DIMP (Pseudonymization)

De-identifies FHIR data using a DIMP service.

What it does

• Sends FHIR Bundles to DIMP service
• Automatically splits large Bundles (>10MB by default)
• Saves pseudonymized data

Configuration

services:
  dimp:
    url: "http://your-dimp-server:32861/fhir"
    bundle_split_threshold_mb: 10  # 1-100 MB, default: 10

pipeline:
  enabled_steps:
    - local_import
    - dimp

Configuration Options

Option	Type	Default	Description
url	string	-	DIMP service URL (required)
bundle_split_threshold_mb	int	10	Split Bundles larger than this (1-100 MB)

Bundle Splitting

Large FHIR Bundles are automatically split to prevent HTTP 413 errors:

• Bundles exceeding the threshold are partitioned into smaller chunks
• Each chunk is sent separately to DIMP
• Results are reassembled after processing
• 100% data preservation during split-reassemble

Output

Pseudonymized files are saved to jobs/<job-id>/pseudonymized/