Trying to add word embedding to my nodes using Unwind.
The code:
result = conn.query("MATCH (m:Word) WHERE NOT exists(m.embedding) RETURN m")
nodes = [x['m'] for x in result]
for dic in nodes:
dic['embedding'] = list(np.round(model_st.encode(dic['name'], show_progress_bar=False), 3))
query_update = f""" UNWIND $nodes as res_dict
MATCH (n:Word {{name: res_dict.name}})
SET n.embedding = res_dict.embedding
"""
self.conn.query(query=query_update, parameters={'nodes': nodes})
But when I try to run it, I get the following error:
Parameters of type float32 are not supported
How can I achieve it in Python? Is there a better way to set an embedding to a node?
I was able to add embedding using the following code
embedding = list(np.round(model_st.encode(name, show_progress_bar=False).astype(np.float64),3))
I still think it's weird neo4j don't have a native way to create embedding (using apoc/gds)