Grace is a WYSIWYG tool to make two-dimensional plots of numerical data. It runs under various (if not all) flavors of Unix with X11 and M*tif (LessTif or Motif). It also runs under VMS, OS/2, and Windows (95/98/NT/2000/XP). Its capabilities are roughly similar to GUI-based programs like Sigmaplot or Microcal Origin plus script-based tools like Gnuplot or Genplot. Its strength lies in the fact that it combines the convenience of a graphical user interface with the power of a scripting language which enables it to do sophisticated calculations or perform automated tasks.
Grace is derived from Xmgr (a.k.a. ACE/gr), originally written by Paul Turner.
From version number 4.00, the development was taken over by a team of volunteers under the coordination of Evgeny Stambulchik. You can get the newest information about Grace and download the latest version at the Grace home page.
When its copyright was changed to GPL, the name was changed to Grace, which stands for ``GRaphing, Advanced Computation and Exploration of data'' or ``Grace Revamps ACE/gr''. The first version of Grace available is named 5.0.0, while the last public version of Xmgr has the version number 4.1.2.
Paul still maintains and develops a non-public version of Xmgr for internal use.
Copyright (©) 1991-1995 Paul J Turner, Portland, OR
Copyright (©) 1996-2007 Grace Development Team
Maintained by Evgeny Stambulchik
All Rights Reserved
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
For certain libraries required to build Grace (which are therefore even included in a suitable version) there may be different Copyright/License statements. Though their License may by chance match the one used for Grace, the Grace Copyright holders can not influence or change them.
Package | License |
cephes library | Free |
T1lib | LGPL |
Xbae | BSD-like |
Tab Widget | BSD-like |
yacc
or, better,
bison
)./usr/local/src/grace-x.y.z
and the compilation
will be performed in /tmp/grace-obj
, do the following:
% mkdir /tmp/grace-obj % cd /tmp/grace-obj % /usr/local/src/grace-x.y.z/ac-tools/shtool mkshadow \ /usr/local/src/grace-x.y.z .
configure
shell script attempts to guess correct
values for various system-dependent variables used during
compilation. It uses those values to create Make.conf
in the
top directory of the package. It also create config.h
file
containing system-dependent definitions. Finally, it creates a shell
script config.status
that you can run in the future to
recreate the current configuration, a file config.cache
that
saves the results of its tests to speed up reconfiguring, and a file
config.log
containing compiler output (useful mainly for
debugging configure
). If at some point config.cache
contains results you don't want to keep, you may remove or edit it../configure --help
to get list of additional switches specific to Grace./configure <options>
. Just an example:
% ./configure --enable-grace-home=/opt/grace --with-extra-incpath=/usr/local/include:/opt/include \ --with-extra-ldpath=/usr/local/lib:/opt/lib --prefix=/usrwould use
/usr/local/include
and
/opt/include
in addition to the default include path
and /usr/local/lib
and /opt/lib
in addition
to the default ld path. As well, all stuff would be put under
the /opt/grace directory and soft links made to
/usr/bin
, /usr/lib
and /usr/include
.
Note: If you change one of the
--with-extra-incpath
or
--with-extra-ldpath
options from one run of
configure to another, remember to delete the
config.cache
file!!!
make
If something goes wrong, try to see if the problem has been
described already in the Grace FAQ (in the
doc
directory).
make tests
This will give you a slide show demonstrating some nice features of Grace.
make install
make links
The later (optional) step will make soft links from some files
under the Grace home directory to the system-wide default
locations (can be changed by the --prefix
option
during the configuration, see above).
Not written yet...
For a jump-in start, you can browse the demos ("Help/Examples" menu tree). These are ordinary Grace projects, so you can play with them and modify them. Also, read the Tutorial.
O.k. Here's a VERY quick introduction:
A project file contains all information necessary to restore a plot created by Grace, as well as some of preferences. Each plot is represented on a single page, but may have an unlimited number of graphs.You create a project file of your current graph with File/Save,Save as.
A parameter file contains the detailed settings of your project. It can be used to transfer these settings to a different plot/project. You generate a parameter file with File/Save menu entry selected from the "Plot/Graph appearance popup". You can load the settings contained in a parameter file with File/Open.
Grace understands several input files formats. The most basic one is ASCII text files containing space and comma separated columns of data. The data fields can be either numeric (Fortran 'd' and 'D' exponent markers are also supported) or alphanumeric (with or without quotes). Several calendar date formats are recognized automatically and you can specify your own reference for numeric date formats. Lines beginnig with "#" are ignored. Blank lines indicate new dataset. Grace also has a command language (see command interpreter), you can include commands in data files using lines having "@" as their first non-blank character, though this is not recommended. Depending on configuration, Grace can also read NetCDF files (see configuration).
A graph consists of (every element is optional): a graph frame, axes, a title and a subtitle, a number of sets and additional annotative objects (time stamp string, text strings, lines, boxes and ellipses).
The idea of "XY Chart" is to plot bars (or symbols in general) of several sets side by side, assuming the abscissas of all the sets are the same (or subsets of the longest set).
A dataset is a collection of points with x and y coordinates, up to four optional data values (which, depending on the set type, can be displayed as error bars or like) and one optional character string.
A set is a way of representing datasets. It consists of a pointer to a dataset plus a collection of parameters describing the visual appearance of the data (like color, line dash pattern etc).
The set type can be any of the following:
Set type | # of num. cols | Description |
XY | 2 | An X-Y scatter and/or line plot, plus (optionally) an annotated value |
XYDX | 3 | Same as XY, but with error bars (either one- or two-sided) along X axis |
XYDY | 3 | Same as XYDX, but error bars are along Y axis |
XYDXDX | 4 | Same as XYDX, but left and right error bars are defined separately |
XYDYDY | 4 | Same as XYDXDX, but error bars are along Y axis |
XYDXDY | 4 | Same as XY, but with X and Y error bars (either one- or two-sided) |
XYDXDXDYDY | 6 | Same as XYDXDY, but left/right and upper/lower error bars are defined separately |
BAR | 2 | Same as XY, but vertical bars are used instead of symbols |
BARDY | 3 | Same as BAR, but with error bars (either one- or two-sided) along Y axis |
BARDYDY | 4 | Same as BARDY, but lower and upper error bars are defined separately |
XYHILO | 5 | Hi/Low/Open/Close plot |
XYZ | 3 | Same as XY; makes no sense unless the annotated value is Z |
XYR | 3 | X, Y, Radius. Only allowed in Fixed graphs |
XYSIZE | 3 | Same as XY, but symbol size is variable |
XYCOLOR | 3 | X, Y, color index (of the symbol fill) |
XYCOLPAT | 4 | X, Y, color index, pattern index (currently used for Pie charts only) |
XYVMAP | 4 | Vector map |
XYBOXPLOT | 6 | Box plot (X, median, upper/lower limit, upper/lower whisker) |
Not all set types, however, can be plotted on any graph type. The following table summarizes it:
Set type | XY Graph | XY Chart | Fixed | Polar | Pie |
XY | + | + | + | + | + |
XYDX | + | - | + | - | - |
XYDY | + | + | + | - | - |
XYDXDX | + | - | + | - | - |
XYDYDY | + | + | + | - | - |
XYDXDY | + | - | + | - | - |
XYDXDXDYDY | + | - | + | - | - |
BAR | + | + | + | - | - |
BARDY | + | + | - | - | - |
BARDYDY | + | + | - | - | - |
XYHILO | + | - | - | - | - |
XYZ | + | - | + | + | - |
XYR | - | - | + | - | - |
XYSIZE | + | + | + | + | - |
XYCOLOR | + | + | + | + | + |
XYCOLPAT | - | - | - | - | + |
XYVMAP | + | - | + | - | - |
XYBOXPLOT | + | - | - | - | - |
Regions are sections of the graph defined by the interior or exterior of a polygon, or a half plane defined by a line. Regions are used to restrict data transformations to a geometric area occupied by region.
Real Time Input refers to the ability Grace has to be fed in real time by an external program. The Grace process spawned by the driver program is a full featured Grace process: the user can interact using the GUI at the same time the program sends data and commands. The process will adapt itself to the incoming data rate.
Hotlinks are sources containing varying data. Grace can be instructed a file or a pipe is a hotlink in which case it will provide specific commands to refresh the data on a mouse click (a later version will probably allow automatic refresh).
Grace allows the user to choose between several output devices to produce its graphics. The current list of supported devices is:
Note that Grace no longer supports GIF due to the copyright policy of Unisys. Grace can also be instructed to launch conversion programs automatically based on file name. As an example you can produce MIF (FrameMaker Interchange Format) or Java applets using pstoedit, or almost any image format using the netpbm suite (see the FAQ).
In many cases, when Grace needs to access a file given with a
relative pathname
, it searches for the file along the
following path:
./pathname:./.grace/pathname:~/.grace/pathname:$GRACE_HOME/pathname
Grace can access external functions present in either system or third-party shared libraries or modules specially compiled for use with it. The term dynamic refers to the possibility Grace has to open the library at run time to find the code of the external function, there is no need to recompile Grace itself (the functions already compiled in Grace are "statically linked").
There are two types of coordinates in Grace: the world coordinates and the viewport coordinates. Points of data sets are defined in the world coordinates. The viewport coordinates correspond to the image of the plot drawn on the canvas (or printed on, say, PS output page). The transformation converting the world coordinates into the viewport ones is determined by both the graph type and the axis scaling.
Actually, there is yet another level in the hierarchy of coordinates - the device coordinates. However, you (as a user of Grace) should not worry about the latter. The mapping between the viewport coordinates and the device coordinates is always set in such a way that the origin of the viewport corresponds to the left bottom corner of the device page, the smallest of the device dimensions corresponds to one unit in the viewport coordinates. Oh, and the most important thing about the viewport -> device transformation is that it is homotetic, i.e. a square is guaranteed to remain a square, not a rectangle, a circle remains a circle (not an ellipse) etc.
With respect to the user interface, there are three modes of
operation that Grace can be invoked in. The full-featured GUI-based
version is called xmgrace
. A batch-printing version is
called gracebat
. A command-line interface mode is called
grace
. Usually, a single executable is called in all cases,
with two of the three files being (symbolic) links to a "real" one.
