def annotation_to_fbs_matrix(self, axis, fields=None, labels=None):
with ServerTiming.time(f"annotations.{axis}.query"):
A = self.open_array(str(axis))
# may raise if fields contains unknown key
cxg_fields, anno_fields, return_fields, index_field = self._annotations_field_split(
axis, fields, A, labels)
if cxg_fields is None:
data = A[:]
elif cxg_fields:
data = A.query(attrs=cxg_fields)[:]
else:
data = {}
df = pd.DataFrame.from_dict(data)
if axis == Axis.OBS and labels is not None and not labels.empty:
if anno_fields is None:
assert index_field
df = df.join(labels, index_field)
elif anno_fields:
assert index_field
df = df.join(labels[anno_fields], index_field)
if return_fields:
df = df[return_fields]
with ServerTiming.time(f"annotations.{axis}.encode"):
fbs = encode_matrix_fbs(df, col_idx=df.columns)
return fbs
def layout_to_fbs_matrix(self, fields):
"""
return specified embeddings as a flatbuffer, using the cellxgene matrix fbs encoding.
* returns only first two dimensions, with name {ename}_0 and {ename}_1,
where {ename} is the embedding name.
* client assumes each will be individually centered & scaled (isotropically)
to a [0, 1] range.
* does not support filtering
"""
embeddings = self.get_embedding_names() if fields is None or len(fields) == 0 else fields
layout_data = []
with ServerTiming.time("layout.query"):
for ename in embeddings:
embedding = self.get_embedding_array(ename, 2)
normalized_layout = DataAdaptor.normalize_embedding(embedding)
layout_data.append(pd.DataFrame(normalized_layout, columns=[f"{ename}_0", f"{ename}_1"]))
with ServerTiming.time("layout.encode"):
if layout_data:
df = pd.concat(layout_data, axis=1, copy=False)
else:
df = pd.DataFrame()
fbs = encode_matrix_fbs(df, col_idx=df.columns, row_idx=None)
return fbs
def get_embedding_names(self):
with ServerTiming.time("layout.lsuri"):
pemb = self.get_path("emb")
embeddings = [os.path.basename(p) for (p, t) in self.lsuri(pemb) if t == "array"]
if len(embeddings) == 0:
raise DatasetAccessError("cxg matrix missing embeddings")
return embeddings
def compute_embedding(self, method, obsFilter):
if Axis.VAR in obsFilter:
raise FilterError(
"Observation filters may not contain variable conditions")
if method != "umap":
raise NotImplementedError(
f"re-embedding method {method} is not available.")
try:
shape = self.get_shape()
obs_mask = self._axis_filter_to_mask(Axis.OBS, obsFilter["obs"],
shape[0])
except (KeyError, IndexError):
raise FilterError("Error parsing filter")
with ServerTiming.time("layout.compute"):
X_umap = scanpy_umap(self.data, obs_mask)
normalized_layout = DataAdaptor.normalize_embedding(X_umap)
# Server picks reemedding name, which must not collide with any other
# embedding name generated by this backed.
name = f"reembed:{method}_{datetime.now().isoformat(timespec='milliseconds')}"
dims = [f"{name}_0", f"{name}_1"]
df = pd.DataFrame(normalized_layout, columns=dims)
fbs = encode_matrix_fbs(df, col_idx=df.columns, row_idx=None)
schema = {"name": name, "type": "float32", "dims": dims}
return (schema, fbs)
def compute_embedding(self, method, obsFilter):
if Axis.VAR in obsFilter:
raise FilterError("Observation filters may not contain variable conditions")
if method != "umap":
raise NotImplementedError(f"re-embedding method {method} is not available.")
try:
shape = self.get_shape()
obs_mask = self._axis_filter_to_mask(Axis.OBS, obsFilter["obs"], shape[0])
except (KeyError, IndexError):
raise FilterError("Error parsing filter")
with ServerTiming.time("layout.compute"):
X_umap = scanpy_umap(self.data, obs_mask)
# Server picks reemedding name, which must not collide with any other
# embedding name generated by this backend.
name = f"reembed:{method}_{datetime.now().isoformat(timespec='milliseconds')}"
dims = [f"{name}_0", f"{name}_1"]
layout_schema = {"name": name, "type": "float32", "dims": dims}
self.schema["layout"]["obs"].append(layout_schema)
self.data.obsm[f"X_{name}"] = X_umap
return layout_schema