Open-source migration framework for legacy GLAM exports to AtoM CSV

[matthewb...@gmail.com]

Hi all,

I’ve just published an early version of a new open-source tool called Open Collections Migration Framework (OCMF):

https://github.com/matthewbrutonall/open-collections-migration-framework

In plain terms, OCMF is designed to help move catalogue/archival data out of older collections systems and into AtoM.

At the moment it focuses on exported data rather than direct database connections. The current version supports:

  - PastPerfect-style CSV exports
  - Axiell/Adlib XML exports
  - AtoM-compatible ISAD(G) information_objects.csv output

  The basic workflow is:

  1. Create an Excel mapping workbook.
  2. Map source fields to AtoM-compatible target fields.
  3. Run the migration pipeline.
  4. OCMF builds an internal graph of the records and relationships.
  5. It reconstructs hierarchy from reference codes.
  6. It runs preflight checks for issues like missing identifiers, duplicate identifiers, empty titles, orphan records, and hierarchy cycles.
 7. It emits an AtoM-ready CSV plus an audit pack.

  The audit pack includes:

  - unmapped source fields
  - mapped fields missing from the export
  - synthesized hierarchy nodes
  - interpretive transformations, such as collapsed date ranges
  - a load playbook with suggested AtoM validation/import commands

This is not intended to replace proper migration review or AtoM’s own CSV validation. The aim is to provide a reusable, testable framework for the messy transformation stage before AtoM import, where a lot of migration projects currently rely on one-off scripts.

The project is still early, but it has tests around hierarchy reconstruction, date parsing, event-field alignment, source- field auditing, and failure cases. The AtoM writer currently targets the AtoM 2.8-compatible ISAD(G) CSV template structure, which should also be relevant to later/compatible CSV import workflows, but I’d welcome corrections from anyone with deeper experience across versions.

I’d be very interested in feedback from AtoM users, archivists, and migration consultants, especially on:

  - missing AtoM CSV fields or field-name issues
  - PastPerfect/Adlib export edge cases
  - hierarchy reconstruction assumptions
  - useful audit/reporting features
  - whether this kind of framework would be useful in real migration work

  Thanks,

  Matthew Bruton

[Johan Pieterse]

Looks good Matthew