def main(input_files, output_file, levels):
fig = plt.figure()
ax = plt.subplot(111)
x_fun_name = "included_contigs_ratio"
y_fun_name = "precision"
axis_funs = RocAxisFuns(x_fun_name,y_fun_name)
for input_file in input_files:
data = ExperimentData(axis_funs)
data.load_data_frame(input_file)
data.standardize()
for level in levels:
data.classify(level)
data.calculate_roc()
# Plot ROC curve
for level in levels:
x = data.roc_data[level].x
y = data.roc_data[level].y
file_name = input_file.split('/')[-1]
plt.plot(x,y,label=file_name + ", " + level)
box = ax.get_position()
ax.set_position([box.x0,box.y0,box.width,box.height*0.8])
plt.legend(loc='upper center', bbox_to_anchor=(0.5, 1.35),
ncol=2, fancybox=True, shadow=True)
plt.xlabel(x_fun_name)
plt.ylabel(y_fun_name)
plt.savefig(output_file)
def main(input_files, level, fun_name):
fun = RocAxisFun(fun_name).fun
m_row = []
for input_file in input_files:
data = ExperimentData(fun)
data.load_data_frame(input_file)
data.standardize()
data.classify(level)
df = data.classification[level]
real_classif = [row[1] for row in df.values]
m_row.append(fun(real_classif, []))
formated_matrix_row = " & ".join([str(n) for n in m_row])
sys.stdout.write(formated_matrix_row)
def test_experiment_data(self):
axis_funs = RocAxisFuns("included_contigs_ratio", "precision")
d = ExperimentData(axis_funs)
cur_dir = os.path.dirname(__file__)
input_file = os.path.join(cur_dir, "fixtures/score_mul_test.tsv")
d.load_data_frame(input_file)
d.standardize()
# Random sample to test if the standardization works
assert_almost_equal(d.df[0:1].p_value_standardized[0], 0.7767953)
assert_almost_equal(d.df[5:6].p_value_standardized[5], -0.67591835)
assert_almost_equal(d.df[10:11].p_value_standardized[10], -0.1388929)
assert_almost_equal(d.df[12:13].p_value_standardized[12], -1.13672827)
d.classify("family")
d.classify("genus")
d.classify("genome")
df = d.classification["family"]
assert_equal(df[df.index == 1010].real_classif.values[0], False)
assert_equal(list(df.real_classif.values).count(True), 28)
df = d.classification["genus"]
assert_equal(df[df.index == 1010].real_classif.values[0], False)
assert_equal(df[df.index == 1013].real_classif.values[0], False)
assert_equal(df[df.index == 1027].real_classif.values[0], False)
assert_equal(list(df.real_classif.values).count(True), 26)
df = d.classification["genome"]
assert_equal(df[df.index == 1000].real_classif.values[0], False)
assert_equal(df[df.index == 1001].real_classif.values[0], False)
assert_equal(df[df.index == 1003].real_classif.values[0], False)
assert_equal(df[df.index == 1004].real_classif.values[0], False)
assert_equal(df[df.index == 1007].real_classif.values[0], False)
assert_equal(df[df.index == 1010].real_classif.values[0], False)
assert_equal(df[df.index == 1013].real_classif.values[0], False)
assert_equal(df[df.index == 1018].real_classif.values[0], False)
assert_equal(df[df.index == 1027].real_classif.values[0], False)
assert_equal(list(df.real_classif.values).count(True), 20)
#!/usr/bin/env python
import os
from corrbin.score import ExperimentData, RocAxisFun
file_path = os.path.realpath(__file__)
input_file = os.path.abspath(os.path.join(file_path,"..","..","fixtures/score_500000.tsv"))
fun = RocAxisFun("precision").fun
data = ExperimentData(fun)
data.load_data_frame(input_file)
data.standardize()
@profile
def my_fun(data):
data.classify("genome")
my_fun(data)
def my_fun():
fun = RocAxisFun("precision").fun
data = ExperimentData(fun)
data.load_data_frame(input_file)
data.standardize()