def test_no_custom_modules(self, deployable_entity, model_for_deployment):
deployable_entity.log_model(model_for_deployment['model'])
custom_module_filenames = {"__init__.py", "_verta_config.py"}
for path in sys.path:
# skip std libs and venvs
# This logic is from verta.client._log_modules().
lib_python_str = os.path.join(os.sep, "lib", "python")
i = path.find(lib_python_str)
if i != -1 and glob.glob(os.path.join(path[:i], "bin", "python*")):
continue
for parent_dir, dirnames, filenames in os.walk(path):
# only Python files
filenames[:] = [
filename for filename in filenames
if filename.endswith(('.py', '.pyc', '.pyo'))
]
if not _utils.is_in_venv(path) and _utils.is_in_venv(
parent_dir):
continue
custom_module_filenames.update(map(os.path.basename,
filenames))
custom_modules = deployable_entity.get_artifact(
_artifact_utils.CUSTOM_MODULES_KEY)
with zipfile.ZipFile(custom_modules, 'r') as zipf:
assert custom_module_filenames == set(
map(os.path.basename, zipf.namelist()))
def test_log_model(self, deployable_entity):
np = pytest.importorskip("numpy")
sklearn = pytest.importorskip("sklearn")
from sklearn.linear_model import LogisticRegression
classifier = LogisticRegression()
classifier.fit(np.random.random((36, 12)), np.random.random(36).round())
original_coef = classifier.coef_
deployable_entity.log_model(classifier)
# retrieve the classifier:
retrieved_classfier = deployable_entity.get_model()
assert np.array_equal(retrieved_classfier.coef_, original_coef)
# check model api:
assert _artifact_utils.MODEL_API_KEY in deployable_entity.get_artifact_keys()
for artifact in deployable_entity._msg.artifacts:
if artifact.key == _artifact_utils.MODEL_API_KEY:
assert artifact.filename_extension == "json"
# overwrite should work:
new_classifier = LogisticRegression()
new_classifier.fit(np.random.random((36, 12)), np.random.random(36).round())
deployable_entity.log_model(new_classifier, overwrite=True)
retrieved_classfier = deployable_entity.get_model()
assert np.array_equal(retrieved_classfier.coef_, new_classifier.coef_)
# when overwrite = false, overwriting should fail
with pytest.raises(ValueError) as excinfo:
deployable_entity.log_model(new_classifier)
assert "already exists" in str(excinfo.value)
# Check custom modules:
custom_module_filenames = {"__init__.py", "_verta_config.py"}
for path in sys.path:
# skip std libs and venvs
# This logic is from verta.client._log_modules().
lib_python_str = os.path.join(os.sep, "lib", "python")
i = path.find(lib_python_str)
if i != -1 and glob.glob(os.path.join(path[:i], "bin", "python*")):
continue
for parent_dir, dirnames, filenames in os.walk(path):
# only Python files
filenames[:] = [
filename
for filename in filenames
if filename.endswith((".py", ".pyc", ".pyo"))
]
if not _utils.is_in_venv(path) and _utils.is_in_venv(parent_dir):
continue
custom_module_filenames.update(map(os.path.basename, filenames))
custom_modules = deployable_entity.get_artifact(_artifact_utils.CUSTOM_MODULES_KEY)
with zipfile.ZipFile(custom_modules, "r") as zipf:
assert custom_module_filenames == set(
map(os.path.basename, zipf.namelist())
)
def _custom_modules_as_artifact(self, paths=None):
if isinstance(paths, six.string_types):
paths = [paths]
