Hyercube viewer: New displays and new data-types

Dataset

A dataset is initialized by an io_func and creates buffers as it needs them. Currently there are four basic buffer types. The simplest is inherited from incore_data. As the name implies this data type is stored completely in memory and has the same number of axes as the controlling grid. The created_data class is for data that changes depending on the user's interaction with the program. An example of this is the semblance and NMO corrected datasets in the velocity analysis example. If a dataset does not contain all of the axes of the grid, it should be stored using the partial_data class.

The outcore_data class, as the name implies, is for datasets that are not stored completely in memory. This class reads in a 3-D subset of the domain based on the order the axes are being displayed in a given view. For example, given a 4-D volume and a view that is displaying the first, third, and fourth axis, a 3-D subset of the data will be read centered at the current position along the second axis. As long as the position is only being changed in these three dimension no further read requests are necessary. Several of these out-of-core buffers are created for each dataset allowing multiple views display multiple subsets and allowing switching back and forth between positions without having constant data read delays.

Hyercube viewer: New displays and new data-types