http://readthedocs.org/projects/ifiscripts/badge/?version=latest

Introduction

Summary

These are Python 2.7 scripts that facilitate collections management workflows within the IFI Irish Film Archive. These scripts have been tested in OSX, Windows 7 & 10, Ubuntu 14.04 & 16.04. They are located here on github: https://github.com/kieranjol/IFIscripts

Most scripts take either a file or a directory as their input, for example makeffv1.py filename.mov or premis.py path/to/folder_of_stuff. (It’s best to just drag and drop the folder or filename into the terminal as this provides the absolute path).

We want the project to be as reuseable as possible in different institutions and contexts. Some scripts, particularly anything to do with Object Entry or Accessioning will be quite IFI specific, but other scripts such as makeffv1.py, dcpaccess.py and many others have been used in a variety of contexts in several different countries.

The project uses the MIT license, and we encourage the reuse, modification and study of the scripts. It’s always nice to hear when the scripts have been reused in some way, but it’s not necessary to let us know.

Purpose

These python scripts facilitate much of our collections management procedures for digitised and born digital objects in the Irish Film Institute. We utilise a lot of open source tools, so we wanted to make these scripts as open as possible. This is why this project has the MIT License.

The Irish Film Institute has followed the SPECTRUM museum collections management standard for several years. These scripts attempt to follow SPECTRUM procedures while also utilising some of the concepts of the Open Archival Information System (OAIS). Initially the scripts only handled single video files, but they are now capable of handling:

  • Digital Cinema Packages
  • XDCAM cards
  • DPX/TIFF image sequences
  • Documents (.doc, .pdf etc)
  • Images (.jpg, .TIFF etc)

An example workflow might be:

  • A digital object is created or acquired by the IFI, and ingest begins.
  • sipcreator.py is run on the object. This script:
    • generates an Object Entry identifier (eg OE-1234)
    • creates a folder structure for logs, metadata, objects
    • generates a UUID, extracts technical metadata
    • generates a md5 checksum manifest
    • and see the usage section for more.
  • All of these preservation events are logged in a log file located in the logs directory. This log file tries to use PREMIS (PREservation Metadata Implementation Strategies) terminology as much as possible.
  • Even though the package has yet to be accessioned, temporary backups are required. copyit.py will generate backups, and it will use the checksum manifest generated by sipcreator.py to verify the integrity of the file transfer.
  • If the package contains FFV1 or Matroska files, perhaps ffv1mkvvalidate.py could run, which uses mediaconch to verify the compliance of the files, and stores the information in the logfile.
  • If the package passes our Quality Control Procedures, then it will be accessioned. accession.py will generate an accession number, rename the OE number with the accession number, generate a SHA-512 manifest and update the log file to document these new preservation events.
  • A large batch of items can be accessioned using batchaccession.py. If the -pbcore command line argument is used with the accessioning scripts, technical metadata based on the PBCore standard will be generated in CSV format. This process can be run seperately by using makepbcore.py. CSV was chosen instead of XML as this allows us to immediately import the CSV into our database system so that we have item level records.
  • Access copies may be needed, so low-res watermarked proxies can be generated with bitc.py, or high res mezzanines with prores.py.
  • The accessioned package can then be written to preservation storage, again using the copyit.py command.

So this is just one way of using the scripts from acquisition to preservation storage, but there are many other scripts for specific workflows, which you can investigate further down in the documentation.