Making Bitmaps from Vector Fields: avf2ppm

The avf2ppm utility converts a collection of vector field files (e.g., .omf, .ovf) into color bitmaps suitable for inclusion into documents or collating into movies. The command line arguments control filename and format selection, while plain-text configuration files, modeled after the mmDisp configuration dialog box, specify conversion parameters.

Launching
The avf2ppm launch command is:

tclsh oommf.tcl avf2ppm [standard options] [-config file] [-f] \
   [-filter program] [-format <P3|P6|B24>] [-ipat pattern] \
   [-opatexp regexp] [-opatsub sub] [-v level] [infile ...]

where

-config file

User configuration file that specifies the image conversion parameters. This file is discussed in detail below.

-f

Force overwriting of existing (output) files. By default, if avf2ppm tries to create a file, say foo.ppm, that already exists, it generates instead a new name of the form foo.ppm-000, or foo.ppm-001, ..., or foo.ppm-999, that doesn't exist and writes to that instead. The -f flag disallows alternate filename generation, and overwrites foo.ppm instead.

-filter program

Post-processing application to run on each app2ppm output file. May be a pipeline of many programs.

-format <P3|P6|B24>

Specify the output image file format. Currently supported formats are the true color Portable Pixmap (PPM) formats P3 (ASCII text) and P6 (binary), and the uncompressed Windows BMP 24 bits-per-pixel format. The default is P6.

-ipat pattern

Specify input files using a pattern including ``glob-style'' wildcards. Mostly useful in DOS.

-opatexp regexp

Specify the ``regular expression'' applied to input filenames to determine portion to be replaced in generation of output filenames. Default: (\.[^.]?[^.]?[^.]?$|$)

-opatsub sub

The string with which to replace the portion of input filenames matched by the -opatexp regexp during output filename generation. The default is .ppm for type P3 and P6 file output, .bmp for B24 file output.

-v level

Verbosity (informational message) level, with 0 generating only error messages, and larger numbers generating additional information. The level value is an integer, defaulting to 1.

infile ...

List of input files to process.

The file specification options require some explanation. Input files may be specified either by an explicit list (infile ...), or by giving a wildcard pattern, e.g., -ipat *.omf, which is expanded in the usual way by avf2ppm (using the Tcl command glob). Unix shells (sh, csh, etc.) automatically expand wildcards before handing control over to the invoked application, so the -ipat option is not needed (although it is useful in case of a ``command-line too long'' error). DOS does not do this expansion, so you must use -ipat to get wildcard expansion in Windows.

As each input file is processed, a name for the output file is produced from the input filename by rules determined by handing the -opatexp and -opatsub expressions to the Tcl regsub command. Refer to the Tcl regsub documentation for details, but essentially whatever portion of the input filename is matched by the -opatexp expression is removed and replaced by the -opatsub string. The default -opatexp expression matches against any filename extension of up to 3 characters, and the default -opatsub string replaces this with the extension .ppm or .bmp.

If you have command line image processing ``filter'' programs, e.g., ppmtogif (part of the NetPBM package), then you can use the -filter option to pipe the output of avf2ppm through that filter before it is written to the output file specified by the -opat* expressions. If the processing changes the format of the file, (e.g., ppmtogif converts from PPM to GIF), then you will likely want to specify a -opatsub different from the default.

Here is an example that processes all files with the .omf extension, sending the output through ppmtogif before saving the results in a files with the extension .gif:

tclsh oommf.tcl avf2ppm -ipat *.omf -opatsub .gif -filter ppmtogif

(On Unix, either drop the -ipat flag, or use quotes to protect the input file specification string from expansion by the shell, as in -ipat '*.omf'.) You may also pipe together multiple filters, e.g., -filter "ppmquant 256 | ppmtogif".

Configuration files
The details of the conversion process are specified by plain-text configuration files, with fields analogous to the entries in the mmDisp configuration dialog box. Each of the parameters is an element in an array named plot_config. The default values for this array are taken from the default configuration file oommf/app/mmdisp/scripts/avf2ppm.def, which is a Tcl script read during avf2ppm initialization.

There are several places in the configuration file where colors are specified. Colors may be represented using the symbolic names in oommf/config/colors.def, in any of the Tk hexadecimal formats, e.g., #RRGGBB, or as a shade of gray using the format ``grayD'' (or ``greyD''), where D is a decimal integer from 0-100, inclusive. Examples in the latter two formats are #FFFF00 for yellow, gray0 for black and gray100 or #FFFFFF for white.

