% pip install csv2graph2web
or
% pip install git+https://github.com/mastnk/csv2graph2web
or
% git clone https://github.com/mastnk/csv2graph2web
% cd csv2grap2web
% pip install .
% pip uninstall csv2graph2web
% cgw_copy --server yourhost.com --port 22 --login_name yourname --dir www/cgw
-
server specify your host server name.
-
port port for ssh (defulat: 22).
-
login_name specify your login name for the host server.
-
dir specify the directory which you want to install on the host sever.
% cgw_copy
% ls -l
Then, you can find the cgi file cgw.cgi and the directory graphs. Please upload those two to your web server.
% cgw_copy
% ls -l
Then, you can find the cgi file cgw.cgi and the directory graphs. Please upload those two to your web server.
from csv2graph2web import csv2graph2web as cgw
URL = '' # specify the URL for the cgw.cgi
csv_filename = 'a.csv'
line = 'epoch, loss, acc, val_loss, val_acc'
with open( csv_filename, 'w' ) as fout:
fout.write( line + '\n' )
graph_params = []
opt = {'xlabel':'Epoch', 'legend':True}
opt['title'] = 'acc'
graph_params.append( ([('epoch', 'acc'), ('epoch', 'val_acc')], opt.copy()) )
opt['title'] = 'loss'
graph_params.append( ([('epoch', 'loss'), ('epoch', 'val_loss')], opt.copy()) )
for epoch in range(300):
# some processes
# dummy code
loss = epoch
acc = epoch / 300
val_loss = loss * 0.9
val_acc = acc * 0.9
# dummy code
if( epoch % 10 ):
line = '{}, {}, {}, {}, {}'.format( epoch, loss, acc, val_loss, val_acc )
with open( csv_filename, 'a' ) as fout:
fout.write( line + '\n' )
try:
cgw.csv2graph2web( URL=URL, dir_name='tmp', csv_filename=csv_filename, graph_params=graph_params )
except:
pass
Then, you can see the graphs in the *URL*.
It generate the cgi for the web server.
% cgw_copy -h
-
URL
The url which represents the cgw.cgi. You can also check the graphs from this URL.
-
dir_name
The title of graphs. It appears in the html.
-
csv_filename
-
graph_params
graph_params = [ list_params, opt ]
-
list_params
list_params = [(x0,y0), (x1,y1), ...]
xn: The index of x axis
yn: The index of y axis
-
opt dictionary of the options: title, xlabel, ylabel
-