Override any parameter file settings
Turn off all toolbars
Execute batch_file on start up (i.e., after all other options have been processed and the UI initialized)
Assume data file is block data
Form a set from the current block data set using the current set type from columns given in the argument
Set the hint for dates analysis
Read data from descriptor (anonymous pipe) on startup
Set canvas size fixed to width*height
Use free page layout
Set the current graph number
Set the type of the current graph
No interactive session, just print and quit
Set default hardcopy device
Install private colormap
Turn the graph legend on
Set the axis scaling of the current graph to logarithmic
Set the maximal drawing path length
Run Grace in monochrome mode (affects the display only)
Assume data file is in netCDF format. This option is present only if the netCDF support was compiled in
If -netcdf was used previously, read from the netCDF file X_var Y_var variables and create a set. If X_var name is "null" then load the index of Y to X. This option is present only if the netCDF support was compiled in
Assume the answer is yes to all requests - if the operation would overwrite a file, Grace will do so without prompting
Don't use private colormap
In batch mode, do not print
Disable safe mode
Don't catch signals
Read data from named pipe on startup
Assume data file is in X Y1 Y2 Y3 ... format
Load parameters from parameter_file to the current graph
Interpret string as a parameter setting
Read data from stdin on startup
file Save print output to file
Remove data file after read
Write results of some data manipulations to results_file
Exchange the color indices for black and white
Run in the safe mode (default) - no file system modifications are allowd through the batch language
Save all graphs to save_file
Integer seed for random number generator
Set the type of the next data file
Source type of next data file
Set allowed time slice for real time inputs to delay ms
Add timestamp to plot
Show the program version
Set the viewport for the current graph
Set the working directory
Set the world coordinates for the current graph
This message
Set the location of Grace. This will be where help files, auxiliary programs, and examples are located. If you are unable to find the location of this directory, contact your system administrator.
Print command. If the variable is defined but is an empty string, "Print to file" will be selected as default.
The editor used for manual editing of dataset values.
The shell command to run an HTML viewer for on-line browsing of the help documents. Must include at least one instance of "%s" which will be replaced with the actual URL by Grace.
These flags control behavior of the FFTW planner (see FFTW tuning for detailed info)
Upon start-up, Grace loads its init file, gracerc
. The file
is searched for in the magic path (see
magic path); once found, the rest of the
path is ignored. It's recommended that in the gracerc
file,
one doesn't use statements which are part of a project file - such
defaults, if needed, should be set in the default template (see
default template).
Whenever a new project is started, Grace loads the default template,
templates/Default.agr
. The file is searched for in the magic
path (see
magic path); once found, the
rest of the path is ignored. It's recommended that in the default
template, one doesn't use statements which are NOT part of a project
file - such defaults, if needed, should be set in the
gracerc
(see
init file).
The following Grace-specific X resource settings are supported:
It is also possible to customize menus by assigning key accelerators to any item.
You'll need to derive the item's X resource name from the respective menu label, which is easily done following these rules:
For example, in order to make Grace popup the Non-linear curve fitting by pressing Control+F, you would add the following two lines
XMgrace*transformationsMenu.nonLinearCurveFittingButton.acceleratorText: Ctrl+F
XMgrace*transformationsMenu.nonLinearCurveFittingButton.accelerator: Ctrl<Key>f
to your .Xresources
file (the file which is read when an X
session starts; it could be .Xdefaults
, .Xsession
or
some other file - ask your system administrator when in doubt).
Similarly, it may be desirable to alter default filename patterns of file selection dialogs. The recipe for the dialog's name is like for menu buttons outlined above, with "Button" being replaced with "FSB". E.g., to list all files in the "Open project" dialog ("File/Open..."), set the following resource:
XMgrace*openProjectFSB.pattern: *
This section describes interface controls - basic building blocks, used in many popups.
Whenever the user is expected to provide a filename, either for reading in or writing some data, a file selection dialog is popped up. In addition to the standard entries (the directory and file lists and the filter entry), there is a pulldown menu for quick directory change to predefined locations (the current working directory, user's home directory and the file system root). Also, a "Set as cwd" button is there which allows to set any directory as you navigate through the directory tree as the current working directory (cwd). Once defined, it can be used in any other file selection dialog to switch to that directory quickly.
Various selectors are available in several popups. They all display lists of objects (graphs, sets, ...) and can be used to perform simple operations on these objects (copying, deleting, ...). The operations are available from a popup menu that appears when pressing mouse button 3 on them. Depending on the required functionality, they may allow multiple choices or not. The following shortcuts are enabled (if the result of an action would contradict the list's selection policy, this would be ignored):
The operations that can be performed on graphs through the graph selector's popup menu are:
Double-clicking on a list entry will switch the focus to that graph.
The operations that can be performed on sets through the set selector's popup menu are:
Double-clicking on a list entry will open the spreadsheet editor (see Spreadsheet data set editor) on the set data.
When the pointer focus is on the canvas (where the graph is drawn), there are some shortcuts to activate several actions. They are:
A single click inside a graph switches focus to that graph. This is the default policy, but it can be changed from the "Edit/Preferences" popup.
Double clicking on parts of the canvas will invoke certain actions or raise some popups:
The double clicking actions can be enabled/disabled from the "Edit/Preferences" popup.
Along the left-hand side of the canvas (if shown) is the ToolBar. It is armed with several buttons to provide quick and easy access to the more commonly used Grace functions.
Draw
: This will redraw the canvas and sets.
Useful if "Auto Redraw" has been deselected in the Edit|Preferences
dialog or after executing commands directly from the Window|Commands
interpreter.
Lens
: A zoom lens. Click on the lens, then select the
area of interest on the graph with the "rubber band". The region
enclosed by the rubber band will fill the entire graph. AS
: AutoScale. Autoscales the graph to contain all
data points of all visible (not hidden) sets. Z/z
: Zoom in/out by 5%. The zoom percentage can be
set in the Edit/Preferences dialog.Arrows
: Scroll active graph by 5% in the arrow's
direction. The scroll percentage can be set in the
Edit/Preferences dialog.
AutoT
: AutoTick Axes. This will find the optimum
number of major and minor tick marks for both axes.AutoO
: Autoscale On set. Click the AutoO
button, then click on the graph near the set you wish to use for
determining the autoscale boundaries of the graph.
ZX,ZY
: Zoom along an axis. These buttons work like the
zoom lens above but are restricted to a single axis. AX,AY
: Autoscale one axis only.
The following buttons deal with the graph stack and there is a good example under Help/Examples/General Intro/World Stack.
Pu/Po
: Push and pop the current world settings to/from
the graph stack. When popping, makes the new stack top current.PZ
: Push before Zooming. Functions as the zoom lens,
but first pushes the current world settings to the stack. Cy
: Cycles through the stack settings of the active
graph. Each graph may have up to twenty layers on the stack.
Exit
: Pretty obvious, eh?The file menu contains all entries related to the input/output features of Grace.
Reset the state of Grace as if it had just started (one empty graph ranging from 0 to 1 along both axes). If some work has been done and not yet saved, a warning popup is displayed to allow canceling the operation.
Open an existing project file. A popup is displayed that allow to browse the file system.
Save the current work in a project file, using the name that was used for the last open or save. If no name has been set (i.e., if the project has been created from scratch) act as save as.
Save the current work in a project file with a new name. A popup allows to browse the file system and set the name, the format to use for saving data points (the default value is "%16.8g"), and a textual description of the project. A warning is displayed if a file with the same name already exists.
Abandon all modifications performed on the project since the last save. A confirmation popup is fired to allow the user canceling the operation.
Set the properties of the printing device. Each device has its own set of specific options (see Device-specific settings). According to the device, the output can be sent either directly to a printer or directed to a file. The global settings available for all devices are the sizing parameters. The size of the graph is fixed. Changing the 'Page' settings changes the size of the canvas underneath the graph. Switching between portrait and landscape rotates the canvas. Make sure the canvas size is large enough to hold your graph. Otherwise you get a 'Printout truncated' warning. If your canvas size cannot easily be changed because, for example, you want to print on letter size paper, you need to adjust the size of your graph ('Viewport' in Plot/Graph Appearance).
Print the project using the current printer settings
Exit from Grace. If some work has been done and not saved, a warning popup will be displayed to allow the user to cancel the operation.
Using the data set popup, you can view the properties of datasets. This include its type, length, associated comment and some statistics (min, max, mean, standard deviation). A horizontal scrollbar at the bottom allows to get the two last properties, they are not displayed by default. Also note that if you find some columns are too narrow to show all significant digits, you can drag the vertical rules using Shift+Button 2.
Using the menu on the top of this dialog, you can manipulate existing sets or add new ones. Among the most important entries in the menu, are options to create or modify a set using the spreadsheet data set editor (see Spreadsheet data set editor).
The dialog presents an editable matrix of numbers, corresponding to the data set being edited. The set type (and hence, the number of data columns) can be changed using the "Type:" selector. Clicking on a column label pops up a dialog allowing to adjust the column formatting. Clicking on the row labels toggles the respective row state (selected/unselected). The selected rows can be deleted via the dialog's "Edit" menu. Another entry in this menu lets you add a row; the place of the new row is determined by the row containing a cell with the keyboard focus on. As well, just typing in an empty cell will add one or several rows (filling the intermediate rows with zeros).
To resize columns, drag the vertical rules using Shift+Button 2.
The set operations popup allows you to interact with sets as a whole. If you want to operate on the data ordering of the sets, you should use the data set operations popup from the Data menu. The popup allows you to select a source (one set within one graph) and a destination and perform some action upon them (copy, move, swap). This popup also give you a quick access to several graph and set selectors if you want to perform some other operation like hiding a graph or creating a new set from block data.
This entry fires up a popup to lay out several graphs in a regular grid given by M rows and N columns.
The graph selector at the top allows one to select a number of graphs the arrangement will operate on. If the number of selected graphs isn't equal to M times N, new graphs may be created or extra graphs killed if needed. These options are controlled by the respective checkboxes below the graph selector.
The order in which the matrix is filled in with the graphs can be selected (first horizontally then vertically or vise versa, with either of them inverted). Additionaly, one may choose to fill the matrix in the snake-like manner (adjacent "strokes" are anti-parallel).
