AGILITY now supports multiplexing multiple models per pipeline, which allows running different models based on the input service or folder context. This feature provides flexibility in how data is processed, offering a range of analysis modes, from service-specific  to a more general brute-force approach.
Pipeline Service Multiplexing Logic:

1. Service as Input Parameter:

• If the user provides a service as an input parameter, the pipeline will run the analysis using that specific service model.

2. No Service as Input Parameter:

• If no service is provided, the pipeline will check the folder name where the input data is stored.

3. a. Folder Name = Service Key:

• If the folder name matches a service_key, the pipeline will run the analysis using the model for that specific service.

4. b. Folder Name != Service Key:

• If the folder name does not match any service_key, the pipeline will fall back to a brute force mode, running a generic model that attempts to analyze the data without any service-specific configuration.

Constraints:

• Service Key Uniqueness: Each folder used in this pipeline configuration must have a unique service_key associated with a model.

• Brute Force Mode: When no service is identified, the brute-force mode attempts to handle the data, but may produce less optimal results compared to using a service-specific model.

• Performance Considerations: Running multiple models across services may increase pipeline runtime. Ensure appropriate resource allocation and model selection to avoid performance bottlenecks.