tests.data.inference.test_inference.TestInference.try_inference() - Code Metrics - Inspection of "Refactor inference classes" - fepegar/torchio - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Pull Request — master (#182)

by Fernando

created 2020-06-07 16:49 UTC

TestInference.try_inference() A

↳ Parent: tests.data.inference.test_inference

Complexity

Conditions

Size

Total Lines	25
Code Lines	22

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
cc	4
eloc	22
nop	2
dl	0
loc	25
rs	9.352
c	0
b	0
f	0

import torch
from torch.utils.data import DataLoader
from tqdm import tqdm
from torchio import LOCATION, DATA
from torchio.data.inference import GridSampler, GridAggregator
from ...utils import TorchioTestCase


class TestInference(TorchioTestCase):
    """Tests for `inference` module."""
    def test_inference_no_padding(self):
        self.try_inference(None)

    def test_inference_padding(self):
        self.try_inference(3)

    def try_inference(self, padding_mode):
        for n in 17, 27:
            patch_size = 10, 15, n
            patch_overlap = 4, 6, 8
            batch_size = 6

            grid_sampler = GridSampler(
                self.sample,
                patch_size,
                patch_overlap,
                padding_mode=padding_mode,
            )
            aggregator = GridAggregator(grid_sampler)
            patch_loader = DataLoader(grid_sampler, batch_size=batch_size)
            with torch.no_grad():
                for patches_batch in tqdm(patch_loader):
                    input_tensor = patches_batch['t1'][DATA]
                    locations = patches_batch[LOCATION]
                    logits = model(input_tensor)  # some model
                    outputs = logits
                    aggregator.add_batch(outputs, locations)

            output = aggregator.get_output_tensor()
            assert (output == -5).all()
            assert output.shape == self.sample.t1.shape


def model(tensor):
    tensor[:] = -5
    return tensor


1			import torch
2			from torch.utils.data import DataLoader
3			from tqdm import tqdm
4			from torchio import LOCATION, DATA
5			from torchio.data.inference import GridSampler, GridAggregator
6			from ...utils import TorchioTestCase
7
8
9			class TestInference(TorchioTestCase):
10			"""Tests for `inference` module."""
11			def test_inference_no_padding(self):
12			self.try_inference(None)
13
14			def test_inference_padding(self):
15			self.try_inference(3)
16
17			def try_inference(self, padding_mode):
18			for n in 17, 27:
19			patch_size = 10, 15, n
20			patch_overlap = 4, 6, 8
21			batch_size = 6
22
23			grid_sampler = GridSampler(
24			self.sample,
25			patch_size,
26			patch_overlap,
27			padding_mode=padding_mode,
28			)
29			aggregator = GridAggregator(grid_sampler)
30			patch_loader = DataLoader(grid_sampler, batch_size=batch_size)
31			with torch.no_grad():
32			for patches_batch in tqdm(patch_loader):
33			input_tensor = patches_batch['t1'][DATA]
34			locations = patches_batch[LOCATION]
35			logits = model(input_tensor) # some model
36			outputs = logits
37			aggregator.add_batch(outputs, locations)
38
39			output = aggregator.get_output_tensor()
40			assert (output == -5).all()
41			assert output.shape == self.sample.t1.shape
42
43
44			def model(tensor):
45			tensor[:] = -5
46			return tensor
47

fepegar / torchio

Pull Request — master (#182)

TestInference.try_inference() A

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like