The sample default configuration script can be used as a starting point for user (per-run) configuration files. Refer to this sample file and the mmDisp documentation as we discuss each element of the array plot_config. (See the Tcl documentation for details of the array set command.)

colormaps

A list of valid colormaps known to the program. This entry is not user-configurable, and should not appear in user configuration files.

arrow,status

Set to 1 to display arrows, 0 to not draw arrows.

arrow,colormap

Select the colormap to use when drawing arrows. Should be one of the strings specified in the colormaps array element.

arrow,colorcount

Number of discretization levels to use from the colormap. A value of zero will color all arrows with the first color in the colormap.

arrow,quantity

Scalar quantity the arrow color is to represent. Supported values include x, y, z, xy-angle, xz-angle, yz-angle, and slice. The mmDisp configuration dialog box will present the complete list of allowed quantities (which may be image dependent).

arrow,colorphase

The phase is a real number between -1 and 1 that provides a translation in the selected colormap. For the xy-angle, xz-angle and yz-angle color quantities, this displays as a rotation of the colormap, for example, setting colorphase to 0.333 would effectively change the Red-Green-Blue-Red colormap into Green-Blue-Red-Green. For the other color quantities, it simply shifts the display band, saturating at one end. For example, setting colorphase to 0.5 changes the Blue-White-Red colormap into White-Red-Red.

arrow,colorinvert

The colorinvert value should be 1 or 0, signifying to invert or not invert, respectively. If invert is selected, the the colormap ordering is reversed, changing for example Blue-White-Red into Red-White-Blue. If both inversion and phase adjustment are requested, then inversion is applied first.

arrow,autosample

If 1, then ignore the value of arrow,subsample and automatically determine a ``reasonable'' subsampling rate for the arrows. Set to 0 to turn off this feature.

arrow,subsample

If arrow,autosample is 0, then subsample the input vectors at this rate when drawing arrows. A value of 0 for arrow,subsample is interpreted specially to display all data.

arrow,size

Size of the arrows relative to the default size (represented as 1.0).

arrow,antialias

If 1, then each pixel along the edge of an arrow is drawn not with the color of the arrow, but with a mixture of the arrow color and the background color. This makes arrow boundaries appear less jagged, but increases computation time. Also, the colors used in the anti-aliased pixels are not drawn from the arrow or pixel colormap discretizations, so color allocation in the output bitmap may increase dramatically.

pixel,...

Most of the pixel configuration elements have analogous arrow configuration elements, and are interpreted in the same manner. The exceptions are the pixel,opaque element which is discussed next, the arrow,antialias element which has no corresponding pixel element, and the qualitative difference in that the auto subsampling rate for pixels is considerably denser than for arrows.

pixel,opaque

If the opaque value is 1, then the pixel is drawn in a solid manner, concealing any arrows which are drawn under it. If opaque is 0, then the pixel is drawn only partially filled-in, so objects beneath it can still be discerned.

misc,background

Specify the background color.

misc,drawboundary

If 1, then draw the bounding polygon, if any, as specified in the input vector field file.

misc,boundarywidth

Width of the bounding polygon, in pixels. A value of 0 is treated the same as 1. To disable drawing of the boundary, use the misc,drawboundary option.

misc,boundarycolor

Color of the bounding polygon.

misc,boundarypos

Placement of the bounding polygon, either back or front, i.e., either behind or in front of the rendered arrows and pixel elements.

misc,matwidth

Specifies the width, in pixels, of a mat (frame) around the outer edge of the image. The mat is drawn in front of all other objects. To disable, set matwidth to 0.

misc,matcolor

Color of the mat.

misc,margin

The size of the border margin, in pixels.

misc,width, misc,height

Maximum width and height of the output bitmap, in pixels. If misc,crop is enabled, then one or both of these dimensions may be shortened.

misc,crop

If disabled (0), then any leftover space in the bitmap (of dimensions misc,width by misc,height) after packing the image are filled with the background color. If enabled (1), then the bitmap is cropped to just include the image (with the margin specified by misc,margin). NOTE: Some movie formats require that bitmap dimensions be multiples of 8 or 16. For such purposes, you should disable misc,crop and specify appropriate dimensions directly with misc,width and misc,height.

