if prim == 'PageRank':
# normalize per iteration
is_PR = True
df['m_teps_normalized'] = df['m_teps'] / df['search_depth']
chart = Chart(df).mark_bar().encode(
x=X('dataset:N', axis=Axis(title='Dataset', )),
y=Y(
'm_teps' if not is_PR else 'm_teps_normalized',
axis=Axis(title='MTEPS' if not is_PR else 'MTEPS (normalized)', ),
),
)
print chart.to_dict(data=False)
plotname = '%s_perf'
for fileformat in ['html', 'svg', 'png']:
savefile(chart, name=plotname % prim, fileformat=fileformat)
data_unfiltered = [json.load(open(jf)) for jf in all_json_files_1011nov16]
df = pandas.DataFrame(data_unfiltered)
# All prim-specific data is now stored in the pandas DataFrame "df".
df.loc[df.algorithm == 'DOBFS', 'algorithm'] = 'BFS'
chart = Chart(df).mark_point().encode(
x=X(
'dataset:N',
axis=Axis(title='Dataset', ),
),
column=Column('algorithm:N', axis=Axis(
title='Primitive',
orient='top',
)),
orient='bottom',
)),
y=Y(
'm_teps',
axis=Axis(title='MTEPS', ),
# scale=Scale(type='log'),
),
color=Color(
'traversal_mode:N',
legend=Legend(title='%s / Traversal Mode' % prim),
),
)
print chart.to_dict(data=False)
plotname = '%s_trav_chart'
for fileformat in ['html', 'svg', 'png']:
savefile(chart, name=plotname % prim, fileformat=fileformat)
# Average node degree chart
files = [(json_files_oct16, 'Real-World')]
if (prim == 'BFS'):
files.append((json_files_7nov16, 'RGG+RMAT'))
for file in files:
data_unfiltered = [json.load(open(jf)) for jf in file[0]]
df = pandas.DataFrame(data_unfiltered)
# "graph_type": "grmat",
# "graph_type": "rgg",
df['avg_node_degree'] = df['edges_visited'] / df['nodes_visited']
chart = Chart(df).mark_point().encode(
x=X(
chart = Chart(df).mark_bar().encode(
x=X('dataset:N',
axis=Axis(
title='Dataset',
)
),
y=Y('m_teps' if not is_PR else 'm_teps_normalized',
axis=Axis(
title='MTEPS' if not is_PR else 'MTEPS (normalized)',
),
),
)
print chart.to_dict(data=False)
plotname = '%s_perf'
for fileformat in ['html', 'svg', 'png']:
savefile(chart, name=plotname % prim, fileformat=fileformat)
data_unfiltered = [json.load(open(jf)) for jf in all_json_files_1011nov16]
df = pandas.DataFrame(data_unfiltered)
# All prim-specific data is now stored in the pandas DataFrame "df".
df.loc[df.algorithm == 'DOBFS', 'algorithm'] = 'BFS'
chart = Chart(df).mark_point().encode(
x=X('dataset:N',
axis=Axis(
title='Dataset',
),
),
column=Column('algorithm:N',
axis=Axis(
title='Primitive',
)
),
y=Y('m_teps',
axis=Axis(
title='MTEPS',
),
# scale=Scale(type='log'),
),
color=Color('traversal_mode:N',
legend=Legend(title='%s / Traversal Mode' % prim),
),
)
print chart.to_dict(data=False)
plotname = '%s_trav_chart'
for fileformat in ['html', 'svg', 'png']:
savefile(chart, name=plotname % prim, fileformat=fileformat)
# Average node degree chart
files = [(json_files_oct16, 'Real-World')]
if (prim == 'BFS'):
files.append((json_files_7nov16, 'RGG+RMAT'))
for file in files:
data_unfiltered = [json.load(open(jf)) for jf in file[0]]
df = pandas.DataFrame(data_unfiltered)
# "graph_type": "grmat",
# "graph_type": "rgg",
df['avg_node_degree'] = df['edges_visited'] / df['nodes_visited']
chart = Chart(df).mark_point().encode(
x=X('avg_node_degree',
# Read in the file
files = glob.glob('output/do_ab_random_?????*.svg')
for f in files:
stub = os.path.splitext(os.path.split(f)[1])[0]
print stub
with open(f, 'rb') as file:
filedata = file.read()
# Replace the target string
p = re.compile(r'^<svg class="marks" width="(\d+)" height="(\d+)"')
