vdfcreate - Generate a VDF metadata file
vdfcreate [options] vdffile
vdfcreate generates a .vdf metafile that will define a VAPOR Data Collection (VDC). A VDC consists of a grouping of related 3D and 2D data volumes, most typically these field variables are solution outputs to a single numerical simulation. The metadata contents of a .vdf file establishes common attributes for all data contained in a VDC (e.g. volume dimension) as well as providing information unique to each data volume contained in a VDC (e.g. path names to files containing a particular solution variable).
Data volumes in a VDC are represented hiearchically: each volume may be accessed at varying, power-of-two coarsened approximations. Moreover, two types of VDC files are supported: VDC Type 1 (VDC1) and VDC Type 2 (VDC2). The latter, VDC2, supports level-of-detail access in addtion to hierarchical access. The default type is VDC1.
This option (VDC2 only) specifies a colon delimited list of compression ratios. The minimum compression ratio is 1 (indicating no compression). The maximum is determined by the wavelet familiy (See -wname) and the block dimension (See -bs). The default compression ratio vector is 500:100:10:1. If a compression ratio of 1-to-1 is not included it will be silently added.
The -dimension option specifies the volumetric dimensions of all 3D elements of the VDC. NX is an integer specifying the bounds (in grid points) of the fastest varying dimension. NY specifies the bounds of the second fastest varying dimension, and so on.
Starting time in user time coordinates. This option specifies the time in the user defined coordinate system for the first time step in the VDC. Correct specification of user time for each time step is necessary for accurate time-dependent fields (e.g. velocity).
Time in user coordinates between successive time steps. This option specifies a uniform, sampling interval between successive time steps in the user coordinate system for time.
This option specifies the maximum number of time steps that the VDC may contain. A VDC may contain fewer time steps than what is specified by the .vdf file, but not more. The default number of time steps is one.
The internal storage blocking factor expressed in voxels. In general this should not be changed from the default value (32x32x32 for VDC1, and 64x64x64 for VDC2).
This option specifies the number of refinement levels in a Type 1 VDC data approximation hierarchy. Starting with the native grid resolution, each successive level in the hierarchy is a factor of two smaller along each dimension. If level is 0 no hiearchy will be created (all data will be stored in the VDC at their native resolution). If level is 1 a single approximation will be created, thus the hierachy will have two levels: the native grid resolution, and the first approximation. And so on.
This option specifies a map projection using a PROJ projection string. This string must be enclosed in quotes, and consists of several keyword/value pairs of the form "+key=value". The full set of PROJ key/value pairs can be found at the PROJ wiki http://trac.osgeo.org/proj/wiki . For example, a Cylindrical Equidistant projection on a spherical earth is specified as "+proj=eqc +ellps=WGS84". VAPOR supports Lambert conformal conic, Mercator, Longitude/Latitude, and polar stereographic projections. The PROJ string is not checked for validity.
Specifies the number of wavelet filter coefficients to be used when constructing a Lifting wavelet for transforming data into the VDC. In general this option should not be changed from the default value, one. VDC1 only.
Specifies the number of wavelet lifting coefficients to be used when constructing a Lifting wavelet for transforming data into the VDC. In general this option should not be changed from the default value, one. VDC1 only.
Specifies a comment to be included in the .vdf file.
Specifies the type of the volumetric data grid. Currently supported values for type are: regular, layered, stretched, and block_amr. The default is regular, a uniformly sampled Cartesian grid.
Specifies the path to a file containing a whitespace-delimited list of monotonically increasing user times. Each user time is a unitless floating point value indicating the time associated with a time step. User times are required for correct calculation of time-dependent algorithms such as unsteady flow visualization. If this option is present, the -numts, -deltat, and startt options are ignored.
Specifies the data coordinate system. Currently support values for type are cartesian and spherical. The default is cartesian.
The option specifies a colon-delimited, 6-element vector defining the spatial domain extents of all 3D data volumes in a user defined coordinate system. X0, Y0, and Z0 provided the minimum bounds of the fastest to slowest varying dimensions, respectively. X1, Y1, and Z1 provided the maximum bounds of the fastest to slowest varying dimensions, respectively. Correct domain extents are required for some visualization algorithms such as flow integration. For Cartesian data, if this option is not supplied domain extents will be calculated such that the largest dimension (in grid points) has extents [0..1.0]. The other extents are caluclated such that the minimum bound is 0.0 and the maximum bound assures cubic voxels. For spherical data, the defaults are calculated as:
X0 = 0.0 + (360.0 / NX / 2.0) # min_lon
X1 = 360 - (360 / NX/ 2.0) # max_lon
Y0 = 0.0 + (180 / NY / 2.0) # min_lat
Y1 = 360 - (180 / NY / 2.0) # min_lat
Z0 = 0.0 # min_radius
Z1 = 0.5 # max_radius
where NX and NY are the dimension, in grid points, of the longitude and latitude coordinates, respectively.
A colon-delimited, 3-element vector specifying the permutation ordering of raw data on disk. The default permutation 0:1:2, in the case of spherical data, for example, indicates that Longitude is the fastest varying dimension, followed by lattitude, then radius. The permutation specification 2:1:0 would swap Longitude and Radius dimensions. Note, that this option not only permutes the expected ordering of data, but also permutes option arguments to vdfcreate that specify dimensions. This includes the -dimension, -extents, and -periodic command line options.
A colon-delimited, 3-element boolean vector (0==false, 1==true) specifying the periodicity of the X, Y, and Z coordinate axes. A true value, 1, indicates that the coresponding axis has periodic boundaries. The default is 0:0:0, indicating that all of the boundaries are non-periodic.
This option is deprecated. Use -vars3d.
A colon delimited list of 3D field variable names to be included in the VDF. The default is the single element vector, var1.
A colon delimited list of 2D, XY-plane field variable names to be included in the VDF. The default is to include no 2D XY-plane variables.
A colon delimited list of 2D, XZ-plane field variable names to be included in the VDF. The default is to include no 2D XZ-plane variables.
A colon delimited list of 2D, YZ-plane field variable names to be included in the VDF. The default is to include no 2D YZ-plane variables.
Generate a .vdf file for a VDC Type 2 data collection.
Each option, -xcoords, -ycoords, or -zcoords, specifies the path to a file containing a whitespace-delimited list of monotonically increasing (or decreasing) user coordinates for the X, Y, or Z axis, respectively. Each coordinate is a unitless floating point value indicating the user coordinate of the i'th grid point along the X, Y, or Z axis. For any axis for which a coordinate path file is not specified, the grid coordinates are assumed to have uniform spacing and are computed as described the the -extents option. This option is ignored unless the grid type specified with -gridtype is stretched.
Wavelet family name used for compression (VDC2 only). See the vdfcreate help message for a list of supported wavelets.
Print a usage statement and then exit.
The command
vdfcreate -dimension 512x512x512 -numts 100 -level 3 -varnames vx:vy:vz foo.vdf
will produce a VDC Type 1 VDF file named, foo.vdf. This file will describe a VDC containing up to 100 time steps, starting from 0 and running through 99; three field variables, vx, vy, and vz; each volume would have a spatial resolution of 512^3; and three wavelet transformations would be applied resulting in data approximations at resolutions 64^4, 128^3, and 256^3.
vdfcreate, raw2vdf, vdf2raw
An Overview of VAPOR Data Collections
Last updated on $Date$