TypeError - TS Support?

[morganric]

Hello, I'd like to use this extension within a firebase Node TS application, but I am running into this error:

TypeError: Cannot read properties of undefined (reading 'k')
>      at /Users/richardmorgan/Documents/dev/neuro/mvp/functions/node_modules/genkitx-neo4j/lib/index.js:140:22
>      at Generator.next (<anonymous>)
>      at fulfilled (/Users/richardmorgan/Documents/dev/neuro/mvp/functions/node_modules/genkitx-neo4j/lib/index.js:47:24)
>      at process.processTicksAndRejections (node:internal/process/task_queues:95:5) {
>    traceId: '943d263166424de254d1c94d11064647'
>  }

TIA!

[tomasonjo]

Can you show thecode to replicate this? V sre., 25. jun. 2025, 16:04 je oseba morganric ***@***.***> napisala:

…

*morganric* created an issue (neo4j-partners/genkitx-neo4j#6) <#6> Hello, I'd like to use this extension within a firebase Node TS application, but I am running into this error: TypeError: Cannot read properties of undefined (reading 'k') > at /Users/richardmorgan/Documents/dev/neuro/mvp/functions/node_modules/genkitx-neo4j/lib/index.js:140:22 > at Generator.next (<anonymous>) > at fulfilled (/Users/richardmorgan/Documents/dev/neuro/mvp/functions/node_modules/genkitx-neo4j/lib/index.js:47:24) > at process.processTicksAndRejections (node:internal/process/task_queues:95:5) { > traceId: '943d263166424de254d1c94d11064647' > } TIA! — Reply to this email directly, view it on GitHub <#6>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEYGGTLRFCZT34EDMKJHBT33FKUAVAVCNFSM6AAAAACADNHS3OVHI2DSMVQWIX3LMV43ASLTON2WKOZTGE3TKNZYHE3TMMI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[morganric]

Sure... I did not set the indexer as I have built a graph already in auraDB / neo4j using this tool:
https://llm-graph-builder.neo4jlabs.com/

  
import {genkit} from "genkit";

import {
  neo4j,
  neo4jRetrieverRef,
} from "genkitx-neo4j";

 // Initialize AI configuration
    const ai = genkit({
      plugins: [
        googleAI({apiKey: process.env.GEMINI_API_KEY}),
        neo4j([
          {
            indexId: "index_343aff4e",
            embedder: textEmbedding004,
          },
        ]),
      ],
      model: gemini15Flash,
    });

// KAG

      // To specify an index:
      const kagFactsRetriever = neo4jRetrieverRef({
        indexId: "index_343aff4e", // "index_343aff4e"
      });
      console.log("kagFactsRetriever: ", kagFactsRetriever);  

      // To use the index you configured when you loaded the plugin:
      const KAGdocs = await ai.retrieve({
        retriever: kagFactsRetriever,
        query: "What is the best therapy for anxiety?",
      });
      console.log("KAGdocs.length: ", KAGdocs.length);

      console.log("KAGdocs:", KAGdocs);

      /// END KAG

[morganric]

Perhaps I need to create the index too?

e.g.

export const bobFactsIndexer = neo4jIndexerRef({ indexId: 'bob-facts', }); await ai.index({ indexer: bobFactsIndexer, documents });

[tomasonjo]

Did you manage to get it working?

…

On Wed, Jun 25, 2025 at 9:27 PM morganric ***@***.***> wrote: *morganric* left a comment (neo4j-partners/genkitx-neo4j#6) <#6 (comment)> Perhaps I need to create the index too? e.g. export const bobFactsIndexer = neo4jIndexerRef({ indexId: 'bob-facts', }); await ai.index({ indexer: bobFactsIndexer, documents }); — Reply to this email directly, view it on GitHub <#6 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEYGGTJDHH7VX7QJ7JTGWSL3FLZYNAVCNFSM6AAAAACADNHS3OVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTAMBVHEYDANRUG4> . You are receiving this because you commented.Message ID: ***@***.***>

[morganric]

@tomasonjo no ... do I need to create an indexer first, even if I have a neo4j auraDB setup? What is the "documents" object I should pass?

I got the index "name" from here ... should it be a different id, perhaps?

[tomasonjo]

you should be using index name "vector". It should probably work. Atm, the genkit is not customizable, so it expects text property on the node. You will get other properties of the node as metadata

[morganric]

@tomasonjo okay thanks. I changed to that but I still get the same error...

seems to be from this file in your lib responsible for the retriever:

[morganric]

Should I be supplying some options?