The rest of the controls of the dialog window deal with the matrix spacing: left/right/top/bottom page offsets (in the viewport coordinates) and relative inter-cell distances, vertical and horizontal. Next to each of the vertical/horizontal spacing spinboxes, a "Pack" checkbox is found. Enabling it effectively sets the respective inter-cell distance to zero and alter axis tickmark settings such that only bottom/left-most tickmarks are visible.
If you don't want the regular layout this arrangement gives you, you can change it afterwards using the mouse (select a graph and double click on the focus marker, see clicks and double clicks).
You can overlay a graph on top of another one. The main use of this feature is to plot several curves using different scales on the same (apparently) graph. The main difficulty is to be sure you operate on the graph you want at all times (you can hide one for a moment if this becomes too difficult).
Using this entry, you can autoscale one graph or all graphs according to the specified sets only. This is useful if you need either to have truly comparable graphs despite every one contains data of different ranges, or if you want to focus your attention on one set only while it is displayed with other data in a complex graph.
This small popup only displays the current state (type and whether it is active or not) of the existing regions.
You can define a new region (or redefine an existing one), the allowed region types are:
A region can be either linked to the current graph only or to all graphs.
This kills a region.
This popup reports you which sets or points are inside or outside of a region.
You can link a set to a file or a pipe using this feature. Once a link has been established, you can update it (i.e., read data again) by clicking on the update button.
Currently, only simple XY sets can be used for hotlinks.
After having selected this menu entry, you can select a point on a graph that will be used as the origin of the locator display (just below the menu bar). The fixed point is taken into account only when the display type of the locator is set to [DX,DY].
This entry is provided to remove a fixed point set before and use the default again: point [0, 0].
The locator props popup allows you to customize the display of the locator, mainly its type and the format and precision of the display. You can use all the formats that are allowed in the graphs scales.
The preferences popup allows you to set miscellaneous properties of your Grace session, such as GUI behavior, cursor type, date reading hint and reference date used for calendar conversions.
This popup gathers all operations that are related to the ordering of data points inside a set or between sets. If you want to operate on the sets as a whole, you should use the set operations popup from the Edit menu. You can sort according to any coordinate (X, Y, DX, ...) in ascending or descending order, reverse the order of the points, join several sets into one, split one set into several others of equal lengths, or drop a range of points from a set. The set selector of the popup shows the number of points in each set in square brackets like this: G0.S0[63], the points are numbered from 0 to n-1.
The transformations sub-menu gives you access to all data-mining features of Grace.
Using evaluate expression allows you to create a set by applying an explicit formula to another set, or to parts of another set if you use regions restrictions.
All the classical mathematical functions are available (cos, sin, but also lgamma, j1, erf, ...). As usual all trigonometric functions use radians by default but you can specify a unit if you prefer to say cos (x rad) or sin (3 * y deg). For the full list of available numerical functions and operators, see Operators and functions.
In the formula, you can use X, Y, Y1, ..., Y4 to denote any coordinate you like from the source set. An implicit loop will be used around your formula so if you say:
x = x - 4966.5
you will shift all points of your set 4966.5 units to the left.
You can use more than one set in the same formula, like this:
y = y - 0.653 * sin (x deg) + s2.y
which means you use both X and Y from the source set but also the Y coordinate of set 2. Beware that the loop is a simple loop over the indices, all the sets you use in such an hybrid expression should therefore have the same number of points and point i of one set should really be related to point i of the other set. If your sets do not follow these requirements, you should first homogenize them using interpolation.
The histograms popup allows you to compute either standard or cumulative histograms from the Y coordinates of your data. Optionally, the histograms can be normalized to 1 (hence producing a PDF (Probability Distribution Function).
The bins can be either a linear mesh defined by its min, max, and length values, or a mesh formed by abscissas of another set (in which case abscissas of the set must form a strictly monotonic array).
This popup is devoted to direct and inverse Fourier transforms (actually, what is computed is a power spectrum). The default is to perform a direct transform on unfiltered data and to produce a set with the index as abscissa and magnitude as ordinate. You can filter the input data window through triangular, Hanning, Welch, Hamming, Blackman and Parzen filters. You can load magnitude, phase or coefficients and use either index, frequency or period as abscissas. You can choose between direct and inverse Fourier transforms. If you specify real input data, X is assumed to be equally spaced and ignored; if you specify complex input data X is taken as the real part and Y as the imaginary part.
If Grace was configured with the FFTW library (see configuration), then the DFT and FFT buttons really perform the same transform (so there is no speed-up in using FFT in this case). If you want Grace can to use FFTW wisdom files, you should set several environment variables to name them.
The running average popup allows you to compute some values on a sliding window over your data. You choose both the value you need (average, median, minimum, maximum, standard deviation) and the length of the window and perform the operation. You can restrict the operation to the points belonging to (or outside of) a region.
The differences popup is used to compute approximations of the first derivative of a function with finite differences. The only choice (apart from the source set of course) is the type of differences to use: forward, backward or centered.
The seasonal differences popup is used to subtract data from a period to data of the preceding period (namely y[i] - y[i + period]). Beware that the period is entered in terms of index in the set and not in terms of abscissa!
The integration popup is used to compute the integral of a set and optionally to load it. The numerical value of the integral is shown in the text field after computation. Selecting "cumulative sum" in the choice item will create and load a new set with the integral and compute the end value, selecting "sum only" will only compute the end value.
This popup is used to interpolate a set on an array of alternative X coordinates. This is mainly used before performing some complex operations between two sets with the evaluate expression popup.
The sampling array can be either a linear mesh defined by its min, max, and length values, or a mesh formed by abscissas of another set.
Several interpolation methods can be used: linear, spline or Akima spline.
Note that if the sampling mesh is not entirely within the source set X bounds, evaluation at the points beyond the bounds will be performed using interpolation parameters from the first (or the last) segment of the source set, which can be considered a primitive extrapolation. This behaviour can be disabled by checking the "Strict" option on the popup.
The abscissas of the set being interpolated must form a strictly monotonic array.
The regression popup can be used to fit a set against polynomials or some specific functions (y=A*x^B, y=A*exp(B*x), y=A+B*ln(x) and y=1/(A+Bx)) for which a simple transformation of input data can be used to apply linear regression formulas.
You can load either the fitted values, the residuals or the function itself. Choosing to load fitted values or residuals leads to a set of the same length and abscissas as the initial set. Choosing to load the function is almost similar to load the fitted values except that you choose yourself the boundaries and the number of points. This can be used for example to draw the curve outside of the data sample range or to produce an evenly spaced set from an irregular one.
The non linear fit popup can be used for functions outside of the simple regression methods scope. With this popup you provide the expression yourself using a0, a1, ..., a9 to denote the fit parameters (as an example you can say y = a0 * cos (a1 * x + a2)). You specify a tolerance, starting values and optional bounds and run several steps before loading the results.
The fit characteristics (number of parameters, formula, ...) can be saved in a file and retrieved as needed using the file menu of the popup.
In the "Advanced" tab, you can additionally apply a restriction to the set(s) to be fitted (thus ignoring points not satisfying the criteria), use one of preset weighting schemes or define your own (notice that "dY" in the preset "1/dY^2" one actually refers to the third column of the data set; use the "Custom" function if this doesn't make sense for your data set), and choose whether to load the fitted values, the residuals or the function itself. Choosing to load fitted values or residuals leads to a set of the same length and abscissas as the initial set. Choosing to load the function is almost similar to load the fitted values except that you choose yourself the boundaries and the number of points. This can be used for example to draw the curve outside of the data sample range or to produce an evenly spaced set from an irregular one.
This popup can be used to compute autocorrelation of one set or cross correlation between two sets. You only select the set (or sets) and specify the maximum lag. A check box allows one to evaluate covariance instead of correlation. The result is normalized so that abs(C(0)) = 1.
You can use a set as a weight to filter another set. Only the Y part and the length of the weighting set are important, the X part is ignored.
The convolution popup is used to ... convolve two sets. You only select the sets and apply.
You can rotate, scale or translate sets using the geometric transformations popup. You specify the characteristics of each transform and the application order.
This popup provides two sampling methods. The first one is to choose a starting point and a step, the second one is to select only the points that satisfy a boolean expression you specify.
This popup is devoted to reducing huge sets (and then saving both computation time and disk space).
The interpolation method can be applied only to ordered sets: it is based on the assumption that if a real point and an interpolation based on neighboring points are closer than a specified threshold, then the point is redundant and can be eliminated.
The geometric methods (circle, ellipse, rectangle) can be applied to any set, they test each point in turn and keep only those that are not in the neighborhood of previous points.
Given a set of curves in a graph, extract a feature from each curve and use the values of the feature to provide the Y values for a new curve.
Feature | Description |
Y minimum | Minimum Y value of set |
Y maximum | Maximum Y value of set |
Y average | Average Y value of set |
Y std. dev. | Standard deviation of Y values |
Y median | Median Y value |
X minimum | Minimum X value of set |
X maximum | Maximum X value of set |
X average | Average X value of set |
X std. dev. | Standard deviation of X values |
X median | Median X value |
Frequency | Perform DFT (FFT if set length a power of 2) to find largest frequency component |
Period | Inverse of above |
Zero crossing | Time of the first zero crossing, + or - going |
Rise time | Assume curve starts at the minimum and rises to the maximum, get time to go from 10% to 90% of rise. For single exponential curves, this is 2.2*time constant |
Fall time | Assume curve starts at the maximum and drops to the minimum, get time to go from 90% to 10% of fall |
Slope | Perform linear regression to obtain slope |
Y intercept | Perform linear regression to obtain Y-intercept |
Set length | Number of data points in set |
Half maximal width | Assume curve starts from the minimum, rises to the maximum and drops to the minimum again. Determine the time for which the curve is elevated more than 50% of the maximum rise. |
Barycenter X | Barycenter along X axis |
Barycenter Y | Barycenter along Y axis |
X (Y max) | X of Maximum Y |
Y (X max) | Y of Maximum X |
integral | cumulative sum |
Read new sets of data in a graph. A graph selector is used to specify the graph where the data should go (except when reading block data, which are copied to graphs later on).
Reading as "Single set" means that if the source contains only one column of numeric data, one set will be created using the indices (from 1 to the total number of points) as abscissas and read values as ordinates and that if the source contains more than one column of data, the first two numeric columns will be used. Reading as "NXY" means that the first numeric column will provide the abscissas and all remaining columns will provide the ordinates of several sets. Reading as "Block data" means all column will be read and stored and that another popup will allow to select the abscissas and ordinates at will. It should be noted that block data are stored as long as you do not override them by a new read. You can still retrieve data from a block long after having closed all popups, using the set selector.
