Backend API Reference

Complete API documentation for the AI-DAPT Data Cleaning Engine backend.

Authentication

The backend uses Keycloak-based client authentication. Each request requires a valid client token, obtained via your authentication workflow.

Endpoints

POST /run/

Run a processing job (imputation or anomaly detection) on a dataset.

POST /run/
Authorization: Bearer <token>
Content-Type: application/json

{
  "data": "<job_id>",
  "task": "impute | detect",
  "dtype": "tabular | timeseries",
  "method": "<method_name>",
  "date_col": "<datetime_column_name>",   // only for timeseries
  "params": { "<method_param>": "<value>" } // optional
}

POST /approve/

Approve a previously executed job.

POST /approve/
Authorization: Bearer <token>
Content-Type: application/json

{
  "job_id": "<job_id>",
  "approval_info": { "<key>": "<value>" }
}

GET /get-new-jobs/

Retrieve all currently running jobs that have no results stored yet.

GET /get-new-jobs/
Authorization: Bearer <token>

GET /get-unapproved-jobs/

Retrieve results that have not been approved yet.

GET /get-unapproved-jobs/
Authorization: Bearer <token>

GET /get-results/{uid}

Get results for a specific job by uid.

GET /get-results/\{uid}
Authorization: Bearer <token>

See Also