AI Dashboard Runner for FEM-enabled FLCore tasks

This repository contains the code of a dockerized openVRE analysis tool. It implements a wrapper that enable the DT4H AI Dashboard to dispatch federated task across the federated data nodes (FDNs) via the FEM-orchestrator API.

Run Parameters:

--server_node SERVER_NODE Server node --client_node_list CLIENT_NODE_LIST Client nodes, comma sep --tool_name TOOL_NAME --input_params_path INPUT_PARAMS_PATH Path to Application parameters (JSON|YML) --input_dataset_path INPUT_DATASET_PATH Path to Input dataset reference (JSON) --input_variables_path INPUT_VARIABLES_PATH Path to Input variables from Mica search (zip) --target_label TARGET_LABEL Target label for the model (used with Mica variables) --health_check HEALTH_CHECK Perform heartbeat before executing and store at file --job_timeout JOB_TIMEOUT Job timeout duration (seconds) --finish_wait FINISH_WAIT Finish wait duration (seconds) --files_timeout FILES_TIMEOUT Files timeout duration (seconds) --output_path OUTPUT_PATH Output directory path

Getting started

1. Clone this repository
2. Change in the Dockerfile the FROM image with the one for your tool, the rest leave it as it is​
3. Adjust the VRE_Tool.py of the vre_template_tool_dockerized repo and the tests folder​
4. Build the image with the modified Dockerfile​
5. Test with new image​

Current status

Functional: Code under developement. Demostrated at DT4H consortium meeting (Sept 2025)