Archiving Toolkit
- Year
- 2024
- Role
- Sole developer
- Status
- live
A set of small Python tools that removed the repetitive, error-prone parts of preparing a digital deposit for the archive — and the work that turned a data-entry job into a systems one.
- Python
- GIS
- GML
- Windows
Problem
Preparing a deposit for long-term archiving involves a lot of checking that a person can easily miss when large archives come in. Attribute fields have to match the supplied metadata. Filenames have to meet the archive's standards. GIS material has to reach a preservation format, with images for the website.
Done by hand, all of it is slow, and all of it is the kind of work where attention drifts on the two hundredth file.
Approach
Rather than one large application, several small tools that each remove one repetitive job:
- GIS metadata checker — reads the attribute fields in the supplied files, compares them against the metadata provided with the deposit, and highlights what is missing. The check that previously meant opening files and reading across two sources.
- Filename normaliser — brings filenames up to ADS standards, stripping spaces, exclamation marks, plus signs and the other characters that cause trouble later in a preservation pipeline.
- GIS preservation exporter — converts GIS formats to GML 3.2 and generates the screenshots the website needs, collapsing a manual conversion into a run.
Written in Python and packaged as executables, so colleagues who do not program could run them without a Python environment to set up. Distribution matters as much as the tool: something that requires a runtime to be installed first will not be used by the people it was written for.
Outcome
The checking and conversion steps that dominated deposit preparation became things you run rather than things you do. The GIS tools are still in use by the archive team.
The honest part is what happened to the rest. Several of the tools are now deprecated, and the ones that survive are mostly used by longer-serving archivists — newer staff do not know they exist. The tooling outlasted the knowledge of it, which is a documentation and handover problem rather than a technical one, and a mistake I would not repeat: shipping a tool is not the same as embedding it.
This is also the work that helped to change my career. It was built during a Digital Archives Assistant role — nominally data entry — and showing that the team's repetitive work could be automated is what helped to land me in to the System Officer post.