Overview
This node integrates with the Recombee recommendation service to provide item-to-item recommendations. Specifically, it recommends items related to a given item ID, which is useful in scenarios such as e-commerce product suggestions, content recommendation engines, or any system where suggesting similar or complementary items enhances user experience.
Typical use cases include:
- Suggesting products related to a currently viewed product.
- Recommending accessories or add-ons based on an item.
- Enhancing content discovery by recommending articles or media related to a selected piece.
The node supports batching requests for efficiency and allows customization of recommendation parameters like scenario, count, and whether to return detailed properties of recommended items.
Properties
| Name | Meaning |
|---|---|
| Item ID | The ID of the item to recommend other items for. |
| Return Properties | Whether to return the properties/details of the recommended items (true/false). |
| Cascade Create | Whether to create the item in Recombee if it does not already exist (true/false). |
| Filter | Currently unused filter parameter; intended for advanced filtering per Recombee documentation. |
| Count | Number of recommended items to return. |
| Scenario | Optional scenario name to customize recommendation context. |
| Batch Size | Number of requests to send per batch when processing multiple input items. |
| Timeout (Ms) | Request timeout duration in milliseconds. |
| Max Retries | Number of retry attempts on request failure before throwing an error. |
Output
The node outputs an array of JSON objects, each corresponding to an input item processed. Each output object contains:
success: Boolean indicating if the recommendation request succeeded.itemId: The original item ID for which recommendations were requested.userId: (Present in code but not defined in properties; likely unused or optional).count: Number of recommended items requested.scenario: The scenario used for the recommendation.returnProperties: Whether properties were returned.data: The actual recommendation results from Recombee, including recommended item IDs and optionally their properties.error: Present only if the request failed, containing the error message.
If binary data were involved, it would be summarized here, but this node only outputs JSON data.
Dependencies
- Requires an API key credential for Recombee with access to the target database.
- Uses the official Recombee API client library (
recombee-api-client). - Node configuration must include credentials with database ID, private token, region, and optionally a timeout setting.
- Network connectivity to Recombee's API endpoints is necessary.
Troubleshooting
Common Issues:
- Invalid or missing item IDs will cause errors or empty recommendations.
- Network timeouts or connectivity issues may cause retries and eventual failures.
- Incorrect credentials or insufficient permissions will result in authentication errors.
- Using unsupported or malformed filters (currently unused) might cause unexpected behavior if implemented later.
Error Messages:
- Errors thrown by the Recombee API client are caught and either cause the node to fail or, if "Continue On Fail" is enabled, return an error object per item.
- Typical errors include authentication failures, invalid parameters, or request timeouts.
Resolutions:
- Verify that the item IDs exist in the Recombee database or enable cascade creation if appropriate.
- Check API credentials and permissions.
- Adjust timeout and max retry settings to handle network instability.
- Use the "Continue On Fail" option to allow partial success in batch operations.
Links and References
- Recombee API Documentation: https://docs.recombee.com/
- Recombee Filtering and Boosting Docs: https://docs.recombee.com/reql_filtering_and_boosting#reql-filtering