The set type can be one of the predefined set presentation types (see sets).
The data source can be selected as "Disk" or "Pipe". In the first case the text in the "Selection" field is considered to be a file name (it can be automatically set by the file selector at the top of the popup). In the latter case the text is considered to be a command which is executed and should produce the data on its standard output. On systems that allows is, the command can be a complete sequence of programs glued together with pipes.
If the source contains date fields, they should be automatically detected. Several formats are recognized (see appendix dates in grace). Calendar dates are converted to numerical dates upon reading.
The "Autoscale on read" menu controls whether, upon reading in new sets, which axes of the graph should be autoscaled.
This entry exists only if Grace has been compiled with support for the NetCDF data format (see configuration).
Save data sets in a file. A set selector is used to specify the set to be saved. The format to use for saving data points can be specified (the default value is "%16.8g"). A warning is displayed if a file with the same name already exists.
The plot appearance popup let you set the time stamp properties and the background color of the page. The color is used outside of graphs and also on graphs were no specific background color is set. The time stamp is updated every time the project is modified.
The graph appearance popup can be displayed from both the plot menu and by double-clicking on a legend, title, or subtitle of a graph (see Clicks and double clicks). The graph selector at the top allows to choose the graph you want to operate on, it also allows certain common actions through its popup menu (see graph selector). Most of the actions can also be performed using the "Edit" menu available from the popup menubar. The main tab includes the properties you will need more often (title for example), and other tabs are used to fine tune some less frequently used options (fonts, sizes, colors, placements).
If you need special characters or special formatting in your title or subtitle, you can use Grace escape sequences (the sequence will appear verbatim in the text field but will be rendered on the graph), see typesetting. If you don't remember the mapping between alphabetic characters and the glyph you need in some specific fonts (mainly symbol and zapfdingbats), you can invoke the font tool from the text field by hitting CTRL-e. You can change fonts and select characters from there, they will be copied back in the text field when you press the "Accept" button. Beware of the position of the cursor as you enter text or change font in the font tool, the character or command will be inserted at this position, not at the end of the string!
You can save graph appearance parameters or retrieve settings previously saved via the "File" menu of this popup. In the "Save parameters" dialog, you can choose to save settings either for the current graph only or for all graphs.
The set appearance popup can be displayed from both the plot menu and by double-clicking anywhere in a graph (see Clicks and double clicks). The set selector at the top allows to choose the set you want to operate on, it also allows certain common actions through its popup menu (see set selector). The main tab gathers the properties you will need more often (line and symbol properties or legend string for example), and other tabs are used to fine tune some less frequently used options (drop lines, fill properties, annotated values and error bars properties for example).
You should note that despite the legend string related to one set is entered in the set appearance popup, this is not sufficient to display it. Displaying all legends is a graph level decision, so the toggle is in the main tab of the graph appearance popup.
If you need special characters or special formatting in your legend, you can use Grace escape sequences (the sequence will appear verbatim in the text field but will be rendered on the graph), see typesetting. If you don't remember the mapping between alphabetic characters and the glyph you need in some specific fonts (mainly symbol and zapfdingbats), you can invoke the font tool from the text field by hitting CTRL-e. You can change fonts and select characters from there, they will be copied back in the text field when you press the "Accept" button. Beware of the position of the cursor as you enter text or change font in the font tool, the character or command will be inserted at this position, not at the end of the string!
The axis properties popup can be displayed from both the "Plot" menu and by double-clicking exactly on an axis (see Clicks and double clicks). The pulldown menu at the top allows to select the axis you want to operate on. The "Active" toggle globally activates or deactivates the axis (all GUI elements are insensitive for deactivated axes). The start and stop fields depict the displayed range. Three types of scales are available: linear, logarithmic or reciprocal, and you can invert the axis (which normally increases from left to right and from bottom to top). The main tab includes the properties you will need more often (axis label, tick spacing and format for example), and other tabs are used to fine tune some less frequently used options (fonts, sizes, colors, placements, stagger, grid lines, special ticks, ...).
If you need special characters or special formatting in your label, you can use Grace escape sequences (the sequence will appear verbatim in the text field but will be rendered on the graph), see typesetting. If you don't remember the mapping between alphabetic characters and the glyph you need in some specific fonts (mainly symbol and zapfdingbats), you can invoke the font tool from the text field by hitting CTRL-e. You can change fonts and select characters from there, they will be copied back in the text field when you press the "Accept" button. Beware of the position of the cursor as you enter text or change font in the font tool, the character or command will be inserted at this position, not at the end of the string!
Most of the controls in the dialog should be self-explanatory. One that is not (and frequently missed) is the "Axis transform" input field in the "Tick labels" tab. Entering there e.g. "-$t" will make the tick labels show negates of the real coordinates their ticks are placed at. You can use any expression understood by the interpreter (see command interpreter).
Once you have set the options as you want, you can apply them. One useful feature is that you can set several axes at once with the bottom pulldown menu (current axis, all axes current graph, current axis all graphs, all axes all graphs). Beware that you always apply the properties of all tabs, not only the selected one.
This toggle item shows or hides the locator below the menu bar.
This toggle item shows or hides the status string below the canvas.
This toggle item shows or hides the tool bar at the left of the canvas.
Set the properties of the display device. It is the same dialog as in Print setup.
This menu item triggers a redrawing of the canvas.
This menu item causes an update of all GUI controls. Usually, everything is updated automatically, unless one makes modifications by entering commands in the Command tool.
Command driven version of the interface to Grace. Here, commands are typed at the "Command:" text item and executed when <Return> is pressed. The command will be parsed and executed, and the command line is placed in the history list. Items in the history list can be recalled by simply clicking on them with the left mouse button. For a reference on the Grace command interpreter, see Command interpreter.
Not written yet...
Not written yet...
Not written yet...
The console window displays errors and results of some numerical operations, e.g. nonlinear fit (see Non-linear fit). The window is popped up automatically whenever an error occurs or new result messages appear. This can be altered by checking the "Options/Popup only on errors" option.
Click on any element of the interface to get context-sensitive help on it. Only partially implemented at the moment.
Browse the Grace user's guide.
Browse the Grace tutorial.
Frequently Asked Questions with answers.
The list of changes during the Grace development.
The whole tree of submenus each loading a sample plot.
Use this to send your suggestions or bug reports.
Grace licensing terms will be displayed (GPL version 2).
A popup with basic info on the software, including some configuration details. More details can be found when running Grace with the "-version" command line flag.
The interpreter parses its input in a line-by-line manner. There may
be several statements per line, separated by semicolon (;
).
The maximal line length is 4 kbytes (hardcoded). The parser is
case-insensitive and ignores lines beginning with the "#
" sign.
Name | Description | Examples |
expr | Any numeric expression | 1.5 + sin(2) |
iexpr | Any expression that evaluates to an integer | 25, 0.1 + 1.9, PI/asin(1) |
nexpr | Non-negative iexpr | 2 - 1 |
indx | Non-negative iexpr | |
sexpr | String expression | "a string", "a " . "string", "square root of 4 = " . sqrt(4) |
vexpr | Vector expression | "2*x" |
Expression | Description | Types | Example |
GRAPH[id] | graph id | indx id | GRAPH[0] |
Gnn | graph nn | nn: 0-99 | G0 |
Expression | Description | Types | Example |
graph.SETS[id] | set id in graph graph | indx id, graphsel graph | GRAPH[0].SETS[1] |
graph.Snn | set nn in graph graph | nn: 0-99, graphsel graph | G0.S1 |
SET[id] | set id in the current graph | indx id | SET[1] |
Snn | set nn in the current graph | nn: 0-99 | S1 |
S_ | the last implicitly (i.e. as a result of a data transformation) allocated set in the current graph | - | S_ |
S$ | the active set in the current graph | - | S$ |
Expression | Description | Types | Example |
Rn | region n | n: 0-4 | R0 |
Expression | Description | Types | Example |
COLOR "colorname" | a mapped color colorname | - | COLOR "red" |
COLOR id | a mapped color with ID id | nexpr id | COLOR 2 |
Expression | Description | Types | Example |
PATTERN id | pattern with ID id | nexpr id | PATTERN 1 |
Expression | Description | Types | Example |
X | the first column | - | X |
Y | the second column | - | Y |
Yn | (n + 2)-th column | n = 0 - 4 | Y3 |
Not finished yet...
Variable | Description |
datacolumn | data column of current ("active") set |
set.datacolumn | data column of set |
vvar | user-defined array |
vvariable [i:j] | segment of a vector variable (elements from i-th to j-th inclusive, i <= j) |
Variable | Description |
vvariable[i] | i-th element of a vector variable |
var | user-defined variable |
In numerical expressions, the infix format is used. Arguments of both operators and functions can be either scalars or vector arrays. Arithmetic, logical, and comparison operators are given in tables below.
Operator | Description |
+ | addition |
- | substraction |
* | multiplication |
/ | division |
% | modulus |
^ | raising to power |
Operator | Description |
AND or && | logical AND |
OR or || | logical OR |
NOT or ! | logical NOT |
Operator | Description |
EQ or == | equal |
NE or != | not equal |
LT or < | less than |
LE or <= | less than or equal |
GT or > | greater than |
GE or >= | greater than or equal |
Another conditional operator is the "?:" (or ternary) operator, which operates as in C and many other languages.
(expr1) ? (expr2) : (expr3);
This expression evaluates to expr2 if expr1 evaluates to TRUE, and expr3 if expr1 evaluates to FALSE.
