def test_centrality(centrality_f, reverse=False):
node_good = (pnode_good, version_good)
node_bad = (pnode_bad, version_bad)
dg_good, nodes_good = graph.make_graph(good)
name = histogram.get_name(dg_good, node_good)
if reverse:
rank_good = centrality_f(dg_good.reverse(copy=True))
else:
rank_good = centrality_f(dg_good)
counts_good = histogram.aggregate(dg_good, rank_good)
kdes_good = histogram.counts_to_kdes(counts_good)
# REMOVE
print counts_good[name]
print "mean", np.mean(counts_good[name])
print
print
print ">>>>>GOOD ON GOOD"
r_good = rank_good[node_good]
pre_good = histogram.kde_predict(kdes_good[name], rank_good[node_good])
print node_good, name, r_good, pre_good
dg_bad, nodes_bad = graph.make_graph(bad)
name = histogram.get_name(dg_bad, node_bad)
if histogram.SHOULD_HARDCODE_GCC:
name = "/usr/bin/gcc"
print ">>>>>BAD ON GOOD"
if reverse:
rank_bad = centrality_f(dg_bad.reverse(copy=True))
else:
rank_bad = centrality_f(dg_bad)
r_bad = rank_bad[node_bad]
pre_bad = histogram.kde_predict(kdes_good[name], rank_bad[node_bad])
print node_bad, name, r_bad, pre_bad
print ">>>>>COMPARISON"
print "diff (good - bad)", name, "rank diff", r_good - r_bad, "prediction diff", pre_good - pre_bad
def test_centrality(centrality_f, reverse=False):
node_good = (pnode_good, version_good)
node_bad = (pnode_bad, version_bad)
dg_good, nodes_good = graph.make_graph(good)
name = histogram.get_name(dg_good, node_good)
if reverse:
rank_good = centrality_f(dg_good.reverse(copy=True))
else:
rank_good = centrality_f(dg_good)
counts_good = histogram.aggregate(dg_good, rank_good)
kdes_good = histogram.counts_to_kdes(counts_good)
# REMOVE
print counts_good[name]
print "mean", np.mean(counts_good[name])
print
print
print ">>>>>GOOD ON GOOD"
r_good = rank_good[node_good]
pre_good = histogram.kde_predict(kdes_good[name], rank_good[node_good])
print node_good, name, r_good, pre_good
dg_bad, nodes_bad = graph.make_graph(bad)
name = histogram.get_name(dg_bad, node_bad)
if histogram.SHOULD_HARDCODE_GCC:
name = "/usr/bin/gcc"
print ">>>>>BAD ON GOOD"
if reverse:
rank_bad = centrality_f(dg_bad.reverse(copy=True))
else:
rank_bad = centrality_f(dg_bad)
r_bad = rank_bad[node_bad]
pre_bad = histogram.kde_predict(kdes_good[name], rank_bad[node_bad])
print node_bad, name, r_bad, pre_bad
print ">>>>>COMPARISON"
print "diff (good - bad)", name, "rank diff", r_good - r_bad, "prediction diff", pre_good - pre_bad
maxx = 1
numbins = 3
else:
numbins = 10
x = np.linspace(minx, maxx, 100)
axis.hist(l, bins=numbins, color='g')
axis.tick_params(axis='y', colors='g')
axis.set_xlim((minx, maxx))
axis.set_title(name)
# plot kde
ax2 = axis.twinx()
axarr2[row, col] = ax2
kde = h.kde_make(l)
y = [h.kde_predict(kde, xp) for xp in x]
ax2.plot(x, y, color='b')
col = (col + 1) % dim
if col == 0:
row += 1
# And finally we get to plot things...
colors = ('Green', 'Red', 'Blue')
#for ax, color in zip(axes, colors):
# data = np.random.random(1) * np.random.random(10)
# ax.plot(data, marker='o', linestyle='none', color=color)
# ax.set_ylabel('%s Thing' % color, color=color)
# ax.tick_params(axis='y', colors=color)
#ax.axvline(x=7.4145, color='r')
fig.tight_layout()
maxx = 1
numbins = 3
else:
numbins = 10
x = np.linspace(minx, maxx, 100)
axis.hist(l, bins=numbins, color="g")
axis.tick_params(axis="y", colors="g")
axis.set_xlim((minx, maxx))
axis.set_title(name)
# plot kde
ax2 = axis.twinx()
axarr2[row, col] = ax2
kde = h.kde_make(l)
y = [h.kde_predict(kde, xp) for xp in x]
ax2.plot(x, y, color="b")
col = (col + 1) % dim
if col == 0:
row += 1
# And finally we get to plot things...
colors = ("Green", "Red", "Blue")
# for ax, color in zip(axes, colors):
# data = np.random.random(1) * np.random.random(10)
# ax.plot(data, marker='o', linestyle='none', color=color)
# ax.set_ylabel('%s Thing' % color, color=color)
# ax.tick_params(axis='y', colors=color)
# ax.axvline(x=7.4145, color='r')
fig.tight_layout()
