CIBC:Documentation:SCIRun:Reference:ModelCreation:StreamMatrixFromDisk
Contents |
StreamMatrixFromDisk
Information
- Package: ModelCreation
- Catagory: DataStreaming
- Author(s): Jeroen Stinstra
- Status: Supported in latest version
- Version: 3.0
Description
Summary
Stream Data from a file on disk without loading the data into memory.
Detailed Description
Stream Data from a file on disk without loading the data into memory. This module extracts a one or more vectors from a matrix stored on disk. There are three methods of determining which columns or rows need to be extracted:
(1) One can use the movie buttons in the GUI to set a start and end index and as well the current index. One can automatically step through the different vectors using the buttons in the GUI. The GUI also allows for setting the timing of reloading a new vector. This way one can select only one vector at a time.
(2) One can supply a matrix with the indices of the vectors to extract. This can be one index or can be multiple. The matrix with the indices need to be connected to the 'Indices' input port. The order of the indices is kept when assigning them in the output matrix.
(3) One can supply weights that can are used to extract columns or rows and then sum them together using the assigned weights. When using the weights, one does not need to supply indices. The indices of the rows or columns to be retrieved from disk are the columns in the weights matrix with a weight assigned. To obtain multple weighted outputs, one can assign a weights matrix with multiple rows. Internally in the module, the weights matrix will be translated into a sparse matrix. For optimal performance one should input a sparse matrix. This option is intended for mapping matrices, where one slices a small piece of a large mesh and wants to extract the signals assigned to that part of the mesh.
The module outputs the vector or matrix with the selected columns or rows. The module also outputs the selected indices. This port can be used to slave other StreamMatrixFromDisk modules to read the same indexed data from different data files. The module also output the scaled index, which is the index multiplied with the spacing information stored in the data file. This port can be used to extract the time from the file automatically.
The file with the streaming data needs to be in the uncompressed NRRD format. This format uses a text file to describe a data set distributed over multiple files. See teem.sourceforge.net for a description of the NRRD file format.
Frequently Asked Questions
Known Bugs
Recent Changes
Go back to Documentation:SCIRun:Reference:ModelCreation