# If we include a path that is actually a module, then we _must_ add its parent to the
# adjusted sys.path in the end so that we can re-import with the same name.
forced_local_sys_paths = []
if paths is not None:
new_paths = []
for p in paths:
abspath = os.path.abspath(os.path.expanduser(p))
if os.path.exists(abspath):
new_paths.append(abspath)
else:
try:
mod = importlib.import_module(p)
new_paths.extend(mod.__path__)
forced_local_sys_paths.extend(map(os.path.dirname, mod.__path__))
except ImportError:
raise ValueError("custom module {} does not correspond to an existing folder or module".format(p))
paths = new_paths
forced_local_sys_paths = sorted(list(set(forced_local_sys_paths)))
# collect local sys paths
local_sys_paths = copy.copy(sys.path)
## replace empty first element with cwd
## https://docs.python.org/3/library/sys.html#sys.path
if local_sys_paths[0] == "":
local_sys_paths[0] = os.getcwd()
## convert to absolute paths
local_sys_paths = list(map(os.path.abspath, local_sys_paths))
## remove paths that don't exist
local_sys_paths = list(filter(os.path.exists, local_sys_paths))
## remove .ipython
local_sys_paths = list(filter(lambda path: not path.endswith(".ipython"), local_sys_paths))
## remove virtual (and real) environments
local_sys_paths = list(filter(lambda path: not _utils.is_in_venv(path), local_sys_paths))
# get paths to files within
if paths is None:
# Python files within filtered sys.path dirs
paths = local_sys_paths
extensions = ['py', 'pyc', 'pyo']
else:
# all user-specified files
paths = paths
extensions = None
local_filepaths = _utils.find_filepaths(
paths, extensions=extensions,
include_hidden=True,
include_venv=False, # ignore virtual environments nested within
)
## remove .git
local_filepaths = set(filter(lambda path: not path.endswith(".git") and ".git/" not in path,
local_filepaths))
# obtain deepest common directory
# This directory on the local system will be mirrored in `_CUSTOM_MODULES_DIR` in
# deployment.
curr_dir = os.path.join(os.getcwd(), "")
paths_plus = list(local_filepaths) + [curr_dir]
common_prefix = os.path.commonprefix(paths_plus)
common_dir = os.path.dirname(common_prefix)
# replace `common_dir` with `_CUSTOM_MODULES_DIR` for deployment sys.path
depl_sys_paths = list(map(lambda path: os.path.relpath(path, common_dir), local_sys_paths + forced_local_sys_paths))
depl_sys_paths = list(map(lambda path: os.path.join(_CUSTOM_MODULES_DIR, path), depl_sys_paths))
bytestream = six.BytesIO()
with zipfile.ZipFile(bytestream, 'w') as zipf:
for filepath in local_filepaths:
arcname = os.path.relpath(filepath, common_dir) # filepath relative to archive root
try:
zipf.write(filepath, arcname)
except:
# maybe file has corrupt metadata; try reading then writing contents
with open(filepath, 'rb') as f:
zipf.writestr(
_artifact_utils.global_read_zipinfo(arcname),
f.read(),
)
# add verta config file for sys.path and chdir
working_dir = os.path.join(_CUSTOM_MODULES_DIR, os.path.relpath(curr_dir, common_dir))
zipf.writestr(
_artifact_utils.global_read_zipinfo("_verta_config.py"),
six.ensure_binary('\n'.join([
"import os, sys",
"",
"",
"sys.path = sys.path[:1] + {} + sys.path[1:]".format(depl_sys_paths),
"",
"try:",
" os.makedirs(\"{}\")".format(working_dir),
"except OSError: # already exists",
" pass",
"os.chdir(\"{}\")".format(working_dir),
]))
)
# add __init__.py
init_filename = "__init__.py"
if init_filename not in zipf.namelist():
zipf.writestr(
_artifact_utils.global_read_zipinfo(init_filename),
b"",
)
bytestream.seek(0)
return bytestream
def test_is_not_in_venv(self):
# NOTE: assumes there isn't a venv in the client test dir
for root, _, files in os.walk("."):
for filename in files:
abs_filepath = os.path.abspath(os.path.join(root, filename))
assert not _utils.is_in_venv(abs_filepath)
def test_is_in_venv(self, prefix, filepath, in_fake_venv):
abs_filepath = os.path.abspath(os.path.join(prefix, filepath))
assert _utils.is_in_venv(abs_filepath)