Hello all I just had quick question regarding using nebula g NebulaGraph #nebula

Hello all, I just had quick question regarding usi...

Joaquin Ramirez

06/18/2024, 8:37 PM

Hello all, I just had quick question regarding using nebula graph index with llamaindex, I was wondering if it was possible to know if I can query directly with the nebula graph index that I created with this script rather than use a persistent store. I would like to eventually deploy an small application where if a user inputted a question it would be directly used within the nebula graph index on the server side to return the most relevant and fastest response. So far I have inserted my data (nodes/related nodes and edges)on my index within my space on nebula graph, however, I am having some difficulty actually being able to query with this graph index from my python script! Can anyone point me to relevant documentation for this?

Wenting

06/19/2024, 1:21 PM

@wey Can you please take a look at this question? Thank you!

❤️ 2

wey

06/20/2024, 3:53 AM

Dear @Joaquin Ramirez, please take a look at this example: https://siwei.io/tutors/GraphRAG/101.html 😄

❤️ 1

Joaquin Ramirez

06/20/2024, 1:04 PM

Thank you Wenting and thank you Wey for the help! 🙏

😉 1

Joaquin Ramirez

06/20/2024, 7:35 PM

Hey @wey, sorry for the bother, one last question! Would it be possible to load a Node Structure like the following into my graph index directly where my LLM instant will map out the relationships or would I need to reformat these nodes into a suitable structure specifically for the graph index insertion? Thanks for your time again!

Copy code

[
    {
        "id_": "3c25b01b-b555-4ab3-abd8-313e043163bf",
        "embedding": 
            0.03493161126971245,
            ...
        ],
        "metadata": {
            "lecture_number": "#",
            "lecture_title": "#",
            "topic_title": "#",
            "sub_topic_title": "#",
            "content_type": "text"
        },
        "excluded_embed_metadata_keys": [],
        "excluded_llm_metadata_keys": [],
        "relationships": {
            "1": {
                "node_id": "source_id",
                "node_type": null,
                "metadata": {},
                "hash": null,
                "class_name": "RelatedNodeInfo"
            }
        },
        "text": "#",
        "start_char_idx": null,
        "end_char_idx": null,
        "text_template": "{metadata_str}\n\n{content}",
        "metadata_template": "{key}: {value}",
        "metadata_seperator": "\n",
        "class_name": "TextNode"
    },
    {
        "id_": "f12e1dc8-748d-4c75-a647-fff0d1f79ef3",
        "embedding": [
            0.00812634825706482,
            ...
        ],
        "metadata": {
            "lecture_number": "#",
            "lecture_title": "#",
            "topic_title": "#",
            "sub_topic_title": "#",
            "content_type": "equation",
            "equation_path": "#"
        },
        "excluded_embed_metadata_keys": [],
        "excluded_llm_metadata_keys": [],
        "relationships": {
            "1": {
                "node_id": "source_id",
                "node_type": null,
                "metadata": {},
                "hash": null,
                "class_name": "RelatedNodeInfo"
            }
        },
        "text": "#",
        "start_char_idx": null,
        "end_char_idx": null,
        "text_template": "{metadata_str}\n\n{content}",
        "metadata_template": "{key}: {value}",
        "metadata_seperator": "\n",
        "class_name": "TextNode"
    },
...
]

Joaquin Ramirez

06/23/2024, 2:45 PM

@Wenting Hey Wenting, sorry for the bother again, but perhaps, you could address this question?

Open in Slack

Previous Next