Function | Description |
abs(x) | absolute value |
acos(x) | arccosine |
acosh(x) | hyperbolic arccosine |
asin(x) | arcsine |
asinh(x) | hyperbolic arcsine |
atan(x) | arctangent |
atan2(y,x) | arc tangent of two variables |
atanh(x) | hyperbolic arctangent |
ceil(x) | greatest integer function |
cos(x) | cosine |
cosh(x) | hyperbolic cosine |
exp(x) | e^x |
fac(n) | factorial function, n! |
floor(x) | least integer function |
irand(n) | random integer less than n |
ln(x) | natural log |
log10(x) | log base 10 |
log2(x) | base 2 logarithm of x |
maxof(x,y) | returns greater of x and y |
mesh(n) | mesh array (0 ... n - 1) |
mesh(x1, x2, n) | mesh array of n equally spaced points between x1 and x2 inclusive |
minof(x,y) | returns lesser of x and y |
mod(x,y) | mod function (also x % y) |
pi | the PI constant |
rand | pseudo random number distributed uniformly on (0.0,1.0) |
rand(n) | array of n random numbers |
rint(x) | round to closest integer |
rsum(x) | running sum of x |
sgn(x) | signum function |
sin(x) | sine function |
sinh(x) | hyperbolic sine |
sqr(x) | x^2 |
sqrt(x) | x^0.5 |
tan(x) | tangent function |
tanh(x) | hyperbolic tangent |
Function | Description |
chdtr(df, x) | chi-square distribution |
chdtrc(v, x) | complemented Chi-square distribution |
chdtri(df, y) | inverse of complemented Chi-square distribution |
erf(x) | error function |
erfc(x) | complement of error function |
fdtr(df1, df2, x) | F distribution function |
fdtrc(x) | complemented F distribution |
fdtri(x) | inverse of complemented F distribution |
gdtr(a, b, x) | gamma distribution function |
gdtrc(a, b, x) | complemented gamma distribution function |
ndtr(x) | Normal distribution function |
ndtri(x) | inverse of Normal distribution function |
norm(x) | gaussian density function |
pdtr(k, m) | Poisson distribution |
pdtrc(k, m) | complemented Poisson distribution |
pdtri(k, y) | inverse Poisson distribution |
rnorm(xbar,s) | pseudo random number distributed N(xbar,s) |
stdtr(k, t) | Student's t distribution |
stdtri(k, p) | functional inverse of Student's t distribution |
Function | Description |
ai(x), bi(x) | Airy functions (two independent solutions of the differential equation y''(x) = xy ) |
beta(x) | beta function |
chi(x) | hyperbolic cosine integral |
ci(x) | cosine integral |
dawsn(x) | Dawson's integral |
ellie(phi, m) | incomplete elliptic integral of the second kind |
ellik(phi, m) | incomplete elliptic integral of the first kind |
ellpe(m) | complete elliptic integral of the second kind |
ellpk(m) | complete elliptic integral of the first kind |
expn(n, x) | exponential integral |
fresnlc(x) | cosine Fresnel integral |
fresnls(x) | sine Fresnel integral |
gamma(x) | gamma function |
hyp2f1(a, b, c, x) | Gauss hyper-geometric function |
hyperg(a, b, x) | confluent hyper-geometric function |
i0e(x) | modified Bessel function of order zero, exponentially scaled |
i1e(x) | modified Bessel function of order one, exponentially scaled |
igam(a, x) | incomplete gamma integral |
igamc(a, x) | complemented incomplete gamma integral |
igami(a, p) | inverse of complemented incomplete gamma integral |
incbet(a, b, x) | incomplete beta integral |
incbi(a, b, y) | Inverse of incomplete beta integral |
iv(v, x) | modified Bessel function of order v |
jv(v, x) | Bessel function of order v |
k0e(x) | modified Bessel function, third kind, order zero, exponentially scaled |
k1e(x) | modified Bessel function, third kind, order one, exponentially scaled |
kn(n, x) | modified Bessel function, third kind, integer order |
lbeta(x) | natural log of |beta(x)| |
lgamma(x) | log of gamma function |
psi(x) | psi (digamma) function |
rgamma(x) | reciprocal gamma function |
shi(x) | hyperbolic sine integral |
si(x) | sine integral |
spence(x) | dilogarithm |
struve(v, x) | Struve function |
voigt(gamma, sigma, x) | Voigt function (convolution of Lorentzian and Gaussian) |
yv(v, x) | Bessel function of order v |
zeta(x, q) | Riemann zeta function of two arguments |
zetac(x) | Riemann zeta function |
Function | Description |
MIN(x) | min value of array x |
MAX(x) | max value of array x |
AVG(x) | average of array x |
SD(x) | standard deviation of array x |
SUM(x) | sum of all elements of array x |
INT(x,y) | integral of y dx |
IMIN(x) | index of min value of array x |
IMAX(x) | index of max value of array x |
Methods of directly manipulating the data corresponding to the Data|Transformation menu are described in table transformations . To evaluate expressions, you can directly submit them to the command interpreter like you would in the evaluate expression window. As an example, S1.X = S1.X * 0.000256 scales the x-axis coordinates. The functionality of the 'Sample points' menu entry can be obtained through RESTRICT.
Statement | Description | Types | Example |
RUNAVG (set, npoints) | running average of set using npoints number of points | nexpr npoints | RUNAVG (S0, 100) |
RUNMED (set, npoints) | running median of set using npoints number of points | nexpr npoints | RUNMED (S0, 100) |
RUNMIN (set, npoints) | running minimum of set using npoints number of points | nexpr npoints | RUNMIN (S0, 100) |
RUNMAX (set, npoints) | running maximum of set using npoints number of points | nexpr npoints | RUNMAX (S0, 100) |
RUNSTD (set, npoints) | running standard deviation of set using npoints number of points | nexpr points | RUNSTD (S0, 100) |
INTERPOLATE (set, mesh, method, strict) | interpolate set on a sampling mesh using method. strict flag controls whether result should be bound within the source set | vexpr mesh, method: one of LINEAR, SPLINE, and ASPLINE, onoff strict | INTERPOLATE (S0, S1.X, ASPLINE, OFF) |
HISTOGRAM (set, bins, cumulative, normalize) | calculate histogram of set on defined bins. cumulative and normalize flags control whether to calculate cumulative and normalized (aka PDF) histograms, respectively. Data points are placed at upper limit of the bin | vexpr bins, onoff cumulative, onoff normalize | HISTOGRAM (S0, MESH(0, 1, 11), OFF, ON) |
XCOR (set1, set2, maxlag, covar) | calculate cross-correlation (or -covariance if the covar flag is set) of set1 with set2 with maximum lag maxlag. | nexpr maxlag, onoff covar | XCOR (S0, S0, 50, OFF) |
LINCONV (set1, set2) | calculate linear convolution of the array of ordinates of set1 with that of set2. | LINCONV (S0, S1) | |
RESTRICT (set, restriction) | filter set according to logical restriction. The original set will be overwritten | vexpr restriction | RESTRICT (S0, S0.X < 0) |
RESTRICT (set, region, negate) | filter set by keeping only points lying inside/outside region. The original set will be overwritten | onoff negate | RESTRICT (S0, R1, OFF) |
Not finished yet...
For producing "hard copy", several parameters can be set via the command interpreter. They are summarized in table Device parameters.
Command | Description |
PAGE SIZE xdim, ydim | set page dimensions (in pp) of all devices |
PAGE RESIZE xdim, ydim | same as above plus rescale the current plot accordingly |
DEVICE "devname" PAGE SIZE xdim, ydim | set page dimensions (in pp) of device devname |
DEVICE "devname" DPI dpi | set device's dpi (dots per pixel) |
DEVICE "devname" FONT onoff | enable/disable usage of built-in fonts for device devname |
DEVICE "devname" FONT ANTIALIASING onoff | enable/disable font aliasing for device devname |
DEVICE "devname" OP "options" | set device specific options (see Device-specific settings) |
HARDCOPY DEVICE "devname" | set device devname as current hardcopy device |
PRINT TO "filename" | set print output to filename (but do not print) |
PRINT TO DEVICE | set print output to hardcopy device (but do not print) |
Statement | Description | Types | Example |
EXIT(status) | cause normal program termination with exit status status | iexpr status | EXIT(0) |
EXIT | cause normal program termination; same as EXIT(0) | EXIT | |
HELP url | open a HTML document pointed to by url | sexpr url | HELP "doc/FAQ.html" |
HELP | open User's Guide | HELP | |
execute print job | |||
AUTOSCALE | scale the graph | AUTOSCALE | |
AUTOSCALE XAXES | scale the graph in x only | AUTOSCALE XAXES | |
AUTOSCALE YAXES | scale the graph in y only | AUTOSCALE YAXES | |
AUTOSCALE set | scale to a specific set | AUTOSCALE S0 | |
AUTOTICKS | autotick all axes | AUTOTICKS | |
REDRAW | refresh the canvas to reflect the current project state | REDRAW | |
SLEEP n | sleep for n seconds | expr n | SLEEP(3) |
UPDATEALL | update the GUI (graph and set selectors etc) to reflect the current project state | UPDATEALL | |
SAVEALL "file" | save project to file | sexpr file | SAVEALL "foo.agr" |
LOAD "file" | load project file | sexpr file | LOAD "foo.agr" |
User-defined variables are set and used according to the syntax described in table User variables.
Statement | Description | Types | Example |
DEFINE var | define new scalar variable var | DEFINE myvar | |
DEFINE vvar[] | define new vector variable vvar of zero length | DEFINE myvvar[] | |
DEFINE vvar[n] | define new vector variable vvar of length n | nexpr n | DEFINE myvvar[10] |
CLEAR var | undefine new variable var and deallocate associated storage | CLEAR myvar | |
vvar LENGTH n | reallocate vector variable vvar | nexpr n | myvvar LENGTH 25 |
Not finished yet...
We divide the commands pertaining to the properties and appearance of graphs into those which directly manipulate the graphs and those that affect the appearance of graph elements---the parameters that can appear in a Grace project file.
General graph creation/annihilation and control commands appear in table Graph operations.
Statement | Description | Types | Example |
FOCUS graph | Makes graph current and unhides it if necessary | graphsel graph | FOCUS G0 |
KILL graph | Kills graph | graphsel graph | KILL G0 |
ARRANGE(nrows, ncols, offset, hgap, vgap) | Arrange existing graphs (or add extra if needed) to form an nrows by ncols matrix, leaving offset at each page edge with hgap and vgap relative horizontal and vertical spacings | nexpr nrows, ncols, expr offset, hgap, vgap | ARRANGE(2, 2, 0.1, 0.15, 0.2) |
ARRANGE(nrows, ncols, offset, hgap, vgap, hvinv, hinv, vinv) | Same as above, plus additional hvinv, hinv, and vinv flags allowing to alter the order of the matrix filling | nexpr nrows, ncols, expr offset, hgap, vgap, onoff hvinv, hinv, vinv | ARRANGE(2, 2, 0.1, 0.15, 0.2, ON, OFF, ON) |
ARRANGE(nrows, ncols, offset, hgap, vgap, hvinv, hinv, vinv, snake) | Same as above, plus additional snake flag allowing to fill the matrix in a snake-like fashion | nexpr nrows, ncols, expr offset, hgap, vgap, onoff hvinv, hinv, vinv, snake | ARRANGE(2, 2, 0.1, 0.15, 0.2, ON, OFF, ON, ON) |
Setting the active graph and its type is accomplished with the commands found in table Graph selection parameters.
