def make_viz_from_df(df, filename):
from clustergrammer import Network
net = Network()
net.df_to_dat(df)
net.swap_nan_for_zero()
# zscore first to get the columns distributions to be similar
net.normalize(axis='col', norm_type='zscore', keep_orig=True)
# filter the rows to keep the perts with the largest normalizes values
net.filter_N_top('row', 2000)
num_coluns = net.dat['mat'].shape[1]
if num_coluns < 50:
# views = ['N_row_sum', 'N_row_var']
views = ['N_row_sum']
net.make_clust(dist_type='cos', views=views)
filename = 'json/' + filename.split('/')[1].replace('.gct',
'') + '.json'
net.write_json_to_file('viz', filename)
def make_viz_json(inst_df, name):
from clustergrammer import Network
net = Network()
filename = 'json/'+name
load_df = {}
load_df['mat'] = inst_df
net.df_to_dat(load_df)
net.swap_nan_for_zero()
net.make_clust(views=[])
net.write_json_to_file('viz', filename, 'no-indent')
def process_GCT_and_export_tsv():
from clustergrammer import Network
filename = 'gcts/LDS-1003.gct'
print('exporting processed GCT as tsv file')
df = load_file(filename)
net = Network()
net.df_to_dat(df)
net.swap_nan_for_zero()
# zscore first to get the columns distributions to be similar
net.normalize(axis='col', norm_type='zscore', keep_orig=True)
# filter the rows to keep the perts with the largest normalizes values
net.filter_N_top('row', 200)
net.write_matrix_to_tsv('txt/example_gct_export.txt')
def process_GCT_and_export_tsv():
from clustergrammer import Network
filename = 'gcts/LDS-1003.gct'
print('exporting processed GCT as tsv file')
df = load_file(filename)
net = Network()
net.df_to_dat(df)
net.swap_nan_for_zero()
# zscore first to get the columns distributions to be similar
net.normalize(axis='col', norm_type='zscore', keep_orig=True)
# filter the rows to keep the perts with the largest normalizes values
net.filter_N_top('row', 200)
net.write_matrix_to_tsv('txt/example_gct_export.txt')
def make_viz_from_df(df, filename):
from clustergrammer import Network
net = Network()
net.df_to_dat(df)
net.swap_nan_for_zero()
# zscore first to get the columns distributions to be similar
net.normalize(axis='col', norm_type='zscore', keep_orig=True)
# filter the rows to keep the perts with the largest normalizes values
net.filter_N_top('row', 2000)
num_coluns = net.dat['mat'].shape[1]
if num_coluns < 50:
# views = ['N_row_sum', 'N_row_var']
views = ['N_row_sum']
net.make_clust(dist_type='cos', views=views)
filename = 'json/' + filename.split('/')[1].replace('.gct','') + '.json'
net.write_json_to_file('viz', filename)
print(new_rows)
for inst_row in new_rows:
inst_row.append('something')
print('\n\n\n')
print(new_rows)
new_rows = [tuple(x) for x in new_rows]
print('\n\n\n')
print(new_rows)
df['mat'].index = new_rows
net.df_to_dat(df)
# for inst_row in all_rows:
# inst_row = list(inst_row)
# inst_row.append('something')
# inst_row = tuple(inst_row)
# print(all_rows)
# cat_list = [tuple(x) for x in cat_list]
# print(net.dat['node_info'])
# net.dat['nodes']['row'] = []
