mirror of
https://github.com/KichangKim/DeepDanbooru.git
synced 2026-10-18 14:27:07 +08:00
rachmadaniHaryono
80 lines
2.3 KiB
Python
80 lines
2.3 KiB
Python
#!/usr/bin/env python
|
|
# -*- coding: utf-8 -*-
|
|
import importlib
|
|
from unittest import mock
|
|
|
|
import numpy
|
|
import pytest
|
|
import six
|
|
from click.testing import CliRunner
|
|
from PIL import Image
|
|
|
|
|
|
def test_import():
|
|
import deepdanbooru
|
|
|
|
|
|
@pytest.mark.parametrize('func_name', ['main', 'evaluate'])
|
|
def test_help(func_name):
|
|
mod = importlib.import_module('deepdanbooru.__main__')
|
|
runner = CliRunner()
|
|
result = runner.invoke(getattr(mod, func_name), ['--help'])
|
|
assert result.exit_code == 0
|
|
assert result.output
|
|
|
|
|
|
@pytest.fixture
|
|
def packages():
|
|
with open('requirements.txt') as f:
|
|
return f.read().splitlines()
|
|
|
|
|
|
def test_package_setup(packages):
|
|
import setuptools
|
|
with mock.patch('setuptools.setup'):
|
|
import setup
|
|
setup_pkgs = setup.install_requires
|
|
tensorflow_pkg = setup.tensorflow_pkg
|
|
assert setup_pkgs == list(
|
|
filter(lambda x: not x.startswith('tensorflow'), packages))
|
|
assert list(
|
|
filter(lambda x: x.startswith('tensorflow'), packages)
|
|
) == [tensorflow_pkg]
|
|
|
|
|
|
def test_readme_pkg(packages):
|
|
with open('README.md') as f:
|
|
text = f.read()
|
|
line_start = 'Following packages are need to be installed.'
|
|
line_end = '\n\n'
|
|
pkg_text = text.split(line_start)[1].rsplit(line_end)[0].strip()
|
|
readme_pkgs = list(map(lambda x: x.split(' ', 1)[1], pkg_text.splitlines()))
|
|
|
|
assert list(
|
|
sorted(filter(lambda x: not x.startswith('tensorflow'), readme_pkgs))
|
|
) == list(
|
|
sorted(filter(lambda x: not x.startswith('tensorflow'), packages))
|
|
)
|
|
|
|
assert list(
|
|
filter(lambda x: x.startswith('tensorflow'), packages)
|
|
) == list(
|
|
filter(lambda x: x.startswith('tensorflow'), readme_pkgs)
|
|
)
|
|
|
|
|
|
@pytest.mark.parametrize('use_bytes_io', [True, False])
|
|
def test_load_image_for_evaluate(tmp_path, use_bytes_io):
|
|
image_path = tmp_path / 'test.jpg'
|
|
image = Image.new('RGB', (300, 300), color='red')
|
|
image.save(image_path)
|
|
from deepdanbooru.data import load_image_for_evaluate
|
|
if not use_bytes_io:
|
|
res = load_image_for_evaluate(image_path.as_posix(), 299, 299)
|
|
else:
|
|
with open(image_path.as_posix(), 'rb') as f:
|
|
image_input = six.BytesIO(f.read())
|
|
res = load_image_for_evaluate(image_input, 299, 299)
|
|
assert isinstance(res, numpy.ndarray)
|
|
assert res.shape == (299, 299, 3)
|