def build_nodes(hetio: Dict, **kwargs) -> List[Node]:
force_rebuild = kwargs.get("force_rebuild", False)
save_checkpoint = kwargs.get("save_checkpoint", True)
if not force_rebuild:
if os.path.exists(NODES_CHECKPOINT):
return Node.deserialize_bunch(NODES_CHECKPOINT)
else:
log.info("Node checkpoint does not exist, building nodes.")
nodes = [
Node(h["identifier"], h["name"], h["kind"],
h["data"].get("source", None) or h["data"].get("sources", []),
h["data"].get("license", None), h["data"].get("url", None))
for h in hetio["nodes"]
]
assert len(nodes) == len(hetio["nodes"])
umls = kwargs.get("umls", None)
do = kwargs.get("do", None)
if umls is not None:
nodes = add_compound_metadata(hetio, nodes, umls)
if do is not None:
nodes = add_disease_metadata(hetio, nodes, do)
if save_checkpoint:
log.info("Checkpointing nodes...")
Node.serialize_bunch(nodes, NODES_CHECKPOINT)
return nodes
def build_nodes(repodb: pd.DataFrame, **kwargs) -> List[Node]:
force_rebuild = kwargs.get("force_rebuild", False)
save_checkpoint = kwargs.get("save_checkpoint", True)
if not force_rebuild:
if os.path.exists(NODES_CHECKPOINT):
return Node.deserialize_bunch(NODES_CHECKPOINT)
else:
log.info("Node checkpoint does not exist, building nodes.")
nodes = []
drug_ids = repodb["drug_id"].unique()
disease_ids = repodb["ind_id"].unique()
for drug_id in drug_ids:
match = repodb[repodb["drug_id"] == drug_id]
drug_names = match["drug_name"].unique()
assert len(drug_names) == 1
node = Node(drug_id,
drug_names[0],
kind="Compound",
sources=["RepoDB"])
nodes.append(node)
for disease_id in disease_ids:
match = repodb[repodb["ind_id"] == disease_id]
disease_names = match["ind_name"].unique()
assert len(disease_names) == 1
node = Node(disease_id,
disease_names[0],
kind="Disease",
sources=["RepoDB"])
nodes.append(node)
log.info(f"Built {len(nodes)} nodes from RepoDB.")
if save_checkpoint:
log.info("Checkpointing nodes...")
Node.serialize_bunch(nodes, NODES_CHECKPOINT)
return nodes
