openBEB organizes data in collections. Collections can contain other collections. The collection root is called “collection.bundle”. A bundle also defines the file system location where the collection hierarchy is stored.
A collection contains four forks, or simply folders in the file-system:
- Fork one contains the raw-data. The raw data is never changed by openBEB, however, openBEB calculates an md5 check sum to monitor the data-integrity and allows data compression by a lossless algorithm.
- The second fork, annotations, contains metadata information, preview and graphically annotated representations of the data. It also contains the embedded metadata of the raw files in a standardized xml format.
- The third fork contains cache information. This information stores cache files for faster browsing or stores processing results. All of this information is redundant and can be recalculated at any time from the information contained in forks one and two. Therefore, the information of the cache fork is not synchronized with the database.
- Fork four is a container hosting sub-collections, allowing an arbitrary hierarchical structure of a collection to be built.
All openBEb related files are standard formats such as XML files or key-value ini files. The data- in the cache fork depend of the viewports (and other plug-ins) used and can be binaries.
The meta-data files are either XML files or PNG files (for graphical representations called “views”).
Curent meta-data format
- Version 2 for data-set metadata
- Version 1 for all other metadata files
Old meta-data formats
- A description of the met-data files version 1 can be found here: Meta-data (Vs. 1)