def testRun_matrixWithOffset(self):
m = list(create_run_matrix(10, 10))
self.assertEqual(len(m), 10 * 10)
for mc, data in m:
self.assertGreaterEqual(mc, 5)
self.assertGreaterEqual(data, 5)
self.assertLessEqual(mc, 15)
self.assertLessEqual(data, 15)
def testRun_matrixWithOffset(self):
m = list(create_run_matrix(10, 10))
self.assertEqual(len(m), 10 * 10)
for mc, data in m:
self.assertGreaterEqual(mc, 5)
self.assertGreaterEqual(data, 5)
self.assertLessEqual(mc, 15)
self.assertLessEqual(data, 15)
def testRun_matrixSimple(self):
m = create_run_matrix(10, 10)
self.assertEqual(len(list(m)), 10 * 10)
for mc, data in m:
self.assertGreaterEqual(mc, 1)
self.assertGreaterEqual(data, 1)
self.assertLessEqual(mc, 10)
self.assertLessEqual(data, 10)
def testRun_matrixSimple(self):
m = create_run_matrix(10, 10)
self.assertEqual(len(list(m)), 10 * 10)
for mc, data in m:
self.assertGreaterEqual(mc, 1)
self.assertGreaterEqual(data, 1)
self.assertLessEqual(mc, 10)
self.assertLessEqual(data, 10)
def split(self, n):
'''
In this case the split function has to establish the size of the
matrix and create jobs with sub-matrices.
Case 1:
- use_n_toy = 10
- offset_toy_mc = 0
- offset_toy_data = 0
- n = 5
Case 2:
- use_n_toy = 10
- offset_toy_mc = 5
- offset_toy_data = 5
- n = 5
For case 1 we can either construct five 4 x 5, 5 x 4 or 2 x 10
matrices. For case 2 we can construct five 1 x 5 or 5 x 1 matrices.
In either case it would make sense to have the smallest unit either
a column or a row.
In order not to double-count, use_n_toy needs to be reduced per job.
'''
import dps.analysis.unfolding_tests.create_unfolding_pull_data as pull
if n == 1:
return self
run_matrix = pull.create_run_matrix(self.n_toy_mc,
self.n_toy_data,
self.offset_toy_mc,
self.offset_toy_data)
self.run_matrix = list(run_matrix)
# after this the run matrix is empty (generator)!
l = list(self.run_matrix)
n_per_job = int(len(l) / n)
run_matrices = []
for _ in range(n):
run_matrix = []
for _ in range(n_per_job):
if l:
run_matrix.append(l.pop())
run_matrices.append(run_matrix)
if l: # if anything is left
for i in l:
run_matrices[-1].append(i)
jobs = []
for r in run_matrices:
offsets_and_ranges = UnfoldingPullJob.get_offsets_and_ranges(r)
offset_toy_mc, offset_toy_data = offsets_and_ranges[:2]
n_toy_mc, n_toy_data = offsets_and_ranges[2:]
j = UnfoldingPullJob(self.input_file_name, self.method,
self.channel, self.centre_of_mass,
self.variable, n_toy_mc,
n_toy_data,
self.output_folder, offset_toy_mc,
offset_toy_data, self.k_value,
self.tau_value, r)
jobs.append(j)
return jobs