def test_args_pretty(testapp, pretty, lines):
"""Sitemap processor has to respect pretty argument."""
timepoint = datetime.datetime.fromtimestamp(0, tz=datetime.timezone.utc)
stream = sitemap.process(
testapp,
[
holocron.WebSiteItem({
"destination": pathlib.Path("1.html"),
"updated": timepoint,
"baseurl": testapp.metadata["url"],
})
],
pretty=pretty,
)
assert isinstance(stream, collections.abc.Iterable)
items = list(stream)
assert items == [
holocron.WebSiteItem({
"destination": pathlib.Path("1.html"),
"updated": timepoint,
"baseurl": testapp.metadata["url"],
}),
holocron.WebSiteItem({
"source": pathlib.Path("sitemap://sitemap.xml"),
"destination": pathlib.Path("sitemap.xml"),
"content": unittest.mock.ANY,
"baseurl": testapp.metadata["url"],
}),
]
assert len(items[-1]["content"].splitlines()) == lines
def test_args_save_as_unsupported(testapp, document_path, sitemap_path):
"""Sitemap process has to check enlisted URLs for compatibility."""
timepoint = datetime.datetime.fromtimestamp(0, tz=datetime.timezone.utc)
stream = sitemap.process(
testapp,
[
holocron.WebSiteItem({
"destination": document_path,
"updated": timepoint,
"baseurl": testapp.metadata["url"],
})
],
save_as=str(sitemap_path),
)
assert isinstance(stream, collections.abc.Iterable)
with pytest.raises(ValueError) as excinfo:
next(stream)
excinfo.match(
"The location of a Sitemap file determines the set of URLs "
"that can be included in that Sitemap. A Sitemap file located "
"at .* can include any URLs starting with .* but can not "
"include .*.")
def test_args_save_as(testapp, save_as):
"""Sitemap processor has to respect save_as argument."""
timepoint = datetime.datetime.fromtimestamp(0, tz=datetime.timezone.utc)
stream = sitemap.process(
testapp,
[
holocron.WebSiteItem({
"destination": pathlib.Path("posts", "1.html"),
"updated": timepoint,
"baseurl": testapp.metadata["url"],
})
],
save_as=str(save_as),
)
assert isinstance(stream, collections.abc.Iterable)
assert list(stream) == [
holocron.WebSiteItem({
"destination": pathlib.Path("posts", "1.html"),
"updated": timepoint,
"baseurl": testapp.metadata["url"],
}),
holocron.WebSiteItem({
"source": pathlib.Path("sitemap://", save_as),
"destination": pathlib.Path(save_as),
"content": unittest.mock.ANY,
"baseurl": testapp.metadata["url"],
}),
]
def test_item_many(testapp, amount):
"""Sitemap processor has to work with stream."""
timepoint = datetime.datetime.fromtimestamp(0, tz=datetime.timezone.utc)
stream = sitemap.process(
testapp,
[
holocron.WebSiteItem({
"destination": pathlib.Path(str(i)),
"updated": timepoint,
"baseurl": testapp.metadata["url"],
}) for i in range(amount)
],
)
assert isinstance(stream, collections.abc.Iterable)
assert list(stream) == list(
itertools.chain(
[
holocron.WebSiteItem({
"destination": pathlib.Path(str(i)),
"updated": timepoint,
"baseurl": testapp.metadata["url"],
}) for i in range(amount)
],
[
holocron.WebSiteItem({
"source":
pathlib.Path("sitemap://sitemap.xml"),
"destination":
pathlib.Path("sitemap.xml"),
"content":
_pytest_xmlasdict(
{
"urlset": {
"@xmlns":
"http://www.sitemaps.org/schemas/sitemap/0.9",
"url": [{
"loc": "https://yoda.ua/%d" % i,
"lastmod": "1970-01-01T00:00:00+00:00",
} for i in range(amount)],
}
},
force_list=["url"],
),
"baseurl":
testapp.metadata["url"],
})
],
))
def test_args_gzip(testapp):
"""Sitemap processor has to respect gzip argument."""
timepoint = datetime.datetime.fromtimestamp(0, tz=datetime.timezone.utc)
stream = sitemap.process(
testapp,
[
holocron.WebSiteItem({
"destination": pathlib.Path("1.html"),
"updated": timepoint,
"baseurl": testapp.metadata["url"],
})
],
gzip=True,
)
assert isinstance(stream, collections.abc.Iterable)
assert list(stream) == [
holocron.WebSiteItem({
"destination": pathlib.Path("1.html"),
"updated": timepoint,
"baseurl": testapp.metadata["url"],
}),
holocron.WebSiteItem({
"source":
pathlib.Path("sitemap://sitemap.xml.gz"),
"destination":
pathlib.Path("sitemap.xml.gz"),
"content":
_pytest_xmlasdict(
{
"urlset": {
"@xmlns":
"http://www.sitemaps.org/schemas/sitemap/0.9",
"url": {
"loc": "https://yoda.ua/1.html",
"lastmod": "1970-01-01T00:00:00+00:00",
},
}
},
ungzip=True,
),
"baseurl":
testapp.metadata["url"],
}),
]
def test_item(testapp, filename, escaped):
"""Sitemap processor has to work!"""
timepoint = datetime.datetime.fromtimestamp(0, tz=datetime.timezone.utc)
stream = sitemap.process(
testapp,
[
holocron.WebSiteItem({
"destination": pathlib.Path(filename),
"updated": timepoint,
"baseurl": testapp.metadata["url"],
})
],
)
assert isinstance(stream, collections.abc.Iterable)
assert list(stream) == [
holocron.WebSiteItem({
"destination": pathlib.Path(filename),
"updated": timepoint,
"baseurl": testapp.metadata["url"],
}),
holocron.WebSiteItem({
"source":
pathlib.Path("sitemap://sitemap.xml"),
"destination":
pathlib.Path("sitemap.xml"),
"content":
_pytest_xmlasdict({
"urlset": {
"@xmlns": "http://www.sitemaps.org/schemas/sitemap/0.9",
"url": {
"loc": "https://yoda.ua/" + escaped,
"lastmod": "1970-01-01T00:00:00+00:00",
},
}
}),
"baseurl":
testapp.metadata["url"],
}),
]
def test_item_many_zero(testapp):
"""Sitemap processor has to work with stream of zero items."""
stream = sitemap.process(testapp, [])
assert isinstance(stream, collections.abc.Iterable)
assert list(stream) == [
holocron.WebSiteItem({
"source":
pathlib.Path("sitemap://sitemap.xml"),
"destination":
pathlib.Path("sitemap.xml"),
"content":
_pytest_xmlasdict({
"urlset": {
"@xmlns": "http://www.sitemaps.org/schemas/sitemap/0.9"
}
}),
"baseurl":
testapp.metadata["url"],
})
]
def test_args_bad_value(testapp, args, error):
"""Sitemap processor has to validate input arguments."""
with pytest.raises(ValueError) as excinfo:
next(sitemap.process(testapp, [], **args))
assert str(excinfo.value) == error