Skip to main content

SDK 3.0

What's new?​

Contextual Engine​

The new contextualEngine setting allows you to automate the entire import process, from mapping to transforming and cleaning data, using nuvo’s Contextual Engine.

You can either:

  • Fully automate the process with "enabled".
  • Give users the choice between manual and automated flow using "prompt".

For more information, check out the contextualEngine setting.

Self-Hosted support with baseUrl​

We’ve added support for running nuvo’s backend in your own infrastructure. By setting the new baseUrl property, the importer will send all requests to your self-hosted instance instead of the default nuvo cloud endpoint. This is useful for air-gapped environments, regulated industries, or companies with strict data residency requirements.

For more information, see the baseUrl setting.

Smarter, faster, and more configurable mapping​

With SDK 3.0, we've significantly enhanced our mapping capabilities. In addition to better performance and higher accuracy, you now have fine-grained control over how both columns and category-like values (options) are matched using two new configuration keys: columnMappingConfiguration & optionMappingConfiguration

These settings allow you to:

  • Select where the mapping logic is executed.

    • Use "node" to unlock server-side power like historic and AI mapping, or "browser" to keep processing on the client-side.

  • Define which matching layers are used.

    • "exact" – Performs strict matches by comparing normalized strings (ignores case, spaces, and special characters).

    • "historic" – Reuses successful mappings from previous imports for returning users or recurring templates (Requires "node" as processing mode).

    • "smart" – Uses nuvo’s AI model to understand ambiguous or loosely related names and suggest the most relevant match (Requires "node" as processing mode).

    • "fuzzy" – Matches values based on partial text similarity.

  • Set a confidence threshold (a number between 0 and 1) to fine-tune the sensitivity of matching.

    • A higher threshold makes matching stricter, while a lower value increases flexibility.

These settings apply at two levels:

  • columnMappingConfiguration – Controls how input headers are matched to your target data model (link).

  • optionMappingConfiguration – Controls how unique input values are mapped to the defined options of a category column (e.g., "USA" → "United States") (link).

We recommend using "node" mode whenever you want to:

  • Enable AI ("smart") or historical ("historic") mapping layers.

  • Improve performance when importing large files.

  • Improve performance with large target data models containing hundreds of columns and thousands of dropdown options.

Deprecated properties​

As part of the cleanup and simplification:

  • The optionMappingMode and processingEngine settings have been deprecated.
  • Use optionMappingConfiguration instead to control option mapping behavior.

How to migrate to SDK 3.0​

We now use scoped npm packages for all frameworks. Please update your implementation accordingly:

Install the latest React SDK:

npm install @getnuvo/importer-react

or upgrade it with Yarn:

yarn add @getnuvo/importer-react
info

These new packages replace the older nuvo-* packages used in SDK 1.x and 2.x. To benefit from the latest features, improvements, and ongoing support, we recommend migrating to the new @getnuvo/importer-* packages.