Statement | Description | Types | Example |
WITH graph | Makes graph current | graphsel graph | WITH G0 |
TYPE type | Sets type of current graph | graphtype type | TYPE XY |
graph onoff | (De)Activates selected graph | graphsel graph, onoff | G0 ON |
graph HIDDEN onoff | Hides selected graph | graphsel graph, onoff | G1 HIDDEN TRUE |
graph TYPE type | Sets type of graph | graphsel graph, graphtype type | G0 TYPE XYDY |
The axis range and scale of the current graph as well as its location
on the plot viewport are set with the commands listed in table
Axis parameters.
Statement | Description | Types | Example |
WORLD XMIN xmin | Sets minimum value of current graph's x axis to xmin | expr xmin | WORLD XMIN -10 |
WORLD XMAX xmax | Sets maximum value of current graph's x axis to xmin | expr xmax | WORLD XMAX 22.5 |
WORLD YMIN ymin | Sets minimum value of current graph's y axis to ymin | expr ymin | WORLD YMIN 0 |
WORLD YMAX ymax | Sets maximum value of current graph's y axis to ymax | expr ymax | WORLD YMAX 1e4 |
VIEW XMIN xmin | Sets left edge of current graph at x=xmin in the viewport | expr xmin | VIEW XMIN .2 |
VIEW XMAX xmax | Sets right edge of current graph at x=xmax in the viewport | expr xmax | VIEW XMAX 1.0 |
VIEW YMIN ymin | Sets bottom edge of current graph at y=ymin in the viewport | expr ymin | VIEW YMIN .25 |
VIEW YMAX ymax | Sets top edge of current graph at y=ymax in the viewport | expr ymax | VIEW YMAX .75 |
VIEW xmin, ymin, xmax, ymax | Sets graph's viewport | expr xmin, ymin, xmax, ymax | VIEW 0.15, 0.15, 1.15, 0.85 |
XAXES SCALE type | Set scaling of the x axes to type | type: one of NORMAL, LOGARITHMIC, or RECIPROCAL | XAXES SCALE NORMAL |
YAXES SCALE type | Set scaling of the y axes to type | type: one of NORMAL, LOGARITHMIC, or RECIPROCAL | YAXES SCALE LOGARITHMIC |
XAXES INVERT onoff | If ON, draws xmin to xmax from right to left | onoff | XAXES INVERT OFF |
YAXES INVERT onoff | If ON, draws ymin to ymax from top to bottom | onoff | YAXES INVERT OFF |
AUTOSCALE ONREAD type | Set automatic scaling on read according to type | type: one of NONE, XAXES, YAXES, XYAXES | AUTOSCALE ONREAD NONE |
The commands to set the appearance and textual content of titles and legends are given in table Titles and legends.
Statement | Description | Types | Example |
TITLE title | Sets the title of current graph | sexpr title | TITLE "Foo" |
TITLE FONT font | Selects font of title string | fontsel font | TITLE FONT 1 |
TITLE SIZE size | Sets size of title string | expr size | TITLE SIZE 1.5 |
TITLE COLOR color | Sets color of title string | colorsel color | TITLE COLOR 1 |
SUBTITLE subtitle | Sets the subtitle of current graph | sexpr subtitle | SUBTITLE "Bar" |
SUBTITLE FONT font | Selects font of subtitle string | fontsel font | SUBTITLE FONT "Times-Italic" |
SUBTITLE SIZE size | Sets size of subtitle string | expr size | SUBTITLE SIZE .60 |
SUBTITLE COLOR color | Sets color of subtitle string | colorsel color | SUBTITLE COLOR "blue" |
LEGEND onoff | Toggle legend display | onoff | LEGEND ON |
LEGEND LOCTYPE type | Posistion legend in type coordinates | type: either WORLD or VIEW | LEGEND LOCTYPE WORLD |
LEGEND xloc, yloc | Set location of legend box (upper left corner) | expr xloc, yloc | LEGEND .5,.75 |
LEGEND FONT font | Set legend font type | fontsel font | LEGEND FONT "Helvetica" |
LEGEND CHAR SIZE size | Sets size of legend label characters (1 is normal) | expr size | LEGEND CHAR SIZE .30 |
LEGEND color | Set color of legend text | colorsel color | LEGEND COLOR 1 |
LEGEND VGAP gap | Sets vertical gap between legend entries | nexpr gap | LEGEND VGAP 1 |
LEGEND HGAP gap | Sets horizontal gap between symbol and description | nexpr gap | LEGEND HGAP 4 |
LEGEND LENGTH length | Sets length of legend | nexpr length | LEGEND LENGTH 5 |
LEGEND INVERT onoff | Determines relationship between order of sets and order of legend labels | onoff | LEGEND INVERT true |
LEGEND BOX onoff | Determines if the legend bounding box is drawn | onoff | LEGEND BOX off |
LEGEND BOX COLOR color | Sets color of legend bounding box | colorsel color | LEGEND BOX COLOR 1 |
LEGEND BOX PATTERN pattern | Sets pattern of legend bounding box | patternsel pattern | LEGEND BOX PATTERN 2 |
LEGEND BOX LINESTYLE style | Sets line style of bounding box | nexpr style | LEGEND BOX LINESTYLE 1 |
LEGEND BOX LINEWIDTH width | Sets line width of bounding box | nexpr width | LEGEND BOX LINEWIDTH 2 |
LEGEND BOX FILL onoff | Determines if the legend bounding box is filled | onoff | LEGEND BOX FILL false |
LEGEND BOX FILL COLOR color | Sets color of legend box fill | colorsel color | LEGEND BOX COLOR 3 |
LEGEND BOX FILL pattern | Sets pattern of legend box fill | patternsel pattern | LEGEND BOX FILL PATTERN 1 |
Not finished yet...
Again, as with the graphs, we separate those parser commands that manipulate the data in a set from the commands that determine parameters---elements that are saved in a project file.
Operations for set I/O are summarized in table Set input, output, and creation. (Note that this is incomplete and only lists input commands at the moment.)
Statement | Description | Types | Example |
READ "file" | Reads file as a single set | sexpr file | READ "foo.dat" |
READ settype "file" | Reads file into a single set of type settype | xytype settype, sexpr file | READ xydy "bar.dat" |
READ NXY "file" | Reads file as NXY data | sexpr file | READ NXY "gad.dat" |
READ BLOCK "file" | Reads file as block data | sexpr file | READ BLOCK "zooks.dat" |
KILL BLOCK | Kills the current block data and frees the associated memory | KILL BLOCK | |
BLOCK settype columns | Forms a data set of type settype using columns from current block data file. | xytype settype, sexpr columns | BLOCK xydxdy "0:2:1:3" |
WRITE set | writes set to stdout | setsel set | WRITE G0.S1 |
WRITE set FORMAT "formatstring" | writes set to stdout using format specification formatstring | setsel set sexpr formatstring | WRITE G0.S1 FORMAT "%18.8g" |
WRITE set FILE "file" | writes set to file | setsel set sexpr file | WRITE G0.S1 FILE "data.dat" |
WRITE set FILE "file" FORMAT "formatstring" | writes set to file using format specification formatstring | setsel set sexpr file sexpr formatstring | WRITE G0.S1 FILE "data.dat" FORMAT "%18.8g" |
The parser commands analogous to the Data|Data set operations dialogue
can be found in table
Set operations.
Statement | Description | Types | Example |
COPY src TO dest | Copies src to dest | setsel src,dest | COPY S0 TO S1 |
MOVE src TO dest | Moves src to dest | setsel src,dest | MOVE G0.S0 TO G1.S0 |
SWAP src AND dest | Interchanges src and dest | setsel src,dest | SWAP G0.S0 AND G0.S1 |
KILL set | Kills set | setsel set | KILL G0.S0 |
Not Finished yet...
Not written yet...
For all devices, Grace uses Type1 fonts. Both PFA (ASCII) and PFB (binary) formats can be used.
The file responsible for the font configurations of Grace is
fonts/FontDataBase
. The first line contains a positive integer
specifying the number of fonts declared in that file. All remaining lines
contain declarations of one font each, composed out of three fields:
Here is the default FontDataBase
file:
14
Times-Roman Times-Roman n021003l.pfb
Times-Italic Times-Italic n021023l.pfb
Times-Bold Times-Bold n021004l.pfb
Times-BoldItalic Times-BoldItalic n021024l.pfb
Helvetica Helvetica n019003l.pfb
Helvetica-Oblique Helvetica-Oblique n019023l.pfb
Helvetica-Bold Helvetica-Bold n019004l.pfb
Helvetica-BoldOblique Helvetica-BoldOblique n019024l.pfb
Courier Courier n022003l.pfb
Courier-Oblique Courier-Oblique n022023l.pfb
Courier-Bold Courier-Bold n022004l.pfb
Courier-BoldOblique Courier-BoldOblique n022024l.pfb
Symbol Symbol s050000l.pfb
ZapfDingbats ZapfDingbats d050000l.pfb
For text rastering, three types of files are used.
.pfa
-/.pfb
-files: These contain the character
outline descriptions. The files are assumed to be in the
fonts/type1
directory; these are the filenames
specified in the FontDataBase
configuration file..afm
-files: These contain high-precision font metric
descriptions as well as some extra information, such as kerning
and ligature information for a particular font. It is assumed
that the filename of a font metric file has same basename as the
respective font outline file, but with the .afm
extension; the metric files are expected to be found in the
fonts/type1
directory, too..enc
-files: These contain encoding arrays in a special
but simple form. They are only needed if someone wants to load
a special encoding to re-encode a font. Their place is
fonts/enc
It is possible to use custom fonts with Grace. One mostly needs to use
extra fonts for the purpose of localization. For many European
languages, the standard fonts supplied with Grace should contain all the
characters needed, but encoding may have to be adjusted. This is done by
putting a Default.enc
file with proper encoding scheme into the
fonts/enc
directory. Grace comes with a few encoding files in
the directory; more can be easily found on the Internet. (If the
Default.enc
file doesn't exist, the IsoLatin1 encoding will be
used). Notice that for fonts having an encoding scheme in themselves
(such as the Symbol font, and many nationalized fonts) the default
encoding is ignored.
