Overview
This node integrates with Oracle Cloud Infrastructure (OCI) Generative AI services to generate text embeddings using OCI's On Demand Models. It is designed to transform input text into vector representations (embeddings), which are useful for tasks such as semantic search, clustering, recommendation systems, and natural language understanding.
Common scenarios include:
- Enhancing search functionality by converting queries and documents into embeddings for similarity comparison.
- Grouping or categorizing large sets of text data based on semantic content.
- Feeding embeddings into machine learning models for downstream AI applications.
Practical example: A user can input product descriptions and obtain embeddings that help cluster similar products or improve search relevance in an e-commerce platform.
Properties
| Name | Meaning |
|---|---|
| Compartment ID | The OCI compartment identifier where the generative AI model resides. If not specified, defaults to tenancy OCID. |
| On Demand Model Name or ID | Selects the specific OCI On Demand Model used to generate embeddings. Options are loaded dynamically from OCI and include only models supporting text embeddings. |
Output
The node outputs a JSON object containing the generated embeddings under the json field. The output structure corresponds to the embedding vectors produced by the selected OCI model. The node’s output connection type is specifically for AI embeddings, indicating it is intended to be consumed by subsequent nodes expecting embedding data.
No binary data output is indicated.
Dependencies
- Requires an API key credential for OCI with access to Generative AI services.
- Needs OCI tenancy details including tenancy OCID, user OCID, key fingerprint, private key, passphrase, and region.
- Uses OCI SDK libraries (
oci-commonandoci-generativeai) to authenticate and interact with the OCI Generative AI service. - The node dynamically loads available models supporting text embeddings from OCI within the specified compartment.
Troubleshooting
- Authentication errors: Ensure the OCI API credentials are correctly configured with valid private key and associated details. Private key formatting is adjusted internally; corrupted or invalid keys will cause failures.
- Model selection issues: If no models appear in the dropdown, verify that the OCI account has access to Generative AI models in the specified compartment.
- Compartment ID problems: Providing an incorrect compartment ID may result in empty or failed model listings. If omitted, the tenancy OCID is used by default.
- API rate limits or network errors: These may cause request failures; retry or check OCI service status.
- Error messages related to missing capabilities: The node filters models to those supporting text embeddings; selecting unsupported models manually (via expressions) may cause errors.