m = p.match(filedata)
w = m.group(1)
h = m.group(2)
# w: subtract 125
# h: add 48
whin = 'width="%s" height="%s"' % (w, h)
neww = int(w) - 125
newh = int(h) + 48
print 'w:', w, neww, 'h:', h, newh
whout = 'width="%s" height="%s"' % (neww, newh)
filedata = filedata.replace(whin, whout)
# 'width="590" height="462"', 'width="465" height="510"')
filedata = filedata.replace('translate(404.5,0.5)', 'translate(260,390)')
# Write the file out again
with open(f, 'wb') as file:
file.write(filedata)
savefile('', stub, 'pdf', 'output', patchFunctions=[])
from subprocess import Popen, PIPE, STDOUT, check_output, CalledProcessError
from fileops import savefile
# load data as a pandas DataFrame
cars = load_dataset('cars')
chart = Chart(cars).mark_point().encode(
x='Horsepower',
y='Miles_per_Gallon',
color='Origin',
)
print chart.to_dict(data=False)
savefile(chart, name='example', fileformat='html')
jsondir = '../gunrock-output/'
bfs_json_files = [f for f in os.listdir(jsondir)
if (os.path.isfile(jsondir + f) and
(os.path.splitext(f)[1] == ".json") and
(os.path.basename(f).startswith("BFS") or
os.path.basename(f).startswith("DOBFS")) and
not os.path.basename(f).startswith("_"))]
bfs_data_unfiltered = [json.load(open(jsondir + jf)) for jf in bfs_json_files]
bfs_df = pandas.DataFrame(bfs_data_unfiltered)
# All DOBFS/BFS data is now stored in the pandas DataFrame "bfs_df".
# Unfortunately, some of the runs in the repo have no dataset. Filter them out.
bfs_df = bfs_df[bfs_df['dataset'] != ""]
# Unfortunately, some of the runs in the repo have no dataset. Filter them out.
bfs_df = bfs_df[bfs_df['dataset'] != ""]
# Let's add a new column (attribute), conditional on existing columns.
# We'll need this to pivot later.
def setParameters(row):
return (row['algorithm'] + ', ' + ('un' if row['undirected'] else '') +
'directed, ' + ('' if row['mark_predecessors'] else 'no ') +
'mark predecessors')
bfs_df['parameters'] = bfs_df.apply(setParameters, axis=1)
bfs_chart = Chart(bfs_df).mark_bar().encode(
x=X('dataset', axis=Axis(title='Dataset')),
y=Y(
'm_teps',
axis=Axis(title='MTEPS'),
scale=Scale(type='log'),
),
)
print bfs_chart.to_dict(data=False)
# print bfs_chart.to_dict(data=False, cleanup_data=True)
# print bfs_chart.to_dict(data=True)
savefile(bfs_chart, name='bfs_chart', fileformat='html')
savefile(bfs_chart, name='bfs_chart', fileformat='json')
savefile(bfs_chart, name='bfs_chart', fileformat='md')
savefile_df(bfs_df, name='bfs_chart', fileformat='html')
# Unfortunately, some of the runs in the repo have no dataset. Filter them out.
bfs_df = bfs_df[bfs_df['dataset'] != ""]
# Let's add a new column (attribute), conditional on existing columns.
# We'll need this to pivot later.
def setParameters(row):
return (row['algorithm'] + ', ' +
('un' if row['undirected'] else '') + 'directed, ' +
('' if row['mark_predecessors'] else 'no ') + 'mark predecessors')
bfs_df['parameters'] = bfs_df.apply(setParameters, axis=1)
bfs_chart = Chart(bfs_df).mark_bar().encode(
x=X('dataset',
axis=Axis(title='Dataset')
),
y=Y('m_teps',
axis=Axis(title='MTEPS'),
scale=Scale(type='log'),
),
)
print bfs_chart.to_dict(data=False)
# print bfs_chart.to_dict(data=False, cleanup_data=True)
# print bfs_chart.to_dict(data=True)
savefile(bfs_chart, name='bfs_chart', fileformat='html')
savefile(bfs_chart, name='bfs_chart', fileformat='json')
savefile(bfs_chart, name='bfs_chart', fileformat='md')