If you do need to use extra fonts, you should modify the
FontDataBase
file accordingly, obeying its format. However,
if you are going to exchange Grace project files with other people who
do not have the extra fonts configured, an important thing is to define
reasonable fall-back font names.
For example, let us assume I use Hebrew fonts, and the configuration file has lines like these:
...
Courier-Hebrew Courier courh___.pfa
Courier-Hebrew-Oblique Courier-Oblique courho__.pfa
...
My colleague, who lives in Russia, uses Cyrillic fonts with Grace
configured like this:
...
Cronix-Courier Courier croxc.pfb
Cronix-Courier-Oblique Courier-Oblique croxco.pfb
...
The font mapping information (Font name <-> Font fall-back) is
stored in the Grace project files. Provided that all the localized fonts
have English characters in the lower part of the ASCII table unmodified,
I can send my friend files (with no Hebrew characters, of course) and be
sure they render correctly on his computer.
Thus, with properly configured national fonts, you can make localized annotations for plots intended for internal use of your institution, while being able to exchange files with colleagues from abroad. People who ever tried to do this with MS Office applications should appreciate the flexibility :-).
The grace_np library is a set of compiled functions that allows you to launch and drive a Grace subprocess from your C or Fortran application. Functions are provided to start the subprocess, to send it commands or data, to stop it or detach from it.
Function | Arguments | Description |
int GraceOpenVA | (char *exe, int buf_size, ...) | launch a Grace executable exe and open a communication channel with it using buf_size bytes for data buffering. The remaining NULL-terminated list of options is command line arguments passed to the Grace process |
int GraceOpen | (int buf_size) | equivalent to GraceOpenVA("xmgrace", buf_size, "-nosafe", "-noask", NULL) |
int GraceIsOpen | (void) | test if a Grace subprocess is currently connected |
int GraceClose | (void) | close the communication channel and exit the Grace subprocess |
int GraceClosePipe | (void) | close the communication channel and leave the Grace subprocess alone |
int GraceFlush | (void) | flush all the data remaining in the buffer |
int GracePrintf | (const char* format, ...) | format a command and send it to the Grace subprocess |
int GraceCommand | (const char* cmd) | send an already formated command to the Grace subprocess |
GraceErrorFunctionType GraceRegisterErrorFunction | (GraceErrorFunctionType f) | register a user function f to display library errors |
Function | Arguments | Description |
integer GraceOpenF | (integer buf_size) | launch a Grace subprocess and open a communication channel with it |
integer GraceIsOpenF | (void) | test if a Grace subprocess is currently connected |
integer GraceCloseF | (void) | close the communication channel and exit the Grace subprocess |
integer GraceClosePipeF | (void) | close the communication channel and leave the Grace subprocess alone |
integer GraceFlushF | (void) | flush all the data remaining in the buffer |
integer GraceCommandF | (character*(*) cmd) | send an already formatted command to the Grace subprocess |
GraceFortranFunctionType GraceRegisterErrorFunctionF | (GraceFortranFunctionType f) | register a user function f to display library errors |
There is no Fortran equivalent for the GracePrintf function, you should format all the data and commands yourself before sending them with GraceCommandF.
The Grace subprocess listens for the commands you send and interprets them as if they were given in a batch file. You can send any command you like (redraw, autoscale, ...). If you want to send data, you should include them in a command like "g0.s0 point 3.5, 4.2".
Apart from the fact it monitors the data sent via an anonymous pipe, the Grace subprocess is a normal process. You can interact with it through the GUI. Note that no error can be sent back to the parent process. If your application send erroneous commands, an error popup will be displayed by the subprocess.
If you exit the subprocess while the parent process is still using it, the broken pipe will be detected. An error code will be returned to every further call to the library (but you can still start a new process if you want to manage this situation).
Here is an example use of the library, you will find this program in the distribution.
#include <stdlib.h>
#include <stdio.h>
#include <unistd.h>
#include <grace_np.h>
#ifndef EXIT_SUCCESS
# define EXIT_SUCCESS 0
#endif
#ifndef EXIT_FAILURE
# define EXIT_FAILURE -1
#endif
void my_error_function(const char *msg)
{
fprintf(stderr, "library message: \"%s\"\n", msg);
}
int
main(int argc, char* argv[])
{
int i;
GraceRegisterErrorFunction(my_error_function);
/* Start Grace with a buffer size of 2048 and open the pipe */
if (GraceOpen(2048) == -1) {
fprintf(stderr, "Can't run Grace. \n");
exit(EXIT_FAILURE);
}
/* Send some initialization commands to Grace */
GracePrintf("world xmax 100");
GracePrintf("world ymax 10000");
GracePrintf("xaxis tick major 20");
GracePrintf("xaxis tick minor 10");
GracePrintf("yaxis tick major 2000");
GracePrintf("yaxis tick minor 1000");
GracePrintf("s0 on");
GracePrintf("s0 symbol 1");
GracePrintf("s0 symbol size 0.3");
GracePrintf("s0 symbol fill pattern 1");
GracePrintf("s1 on");
GracePrintf("s1 symbol 1");
GracePrintf("s1 symbol size 0.3");
GracePrintf("s1 symbol fill pattern 1");
/* Display sample data */
for (i = 1; i <= 100 && GraceIsOpen(); i++) {
GracePrintf("g0.s0 point %d, %d", i, i);
GracePrintf("g0.s1 point %d, %d", i, i * i);
/* Update the Grace display after every ten steps */
if (i % 10 == 0) {
GracePrintf("redraw");
/* Wait a second, just to simulate some time needed for
calculations. Your real application shouldn't wait. */
sleep(1);
}
}
if (GraceIsOpen()) {
/* Tell Grace to save the data */
GracePrintf("saveall \"sample.agr\"");
/* Flush the output buffer and close Grace */
GraceClose();
/* We are done */
exit(EXIT_SUCCESS);
} else {
exit(EXIT_FAILURE);
}
}
To compile this program, type
cc example.c -lgrace_np
If Grace wasn't properly installed, you may need to instruct the
compiler about include and library paths explicitly, e.g.
cc -I/usr/local/grace/include example.c -L/usr/local/grace/lib -lgrace_np
When the FFTW capabilities are compiled in, Grace looks at two environment variables to decide what to do with the FFTW 'wisdom' capabilities. First, a quick summary of what this is. The FFTW package is capable of adaptively determining the most efficient factorization of a set to give the fastest computation. It can store these factorizations as 'wisdom', so that if a transform of a given size is to be repeated, it is does not have to re-adapt. The good news is that this seems to work very well. The bad news is that, the first time a transform of a given size is computed, if it is not a sub-multiple of one already known, it takes a LONG time (seconds to minutes).
The first environment variable is GRACE_FFTW_WISDOM_FILE. If this is set to the name of a file which can be read and written (e.g., $HOME/.grace_fftw_wisdom) then Grace will automatically create this file (if needed) and maintain it. If the file is read-only, it will be read, but not updated with new wisdom. If the symbol GRACE_FFTW_WISDOM_FILE either doesn't exist, or evaluates to an empty string, Grace will drop the use of wisdom, and will use the fftw estimator (FFTW_ESTIMATE flag sent to the planner) to guess a good factorization, instead of adaptively determining it.
The second variable is GRACE_FFTW_RAM_WISDOM. If this variable is defined to be non-zero, and GRACE_FFTW_WISDOM_FILE variable is not defined (or is an empty string), Grace will use wisdom internally, but maintain no persistent cache of it. This will result in very slow execution times the first time a transform is executed after Grace is started, but very fast repeats. I am not sure why anyone would want to use wisdom without writing it to disk, but if you do, you can use this flag to enable it.
Grace can access external functions present in either system or third-party shared libraries or modules specially compiled for use with Grace.
One must make sure, however, that the external function is of one
of supported by Grace types:
Grace type | Description |
f_of_i | a function of 1 int variable |
f_of_d | a function of 1 double variable |
f_of_nn | a function of 2 int parameters |
f_of_nd | a function of 1 int parameter and 1 double variable |
f_of_dd | a function of 2 double variables |
f_of_nnd | a function of 2 int parameters and 1 double variable |
f_of_ppd | a function of 2 double parameters and 1 double variable |
f_of_pppd | a function of 3 double parameters and 1 double variable |
f_of_ppppd | a function of 4 double parameters and 1 double variable |
f_of_pppppd | a function of 5 double parameters and 1 double variable |
The return values of functions are assumed to be of the
double
type.
Note, that there is no difference from the point of view of function prototype between parameters and variables; the difference is in the way Grace treats them - an attempt to use a vector expression as a parameter argument will result in a parse error.
Let us consider few examples.
Caution: the examples provided below (paths and compiler flags) are valid for Linux/ELF with gcc. On other operating systems, you may need to refer to compiler/linker manuals or ask a guru.
Suppose I want to use function pow(x,y)
from the Un*x math
library (libm). Of course, you can use the "^" operator defined
in the Grace language, but here, for the sake of example, we
want to access the function directly.
The command to make it accessible by Grace is
USE "pow" TYPE f_of_dd FROM "/usr/lib/libm.so"
Try to plot y = pow(x,2) and y = x^2 graphs (using, for example, "create new -> Formula" from any set selector) and compare.
Now, let us try to write a function ourselves. We will define
function my_function
which simply returns its (second)
argument multiplied by integer parameter transferred as the
first argument.
In a text editor, type in the following C code and save it as "my_func.c":
double my_function (int n, double x)
{
double retval;
retval = (double) n * x;
return (retval);
}
OK, now compile it:
$gcc -c -fPIC my_func.c
$gcc -shared my_func.o -o /tmp/my_func.so
(You may strip it to save some disk space):
$strip /tmp/my_func.so
That's all! Ready to make it visible to Grace as "myf" - we are too lazy to type the very long string "my_function" many times.
