[Question]: Could not create Property Graph Index with Nebula Graph Store

[RithikRaj64]

Question Validation

• [X] I have searched both the documentation and discord for an answer.

Question

I was trying to create a Property Graph Index with Nebula and got this error. ('NebulaGraph query failed:', "SemanticError: To get the property of the vertex ine.Props__', should use the format var.tag.prop'".

When I looked up on the NebulaGraph documentation, this is what I found. From NebulaGraph version 3.0.0, patterns support matching multiple tags at the same time, so you need to specify a tag name when querying properties. The original statement RETURN variable_name.property_name is changed to RETURN variable_name.<tag_name>.property_name.

So how do I fix this issue?

[dosubot[bot]]

Hey @RithikRaj64! I'd love to help you with technical questions, bug fixes, and engineering knowledge. However, it seems like your query is outside the scope of what I can assist with. If you think I should have answered this request, feel free to reach out to the Dosu Team.

To continue talking to Dosu, mention @dosu.

[RithikRaj64]

@dosu Can you look into this?

[dosubot[bot]]

To fix the error "SemanticError: To get the property of the vertex in e.Props__', should use the formatvar.tag.prop'" when creating a Property Graph Index with Nebula Graph Store, you need to ensure that you are using the correct format to access the properties of the vertex. The correct format is var.tag.prop.

Here is the corrected code snippet:

def _get(
    self,
    properties: Optional[dict] = None,
    ids: Optional[List[str]] = None,
) -> List[LabelledNode]:
    """Get nodes."""
    cypher_statement = "MATCH (e:Node__) "
    if properties or ids:
        cypher_statement += "WHERE "
    params = {}

    if ids:
        cypher_statement += f"id(e) in $all_id "
        params[f"all_id"] = ids
    if properties:
        for i, prop in enumerate(properties):
            cypher_statement += f"e.Props__.{prop} == $property_{i} AND "
            params[f"property_{i}"] = properties[prop]
        cypher_statement = cypher_statement[:-5]  # Remove trailing AND

    return_statement = """
    RETURN id(e) AS name,
           e.Node__.label AS type,
           properties(e.Props__) AS properties,
           properties(e) AS all_props
    """
    cypher_statement += return_statement
    cypher_statement = cypher_statement.replace("\n", " ")

    response = self.structured_query(cypher_statement, param_map=params)

    nodes = []
    for record in response:
        if "text" in record["all_props"]:
            node = ChunkNode(
                id_=record["name"],
                label=record["type"],
                text=record["all_props"]["text"],
                properties=remove_empty_values(record["properties"]),
            )
        elif "name" in record["all_props"]:
            node = EntityNode(
                id_=record["name"],
                label=record["type"],
                name=record["all_props"]["name"],
                properties=remove_empty_values(record["properties"]),
            )
        else:
            node = EntityNode(
                name=record["name"],
                type=record["type"],
                properties=remove_empty_values(record["properties"]),
            )
        nodes.append(node)
    return nodes

Make sure to use the correct format e.Props__.prop to access the properties of the vertex ^[1].

To continue talking to Dosu, mention @dosu.

[RithikRaj64]

@dosu Where do I replace this code?

[dosubot[bot]]

You should replace the code in the _get method of the NebulaPropertyGraphStore class. This class is part of the Nebula Graph Store integration within the LlamaIndex repository. You can find it in the file nebula_property_graph.py located in the path llama-index-integrations/graph_stores/llama-index-graph-stores-nebula/llama_index/graph_stores/nebula/ ^[1].

To continue talking to Dosu, mention @dosu.

[RithikRaj64]

@Dosu Still getting the same error even after replacing the code you gave

[dosubot[bot]]

Ensure that the property access in your query uses the correct format, which is var.<tag_name>.property_name. In the context of the _get method, you should specify the tag name when accessing properties. Here's an updated snippet:

def _get(
    self,
    properties: Optional[dict] = None,
    ids: Optional[List[str]] = None,
) -> List[LabelledNode]:
    """Get nodes."""
    cypher_statement = "MATCH (e:Node__) "
    if properties or ids:
        cypher_statement += "WHERE "
    params = {}

    if ids:
        cypher_statement += f"id(e) in $all_id "
        params[f"all_id"] = ids
    if properties:
        for i, prop in enumerate(properties):
            cypher_statement += f"e.Props__.{prop} == $property_{i} AND "
            params[f"property_{i}"] = properties[prop]
        cypher_statement = cypher_statement[:-5]  # Remove trailing AND

    return_statement = """
    RETURN id(e) AS name,
           e.Node__.label AS type,
           properties(e.Props__) AS properties,
           properties(e) AS all_props
    """
    cypher_statement += return_statement
    cypher_statement = cypher_statement.replace("\n", " ")

    response = self.structured_query(cypher_statement, param_map=params)

    nodes = []
    for record in response:
        if "text" in record["all_props"]:
            node = ChunkNode(
                id_=record["name"],
                label=record["type"],
                text=record["all_props"]["text"],
                properties=remove_empty_values(record["properties"]),
            )
        elif "name" in record["all_props"]:
            node = EntityNode(
                id_=record["name"],
                label=record["type"],
                name=record["all_props"]["name"],
                properties=remove_empty_values(record["properties"]),
            )
        else:
            node = EntityNode(
                name=record["name"],
                type=record["type"],
                properties=remove_empty_values(record["properties"]),
            )
        nodes.append(node)
    return nodes

Make sure that the e.Props__.<property_name> format is used correctly, and replace <property_name> with the actual property you are trying to access. If the error persists, double-check that the tag names and property names in your Nebula Graph schema match those used in your query.

To continue talking to Dosu, mention @dosu.