def test_traversal_path():
da = DocumentArray([Document() for _ in range(6)])
assert len(da) == 6
da.traverse_flat(['r'])
with pytest.raises(ValueError):
da.traverse_flat('r')
da.traverse(['r'])
with pytest.raises(ValueError):
for _ in da.traverse('r'):
pass
da.traverse(['r'])
with pytest.raises(ValueError):
for _ in da.traverse('r'):
pass
def test_traversal_path():
da = DocumentArray([Document() for _ in range(6)])
assert len(da) == 6
da.traverse_flat('r')
with pytest.warns(DeprecationWarning):
da.traverse_flat(['r'])
da.traverse('r')
with pytest.warns(DeprecationWarning):
for _ in da.traverse(['r']):
pass
da.traverse('r')
with pytest.warns(DeprecationWarning):
for _ in da.traverse(['r']):
pass
def craft(self, docs: DocumentArray, **kwargs):
filtered_docs = DocumentArray(
d for d in docs.traverse_flat(['c']) if d.mime_type == 'image/jpeg'
)
target_size = 224
for doc in filtered_docs:
doc.load_uri_to_image_blob()
doc.set_image_blob_shape(shape=(target_size, target_size))
doc.set_image_blob_channel_axis(-1, 0)
return filtered_docs
def test_match_inclusive(only_id):
"""Call match function, while the other :class:`DocumentArray` is itself
or have same :class:`Document`.
"""
# The document array da1 match with itself.
da1 = DocumentArray([
Document(embedding=np.array([1, 2, 3])),
Document(embedding=np.array([1, 0, 1])),
Document(embedding=np.array([1, 1, 2])),
])
da1.match(da1, only_id=only_id)
assert len(da1) == 3
traversed = da1.traverse_flat(traversal_paths=['m', 'mm', 'mmm'])
assert len(traversed) == 9
# The document array da2 shares same documents with da1
da2 = DocumentArray(
[Document(embedding=np.array([4, 1, 3])), da1[0], da1[1]])
da1.match(da2, only_id=only_id)
assert len(da2) == 3
traversed = da1.traverse_flat(traversal_paths=['m', 'mm', 'mmm'])
assert len(traversed) == 9
def craft(self, docs: DocumentArray, fn):
filtered_docs = DocumentArray(d for d in docs.traverse_flat(['c'])
if d.mime_type == 'image/jpeg')
for doc in filtered_docs:
getattr(doc, fn)()
raw_img = _load_image(doc.blob, self.channel_axis)
img = self._normalize(raw_img)
# move the channel_axis to target_channel_axis to better fit different models
if self.channel_axis != self.target_channel_axis:
img = np.moveaxis(img, self.channel_axis,
self.target_channel_axis)
doc.blob = img
return filtered_docs
def filter(self, docs: DocumentArray, **kwargs):
filtered_docs = DocumentArray(d for d in docs.traverse_flat(['c'])
if d.mime_type == 'text/plain')
return filtered_docs
def test_traverse_flat_root_itself():
da = DocumentArray([Document() for _ in range(100)])
res = da.traverse_flat('r')
assert id(res) == id(da)
