Embedding in TextNode show undefined

[myhendry]

hi, I am new to llamaindex.

i configure my Settings.embedModel as such:

 Settings.embedModel = new OpenAIEmbedding({
    apiKey: process.env.OPENAI_API_KEY,
    model: "text-embedding-ada-002",
  });

however, in my TextNode, i am seeing embedding undefined in my TextNode. is this the correct behavior? are my docs in llamaindex ts embedded since I have already set the embedModel as OpenAIEmbedding?

  relationships: {
    SOURCE: {
      nodeId: '/Users/lim/Documents/Coding/history/app/data/stevejobs.txt_1',
      metadata: [Object],
      hash: '4YXPisH9BdRViVAufO7kMdLcj7zQD3fJcfpCsIDwdys='
    },
    PREVIOUS: {
      nodeId: '86d2acac-9a99-4318-9d90-df8fdf0ed020',
      metadata: [Object],
      hash: 'DZk8iPyblNBQO9v46ififMxkxu4dF6AKmRzuRZmUVrg='
    },
    NEXT: {
      nodeId: '5844bae0-2e48-4f9c-8691-a3aa71044136',
      metadata: {},
      hash: 'bjECarGwzCAcxHjjNa01uIjy3ShCe2X2nCMp0rcGF2I='
    }
  },
  embedding: undefined,
  text: 'Founding of Apple: In 1976, Steve Jobs, Steve Wozniak, and Ronald Wayne founded Apple Computer Inc., with the goal of developing user-friendly computers.The first product, Apple I, was soon followed by the Apple II, which was a major success and truly brought computers into the homes of everyday people.',
  textTemplate: '',
  metadataSeparator: '\n'
}

[marcusschiesser]

How did you generate your nodes?

[myhendry]

Hi Marcus,

below is sample of my code

sorry, i come from langchain and new to llamaindex

i set my embedding
Settings.embedModel = new OpenAIEmbedding({
apiKey: process.env.OPENAI_API_KEY,
model: "text-embedding-ada-002",
});

Did i miss out anything in order for the data to be embedded before inserting into the qdrant database?

Thanks

export async function POST(req: Request) {
  const { messages }: { messages: Message[] } = await req.json();

  const folderPath = path.join(process.cwd(), "app", "data");
  const fullPath = path.resolve(folderPath);

  //! Read the docs
  const reader = new SimpleDirectoryReader();
  const documents: Document<Metadata>[] = await reader.loadData(folderPath);

  //! Split the docs
  const nodeParser = new SentenceSplitter({
    chunkSize: 100,
    chunkOverlap: 20,
  });

  //! Configure Settings
  Settings.llm = new OpenAI({
    model: "gpt-3.5-turbo",
    apiKey: process.env.OPENAI_API_KEY,
  });
  Settings.embedModel = new OpenAIEmbedding({
    apiKey: process.env.OPENAI_API_KEY,
    model: "text-embedding-ada-002",
  });
  Settings.nodeParser = nodeParser;

  //! Create Quadrant Vectorstore and Index the Data
  const qdrantVs = new QdrantVectorStore({
    url: process.env.QDRANT_URL,
    apiKey: process.env.QDRANT_VECTOR_DB,
    collectionName: "data",
  });
  const ctx = await storageContextFromDefaults({ vectorStore: qdrantVs });
  const index = await VectorStoreIndex.fromDocuments(documents, {
    storageContext: ctx,
  });

  /*
  ! Retrieve closest result from vector database
  .retrieve("user_query") => retrieve the relevant documents from the data
  .query("user_query") => LLM and generate the response
  */

  const responseSynthesizer = getResponseSynthesizer("tree_summarize");
  const query_engine = index.asQueryEngine({ responseSynthesizer });

  const result = await query_engine.retrieve({
    query: messages[messages.length - 1].content,
  });
  result.map((r) => {
    console.log(r.node);
  });

  const prompts = query_engine.getPrompts();
  console.log({ prompts });

  const newTextQaPrompt: TextQAPrompt = new PromptTemplate({
    template: `Context information is below.
    ---------------------
    {context}
    ---------------------
    Given the context information and not prior knowledge, answer the query.
    Answer the query in the style of a funny person.
    Query: {query}
    Answer:`,
  });
  query_engine.updatePrompts({
    "responseSynthesizer:textQATemplate": newTextQaPrompt,
  });

  const stream = await query_engine.query({
    query: messages[messages.length - 1].content,
    stream: true,
  });

  return LlamaIndexAdapter.toDataStreamResponse(stream);
}