Skip to content

Flow 5.1.0

Apply Selective Modules for Faster Indexing

We introduce a significant enhancement to our indexing and analyzing process with the introduction of the modules.apply parameter. By specifying one or more registered modules, you can now restrict image analysis exclusively to these selected image analysis algorithms. This targeted approach to indexing is especially beneficial for use cases where only certain modules are relevant, speeding up the indexing process by focusing on the most pertinent data.

For instance, by using modules.apply=content to focus on visual similarity searches only, you can double your indexing speed by excluding irrelevant modules.

Performance gains with selective module indexing

See the module.apply API reference for details.

Support for ARM64 CPUs

We've introduced support for ARM64 architectures (ARMv8 is supported; older versions like ARMv7 are not supported). Using ARM CPUs significantly reduce power consumption and operational costs, making it a viable choice for both small and large-scale server environments. The adoption of ARM is increasingly popular in server environments due to its cost-effectiveness.

However, it's important to note that while ARM is more cost-effective, it may result in slower image analysis. Nonetheless, this should be negligible for search-related operations.

Multi-Architecture Docker images

We now provide Docker images compatible with both ARM and x86 architectures, making it easier to deploy across a wider range of platforms.

Near Real-Time Search with Optimized Auto-Commit Settings

We've fine-tuned the auto-commit settings in the sample configuration, which is also used in our Docker images, to significantly improve the availability of newly indexed documents for a near real-time search experience.

To ensure that the latest index changes are available for search queries as quickly as possible, a soft commit is automatically triggered every second. In addition, the more resource-intensive hard commit is triggered every minute to persist data to disk.

Support for Solr 9.8

We support the latest Solr version. Solr 9.8 introduces 6 dependency upgrades and 16 bug fixes, complemented by a range of new features and improvements to enhance performance and reliability.

Dropped Support: Solr 7.7

Solr 7 has reached its end of life (EOL) and is no longer maintained. With this release, Flow no longer supports Solr 7.7. If you are still relying on Solr 7.7, we strongly recommend upgrading to a supported Solr version, such as Solr 9.x, to ensure system stability, maintain security, and guarantee continued compatibility with Flow.

Upgrade from Flow 5.0.X

This release is a drop-in replacement for Flow 5.0.X.