From 663caa10de2a1320763d503207e98e8f3944b0a4 Mon Sep 17 00:00:00 2001
From: Andrej Rode <mail@andrejro.de>
Date: Thu, 15 Mar 2018 17:17:40 +0100
Subject: analog: rework qa_fastnoise, add reproducibility test
---
gr-analog/python/analog/qa_fastnoise.py | 115 ++++++++++++++++++++------------
1 file changed, 72 insertions(+), 43 deletions(-)
(limited to 'gr-analog/python/analog/qa_fastnoise.py')
diff --git a/gr-analog/python/analog/qa_fastnoise.py b/gr-analog/python/analog/qa_fastnoise.py
index 605a0552ac..848b045b9a 100644
--- a/gr-analog/python/analog/qa_fastnoise.py
+++ b/gr-analog/python/analog/qa_fastnoise.py
@@ -23,67 +23,96 @@
from gnuradio import gr, gr_unittest, analog, blocks
import numpy
+
class test_fastnoise_source(gr_unittest.TestCase):
def setUp (self):
- self.tb = gr.top_block ()
- def tearDown (self):
- self.tb = None
+ self.num = 2**22
+ self.num_items = 10**6
+ self.default_args = {"samples": self.num, "seed": 43, "ampl": 1}
- def test_001_test_moments(self):
- tb = self.tb
+ def tearDown (self):
+ pass
- NUM = 2**22
- NUM_ITEMS = 10**6
- DEFAULTARGS = {"samples": NUM, "seed": 43, "ampl": 1}
- self.uni_real_src = analog.fastnoise_source_f(type = analog.GR_UNIFORM, **DEFAULTARGS)
- self.uni_cplx_src = analog.fastnoise_source_c(type = analog.GR_UNIFORM, **DEFAULTARGS)
+ def run_test_real(self, form):
+ """ Run test case with float input/output
+ """
+ tb = gr.top_block()
+ src = analog.fastnoise_source_f(type=form, **self.default_args)
+ head = blocks.head(nitems=self.num_items, sizeof_stream_item=gr.sizeof_float)
+ sink = blocks.vector_sink_f()
+ tb.connect(src, head, sink)
+ tb.run()
+ return numpy.array(sink.data())
- self.norm_real_src = analog.fastnoise_source_f(type = analog.GR_GAUSSIAN, **DEFAULTARGS)
- self.norm_cplx_src = analog.fastnoise_source_c(type = analog.GR_GAUSSIAN, **DEFAULTARGS)
+ def run_test_complex(self, form):
+ """ Run test case with complex input/output
+ """
+ tb = gr.top_block()
+ src = analog.fastnoise_source_c(type=form, **self.default_args)
+ head = blocks.head(nitems=self.num_items, sizeof_stream_item=gr.sizeof_gr_complex)
+ sink = blocks.vector_sink_c()
+ tb.connect(src, head, sink)
+ tb.run()
+ return numpy.array(sink.data())
- self.lapl_real_src = analog.fastnoise_source_f(type = analog.GR_LAPLACIAN, **DEFAULTARGS)
+ def test_001_real_uniform_moments(self):
- types = {
- self.uni_real_src: "uniform", self.uni_cplx_src: "uniform",
- self.norm_cplx_src: "norm", self.norm_cplx_src: "norm",
- self.lapl_real_src: "laplace"
- }
+ data = self.run_test_real(analog.GR_UNIFORM)
- real_srcs = [self.uni_real_src, self.norm_real_src, self.lapl_real_src]
- cplx_srcs = [self.uni_cplx_src, self.norm_cplx_src]
+ self.assertAlmostEqual(min(data), -1, places=4)
+ self.assertAlmostEqual(max(data), 1, places=4)
- paths = {}
- for src in real_srcs:
- head = blocks.head(nitems=NUM_ITEMS, sizeof_stream_item=gr.sizeof_float)
- sink = blocks.vector_sink_f()
- paths[src] = (src, head, sink)
+ # mean, variance
+ self.assertAlmostEqual(data.mean(), 0, places=2)
+ self.assertAlmostEqual(data.var(), (1-(-1))**2./12, places=3)
- for src in cplx_srcs:
- head = blocks.head(nitems=NUM_ITEMS, sizeof_stream_item=gr.sizeof_gr_complex)
- sink = blocks.vector_sink_c()
- paths[src] = (src, head, sink)
- for path in paths.itervalues():
- tb.connect(*path)
+ def test_001_real_gaussian_moments(self):
+ data = self.run_test_real(analog.GR_GAUSSIAN)
- tb.run()
- self.data = {src: path[2].data() for src, path in paths.iteritems()}
- data = self.data
- # min, max
- self.assertAlmostEqual(min(data[self.uni_real_src]), -1, places=4)
- self.assertAlmostEqual(max(data[self.uni_real_src]), 1, places=4)
# mean, variance
- data = numpy.array(data[self.uni_real_src])
- self.assertAlmostEqual(data.mean(), 0, places = 2)
- self.assertAlmostEqual(data.var(), (1-(-1))**2./12, places = 3)
- data = numpy.array(self.data[self.norm_real_src])
self.assertAlmostEqual(data.mean(), 0, places=2)
self.assertAlmostEqual(data.var(), 1, places=2)
- data = numpy.array(self.data[self.lapl_real_src])
+
+ def test_001_real_laplacian_moments(self):
+ data = self.run_test_real(analog.GR_LAPLACIAN)
+
+ # mean, variance
self.assertAlmostEqual(data.mean(), 0, places=2)
self.assertAlmostEqual(data.var(), 2, places=2)
+ def test_001_complex_uniform_moments(self):
+ data = self.run_test_complex(analog.GR_UNIFORM)
+
+ # mean, variance
+ self.assertAlmostEqual(data.real.mean(), 0, places=2)
+ self.assertAlmostEqual(data.real.var(), 0.5*(1-(-1))**2./12, places=3)
+
+ self.assertAlmostEqual(data.imag.mean(), 0, places=2)
+ self.assertAlmostEqual(data.imag.var(), 0.5*(1-(-1))**2./12, places=3)
+
+ def test_001_complex_gaussian_moments(self):
+ data = self.run_test_complex(analog.GR_GAUSSIAN)
+
+ # mean, variance
+ self.assertAlmostEqual(data.real.mean(), 0, places=2)
+ self.assertAlmostEqual(data.real.var(), 0.5, places=2)
+
+ self.assertAlmostEqual(data.imag.mean(), 0, places=2)
+ self.assertAlmostEqual(data.imag.var(), 0.5, places=2)
+
+ def test_002_reproducibility(self):
+ data1 = self.run_test_real(analog.GR_UNIFORM)
+ data2 = self.run_test_real(analog.GR_UNIFORM)
+
+ # It's pseudoramdo thus must be equal
+ self.assertTrue(numpy.array_equal(data1, data2))
+
+ data1 = self.run_test_real(analog.GR_GAUSSIAN)
+ data2 = self.run_test_real(analog.GR_GAUSSIAN)
+
+ self.assertTrue(numpy.array_equal(data1, data2))
+
if __name__ == '__main__':
gr_unittest.run(test_fastnoise_source, "test_fastnoise_source.xml")
-
--
cgit v1.2.3