSeriesKMeans
dire = r'/home/b3053674/Documents/pytseries/Microarray2/GOclustering'
microarray_data_file = r'/home/b3053674/Documents/pytseries/Microarray2/MeanMicroarrayDEGS_gt_75.csv'
data = pandas.read_csv(microarray_data_file, index_col=[0])
data.columns = [int(i) for i in data.columns]
data = data.dropna(how='all')
tsg = TimeSeriesGroup(data)
tsg = tsg.interpolate(num=30)
tsg = tsg.norm()
pattern = os.path.join(dire, 'GO*.csv')
files = glob.glob(pattern)
df_dct = {}
for i in files:
go_term = os.path.split(i)[1][:-4]
df_dct[go_term] = pandas.read_csv(i, index_col=[0])
go_df = pandas.concat(df_dct)
go_df.columns = ['genes']