DOMS TestBed 0.2
Delivered for testing against DOMS on 2008-09-12.
DOMS TestBed 0.2 contains full API for DOMS, but with mockup functionality.
The APIs are not completely stable, but no structural changes should occur.
What's included
The Test Bed includes
- A tomcat
- A fully functional and configured Fedora 3.0
- A mockup PIDGenerator webservice
- A mockup Bitstorage webservice
- A mockup Search webservice
- The DOMS Base Collection Content Models
- The DOMS Base Collection base objects
- The Reeltape Test Collection Content Models
- The Reeltape Test Collection test objects
How to install
The Test Bed requires a functional Linux system with Java 1.6 installed.
Unzip the package and read INSTALL for installation instructions.
The APIs
All Fedora APIs
- API-A
- API-M
- REST API
- Resource Index
The usage of the Fedora APIs have the following restrictions (Note: The restrictions are not enforced in TestBed):
Fedora 3.0 API as used by DOMS
BitStorage API
An API for uploading a file to storage:
The mockup functionality stores files locally.
The mockup characterisation does not characterise correctly. The Pronom ID (the file type) is decided purely on the basis of the file extension, with the following table:
{"html", "fmt/100"},
{"tiff", "fmt/10"},
{"tif", "fmt/10"},
{"jpeg", "fmt/44"},
{"jpg", "fmt/44"},
{"pdf", "fmt/20"} All other extensions are considered unknown and invalid.
In the final version the files will be characterised on the basis of the file contents.
The actual characterisation output is pure mockup.
The mockup always returns 0 as diskfree
PID Generator API
The mockup is trivial but fully functional.
Search API
Note the mockup will never sort the results, and search is done on individual Fedora objects, not full posts. Also the mockup searches in the DC datastream, and nothing else.
The mockup search API does no sorting.
The mockup search API does no paging - always shows the first 20 results.
The datamodel
The data model is unchanged since testbed 0.1. That means the test bed still has the following issues:
The data model still has outstanding issues:
- The VIEW datastream will probably be updated to another language better suited to express the same functionality now offered.
The test objects still have outstanding issues:
- The Qualified Dublin Core datastreams definitely need further specialisation. Discussions need to be made about the specific use of XMLSchema too.
- The ORIGIN datastreams are still very poor mockups.
How final is this
Just as in the first Test Bed:
The Test Bed needs to be seen as a first effort, and every piece of it is subject to change.
The ideas and structure is not expected to change radically. However, it is expected that further work with the webinterfaces and data models will lead to changes.
We will be happy to discuss requested changes, of course, and we hope you will be understanding when we need to change things due to our own requirements.
Any major changes will be discussed with you first.