Root Processes

Root processes are named families that group related processes. Every process in the process schema must belong to a root process. Root processes themselves are stored in process.root_process and registered before any other processes in the setup_processes workflow.

Process file

File: ./setup/zzz/ai4sh/setup_processes/json_ai4sh/root_process/root_processes_v10_sql.json

Registered root processes

The AI4SH database registers two root processes:

manage_table_data

{
  "process": "add_root_process",
  "parameters": {
    "root_process_id": "manage_table_data",
    "title": "Manage table data",
    "label": "Root for processes for inserting, updating and deleting table data"
  }
}

All data management operations — inserting, updating, and deleting records in any schema — belong to this root. Every process registered in the subsequent community, utility, observation_utility, observation, landscape, and eDNA process files will reference manage_table_data as its root.

translate_data

{
  "process": "translate_data",
  "parameters": {
    "root_process_id": "translate_data",
    "title": "Translate data",
    "label": "Process for translating data between different formats — note that translate data does not affect the database, it prepares data in different formats to be handled by other processes."
  }
}

Translation processes convert data between formats (e.g. from a tabular spreadsheet to the JSON process file format expected by the framework). Crucially, translation processes do not write to the database directly — they prepare data for subsequent processes that do.

Why root processes matter

The root process is part of the primary key structure for process lookup. When the framework receives a process call it uses both the root_process_id and the process_id to find the correct Python function to execute. This two-level lookup means you can have the same process name in different root families without collision.

Root processes also serve as a grouping mechanism in the user interface — all processes under manage_table_data can be presented together as the “data management” operations for a given schema.