Offline data browser model

[zhujun98]

The idea is to extend the use case of this software to offline reconstructed data browsing. This is motivated by the fact that a whole 3D volume would be required to be reconstructed in most existing pipelines even if the user only wants to check a single 2D slice during offline data analysis.

Since this is a client/server application, in which the GUI typically runs on a console machine in a hutch and the reconstruction server runs on a GPU node, there are two components required: the control (file loading, slice selecting) and the actual file reader. There are two ways of implementing the file reader.

The first one is to implement it as another module in the server code. The disadvantage is that it can only be used in this project.

The second one is to implement it as a third-party library, for example in foamstream. The disadvantages are: 1. Another service is required to run. 2. Data is required to be streamed to the server , which might have a higher latency.