def test_example(self, example_name):
actual_perf = examples.run_example(
EXAMPLE_MODULES,
example_name,
# This should match the invocation in
# zipline/tests/resources/rebuild_example_data
environ={
'ZIPLINE_ROOT': self.tmpdir.getpath('example_data/root'),
},
benchmark_returns=self.benchmark_returns,
)
expected_perf = self.expected_perf[example_name]
# Exclude positions column as the positions do not always have the
# same order
columns = [
column for column in examples._cols_to_check
if column != 'positions'
]
assert_equal(
actual_perf[columns],
expected_perf[columns],
# There is a difference in the datetime columns in pandas
# 0.16 and 0.17 because in 16 they are object and in 17 they are
# datetime[ns, UTC]. We will just ignore the dtypes for now.
check_dtype=False,
)
# Sort positions by SID before comparing
assert_equal(
expected_perf['positions'].apply(sorted, key=itemgetter('sid')),
actual_perf['positions'].apply(sorted, key=itemgetter('sid')),
)
def test_example(self, example_name):
actual_perf = examples.run_example(
example_name,
# This should match the invocation in
# zipline/tests/resources/rebuild_example_data
environ={"ZIPLINE_ROOT": self.tmpdir.getpath("example_data/root")},
)
assert_equal(
actual_perf[examples._cols_to_check],
self.expected_perf[example_name][examples._cols_to_check],
# There is a difference in the datetime columns in pandas
# 0.16 and 0.17 because in 16 they are object and in 17 they are
# datetime[ns, UTC]. We will just ignore the dtypes for now.
check_dtype=False,
)
def test_example(self, example_name):
actual_perf = examples.run_example(
example_name,
# This should match the invocation in
# zipline/tests/resources/rebuild_example_data
environ={
'ZIPLINE_ROOT': self.tmpdir.getpath('example_data/root'),
},
)
assert_equal(
actual_perf[examples._cols_to_check],
self.expected_perf[example_name][examples._cols_to_check],
# There is a difference in the datetime columns in pandas
# 0.16 and 0.17 because in 16 they are object and in 17 they are
# datetime[ns, UTC]. We will just ignore the dtypes for now.
check_dtype=False,
)
def main(ctx, rebuild_input):
"""Rebuild the perf data for test_examples"""
example_path = test_resource_path("example_data.tar.gz")
with tmp_dir() as d:
with tarfile.open(example_path) as tar:
tar.extractall(d.path)
# The environ here should be the same (modulo the tempdir location)
# as we use in test_examples.py.
environ = {"ZIPLINE_ROOT": d.getpath("example_data/root")}
if rebuild_input:
raise NotImplementedError(
"We cannot rebuild input for Yahoo because of "
"changes Yahoo made to their API, so we cannot "
"use Yahoo data bundles anymore. This will be fixed in "
"a future release",
)
# we need to register the bundle; it is already ingested and saved in
# the example_data.tar.gz file
@register("test")
def nop_ingest(*args, **kwargs):
raise NotImplementedError("we cannot rebuild the test buindle")
new_perf_path = d.getpath(
"example_data/new_perf/%s" % pd.__version__.replace(".", "-"),
)
c = dataframe_cache(
new_perf_path,
serialization="pickle:2",
)
with c:
for name in EXAMPLE_MODULES:
c[name] = examples.run_example(
EXAMPLE_MODULES,
name,
environ=environ,
benchmark_returns=read_checked_in_benchmark_data(),
)
correct_called = [False]
console = None
def _exit(*args, **kwargs):
console.raw_input = eof
def correct():
correct_called[0] = True
_exit()
expected_perf_path = d.getpath(
"example_data/expected_perf/%s"
% pd.__version__.replace(".", "-"),
)
# allow users to run some analysis to make sure that the new
# results check out
console = InteractiveConsole(
{
"correct": correct,
"exit": _exit,
"incorrect": _exit,
"new": c,
"np": np,
"old": dataframe_cache(
expected_perf_path,
serialization="pickle",
),
"pd": pd,
"cols_to_check": examples._cols_to_check,
"changed_results": changed_results,
}
)
console.interact(banner)
if not correct_called[0]:
ctx.fail(
"`correct()` was not called! This means that the new"
" results will not be written",
)
# move the new results to the expected path
shutil.rmtree(expected_perf_path)
shutil.copytree(new_perf_path, expected_perf_path)
# Clear out all the temporary new perf so it doesn't get added to the
# tarball.
shutil.rmtree(d.getpath("example_data/new_perf/"))
with tarfile.open(example_path, "w|gz") as tar:
tar.add(d.getpath("example_data"), "example_data")