misc,zoom

Scaling factor for the display. This is the same value as shown in the ``zoom'' box in the mmDisp control bar, and corresponds roughly to the number of pixels per vector in the (original, fully-sampled) vector field. If set to zero, then avf2ppm will automatically set the scaling so the image (with margins) just fits inside the area specified by misc,width and misc,height.

misc,rotation

Counterclockwise rotation in degrees; either 0, 90, 180 or 270.

misc,datascale

Scale for arrow size and colormap range; this element is exactly equivalent to the mmDisp Data Scale control. In general, this should be a positive real value, but a zero or empty value will set the scaling to its default value.

misc,centerpt

This element should either be empty, or else be a three item list of real numbers specifying the center of the display, in input file mesh units. The OOMMF command line program avf2ovf may be invoked with the -info option to determine the range of the file mesh. This information is also contained in the ``Bounding box'' message returned when avf2ppm is run, which should in any event be consulted since any portion of the mesh that is shifted outside the bounding box will not be rendered. This element provides the functionality of the scroll bars on the mmDisp main display window. If an empty value is given, then by default the center of the mesh is used.

viewaxis

Select the view axis, which should be one of +z, -z, +y, -y, +x, or -x. This option is equivalent to the mmDisp View|Viewpoint menu control.

viewaxis,center

The viewaxis,center element corresponds to the mmDisp ``slice'' selection control. It should be a single real number in the range of that portion of the input file mesh bounding box corresponding to the selected view axis. For example, if the view axis is +z or -z, then the view axis center value should lie between the minimum and maximum z-values of the file mesh. If an empty value is given, then the view axis center is set from the corresponding coordinate of the misc,centerpt value.

The portion of the mesh that is rendered depends upon the interaction between the bounding box, the center point, the view axis center, and the axis-selected arrow and pixel spans.

viewaxis,xarrowspan, viewaxis,yarrowspan, viewaxis,zarrowspan

Specifies the thickness of the arrow display slice, for the corresponding view, e.g., if the view axis is +z or -z, then only viewaxis,zarrowspan is active. These values for each of these elements should be either a single real value in units relating to the corresponding coordinate range of the input file mesh, or else an empty string, which is interpreted to be the entire thickness of the mesh in the selected view direction.

viewaxis,xpixelspan, viewaxis,ypixelspan, viewaxis,zpixelspan

Identical interpretation and behavior as the corresponding arrow span elements, but with regards to pixel display.

array set plot_config {
    colormaps  {
        Red-Black-Blue     Blue-White-Red      Teal-White-Red
        Black-Gray-White   Black-White-Black   White-Black-White
        White-Green-Black  Red-Green-Blue-Red
    }
    arrow,status       1               misc,background     #FFFFFF
    arrow,colormap  Black-Gray-White   misc,drawboundary   1
    arrow,colorcount   0               misc,boundarywidth  1
    arrow,quantity     z               misc,boundarycolor  #000000
    arrow,colorphase   0.              misc,boundarypos    front
    arrow,colorinvert  0               misc,matwidth       0
    arrow,autosample   1               misc,matcolor       #FFFFFF
    arrow,subsample    10              misc,margin         10
    arrow,size         1               misc,width          640
    arrow,antialias    1               misc,height         480
                                       misc,crop           1
    pixel,status       1               misc,zoom           0
    pixel,colormap  Teal-White-Red     misc,rotation       0
    pixel,colorcount   225             misc,datascale      0
    pixel,opaque       1               misc,centerpt       {}
    pixel,quantity     x
    pixel,colorphase   0.              viewaxis            {+z}
    pixel,colorinvert  0               viewaxis,center     {}
    pixel,autosample   1               viewaxis,xarrowspan {}
    pixel,subsample    0               viewaxis,xpixelspan {}
    pixel,size         1               viewaxis,yarrowspan {}
                                       viewaxis,ypixelspan {}
                                       viewaxis,zarrowspan {}
                                       viewaxis,zpixelspan {}
}

Figure 4: Sample default configuration script avf2ppm.def. (Description.)

User (per-run) configuration files are specified on the command line with the -config option. To create a user configuration file, make a copy of the default avf2ppm.def configuration file, and edit it as desired in a plain text editor. You may omit any entries that you do not want to change from the default. (Each entry consists of a name + value pair, e.g., misc,width 640.) You may ``layer'' configuration files by specifying multiple user configuration files on the command line. These are processed from left to right, with the last value set for each entry taking precedence.

OOMMF Documentation Team
October 30, 2002