USE "my_function" TYPE f_of_nd FROM "/tmp/my_func.so" ALIAS "myf"
A more serious example. There is a special third-party library available on your system which includes a very important for you yet very difficult-to-program from the scratch function that you want to use with Grace. But, the function prototype is NOT one of any predefined types. The solution is to write a simple function wrapper. Here is how:
Suppose, the name of the library is "special_lib" and the
function you are interested in is called "special_func" and
according to the library manual, should be accessed as void
special_func(double *input, double *output, int parameter)
.
The wrapper would look like this:
double my_wrapper(int n, double x)
{
extern void special_func(double *x, double *y, int n);
double retval;
(void) special_func(&x, &retval, n);
return (retval);
}
Compile it:
$gcc -c -fPIC my_wrap.c
$gcc -shared my_wrap.o -o /tmp/my_wrap.so -lspecial_lib -lblas
$strip /tmp/my_wrap.so
Note that I added -lblas
assuming that the special_lib
library uses some functions from the BLAS. Generally, you have
to add all libraries which your module depends on (and
all libraries those libraries rely upon etc.), as if you wanted
to compile a plain executable.
Fine, make Grace aware of the new function
USE "my_wrapper" TYPE f_of_nd FROM "/tmp/my_wrap.so" ALIAS "special_func"
so we can use it with its original name.
An example of using Fortran modules.
Here we will try to achieve the same functionality as in Example 2, but with the help of F77.
DOUBLE PRECISION FUNCTION MYFUNC (N, X)
IMPLICIT NONE
INTEGER N
DOUBLE PRECISION X
C
MYFUNC = N * X
C
RETURN
END
As opposite to C, there is no way to call such a function from Grace directly - the problem is that in Fortran all arguments to a function (or subroutine) are passed by reference. So, we need a wrapper:
double myfunc_wrapper(int n, double x)
{
extern double myfunc_(int *, double *);
double retval;
retval = myfunc_(&n, &x);
return (retval);
}
Note that most of f77 compilers by default add underscore to
the function names and convert all names to the lower case,
hence I refer to the Fortran function MYFUNC
from my C
wrapper as myfunc_
, but in your case it can be different!
Let us compile the whole stuff:
$g77 -c -fPIC myfunc.f
$gcc -c -fPIC myfunc_wrap.c
$gcc -shared myfunc.o myfunc_wrap.o -o /tmp/myfunc.so -lf2c -lm
$strip /tmp/myfunc.so
And finally, inform Grace about this new function:
USE "myfunc_wrapper" TYPE f_of_nd FROM "/tmp/myfunc.so" ALIAS "myfunc"
In general the method outlined in the examples above can be used on OS/2, too. However you have to create a DLL (Dynamic Link Library) which is a bit more tricky on OS/2 than on most Un*x systems. Since Grace was ported by using EMX we also use it to create the examples; however other development environments should work as well (ensure to use the _System calling convention!). We refer to Example 2 only. Example 1 might demonstrate that DLLs can have their entry points (i.e. exported functions) callable via ordinals only, so you might not know how to access a specific function without some research. First compile the source from Example 2 to "my_func.obj"
gcc -Zomf -Zmt -c my_func.c -o my_func.obj
Then you need to create a linker definition file "my_func.def" which contains some basic info about the DLL and declares the exported functions.
LIBRARY my_func INITINSTANCE TERMINSTANCE
CODE LOADONCALL
DATA LOADONCALL MULTIPLE NONSHARED
DESCRIPTION 'This is a test DLL: my_func.dll'
EXPORTS
my_function
(don't forget about the 8 characters limit on the DLL name!). Finally link the DLL:
gcc my_func.obj my_func.def -o my_func.dll -Zdll -Zno-rte -Zmt -Zomf
(check out the EMX documentation about the compiler/linker flags used here!) To use this new library function within Grace you may either put the DLL in the LIBPATH and use the short form:
USE "my_function" TYPE f_of_nd FROM "my_func" ALIAS "myf"
or put it in an arbitrary path which you need to specify explicitly then:
USE "my_function" TYPE f_of_nd FROM "e:/foo/my_func.dll" ALIAS "myf"
(as for most system-APIs you may use the Un*x-like forward slashs within the path!)
Grace permits quite complex typesetting on a per string basis. Any string displayed (titles, legends, tick marks,...) may contain special control codes to display subscripts, change fonts within the string etc.
Control code | Description |
\f{x} | switch to font named "x" |
\f{n} | switch to font number n |
\f{} | return to original font |
\R{x} | switch to color named "x" |
\R{n} | switch to color number n |
\R{} | return to original color |
\#{x} | treat "x" (must be of even length) as list of hexadecimal char codes |
\t{xx xy yx yy} | apply transformation matrix |
\t{} | reset transformation matrix |
\z{x} | zoom x times |
\z{} | return to original zoom |
\r{x} | rotate by x degrees |
\l{x} | slant by factor x |
\v{x} | shift vertically by x |
\v{} | return to unshifted baseline |
\V{x} | shift baseline by x |
\V{} | reset baseline |
\h{x} | horizontal shift by x |
\n | new line |
\u | begin underline |
\U | stop underline |
\o | begin overline |
\O | stop overline |
\Fk | enable kerning |
\FK | disable kerning |
\Fl | enable ligatures |
\FL | disable ligatures |
\m{n} | mark current position as n |
\M{n} | return to saved position n |
\dl | LtoR substring direction |
\dr | RtoL substring direction |
\dL | LtoR text advancing |
\dR | RtoL text advancing |
\x | switch to Symbol font (same as \f{Symbol}) |
\+ | increase size (same as \z{1.19} ; 1.19 = sqrt(sqrt(2))) |
\- | decrease size (same as \z{0.84} ; 0.84 = 1/sqrt(sqrt(2))) |
\s | begin subscripting (same as \v{-0.4}\z{0.71}) |
\S | begin superscripting (same as \v{0.6}\z{0.71}) |
\T{xx xy yx yy} | same as \t{}\t{xx xy yx yy} |
\Z{x} | absolute zoom x times (same as \z{}\z{x}) |
\q | make font oblique (same as \l{0.25}) |
\Q | undo oblique (same as \l{-0.25}) |
\N | return to normal style (same as \v{}\t{}) |
\\ | print \ |
\n | switch to font number n (0-9) (deprecated) |
\c | begin using upper 128 characters of set (deprecated) |
\C | stop using upper 128 characters of set (deprecated) |
Example:
F\sX\N(\xe\f{}) = sin(\xe\f{})\#{b7}e\S-X\N\#{b7}cos(\xe\f{})
prints roughly
-x
F (e) = sin(e)·e ·cos(e)
x
using string's initial font and e prints as epsilon from the Symbol font.
NOTE:
Characters from the upper half of the char table can be entered directly
from the keyboard, using appropriate xmodmap(1)
settings, or
with the help of the font tool ("Window/Font tool").
Grace can output plots using several device backends. The list of available devices can be seen (among other stuff) by specifying the "-version" command line switch.
Some of the output devices accept several configuration options. You can set the options by passing a respective string to the interpreter using the "DEVICE "devname" OP "options"" command (see Device parameters). A few options can be passed in one command, separated by commas.
Command | Description |
grayscale | set grayscale output |
color | set color output |
level1 | use only PS Level 1 subset of commands |
level2 | use also PS Level 2 commands if needed |
docdata:7bit | the document data is 7bit clean |
docdata:8bit | the document data is 8bit clean |
docdata:binary | the document data may be binary |
xoffset:x | set page offset in X direction x pp |
yoffset:y | set page offset in Y direction y pp |
mediafeed:auto | default input tray |
mediafeed:match | select input with media matching page dimensions |
mediafeed:manual | manual media feed |
hwresolution:on | set hardware resolution |
hwresolution:off | do not set hardware resolution |
Command | Description |
grayscale | set grayscale output |
color | set color output |
level1 | use only PS Level 1 subset of commands |
level2 | use also PS Level 2 commands if needed |
bbox:tight | enable "tight" bounding box |
bbox:page | bounding box coincides with page dimensions |
Command | Description |
PDF1.3 | set compatibility mode to PDF-1.3 |
PDF1.4 | set compatibility mode to PDF-1.4 |
compression:value | set compression level (0 - 9) |
patterns:on | enable use of patterns |
patterns:off | disable use of patterns |
Command | Description |
format:pbm | output in PBM format |
format:pgm | output in PGM format |
format:ppm | output in PPM format |
rawbits:on | "rawbits" (binary) output |
rawbits:off | ASCII output |
Command | Description |
grayscale | set grayscale output |
color | set color output |
optimize:on/off | enable/disable optimization |
quality:value | set compression quality (0 - 100) |
smoothing:value | set smoothing (0 - 100) |
baseline:on/off | do/don't force baseline output |
progressive:on/off | do/don't output in progressive format |
dct:ifast | use fast integer DCT method |
dct:islow | use slow integer DCT method |
dct:float | use floating-point DCT method |
Command | Description |
interlaced:on | make interlaced image |
interlaced:off | don't make interlaced image |
transparent:on | produce transparent image |
transparent:off | don't produce transparent image |
compression:value | set compression level (0 - 9) |
We use two calendars in Grace: the one that was established in 532 by Denys and lasted until 1582, and the one that was created by Luigi Lilio (Alyosius Lilius) and Christoph Klau (Christophorus Clavius) for pope Gregorius XIII. Both use the same months (they were introduced under emperor Augustus, a few years after Julian calendar introduction, both Julius and Augustus were honored by a month being named after each one).
The leap years occurred regularly in Denys's calendar: once every four years, there is no year 0 in this calendar (the leap year -1 was just before year 1). This calendar was not compliant with earth motion and the dates were slowly shifting with regard to astronomical events.
This was corrected in 1582 by introducing Gregorian calendar. First a ten days shift was introduced to reset correct dates (Thursday October the 4th was followed by Friday October the 15th). The rules for leap years were also changed: three leap years are removed every four centuries. These years are those that are multiple of 100 but not multiple of 400: 1700, 1800, and 1900 were not leap years, but 1600 and 2000 were (will be) leap years.
We still use Gregorian calendar today, but we now have several time scales for increased accuracy. The International Atomic Time (TAI) is a linear scale: the best scale to use for scientific reference. The Coordinated Universal Time (UTC, often confused with Greenwich Mean Time) is a legal time that is almost synchronized with earth motion. However, since the earth is slightly slowing down, leap seconds are introduced from time to time in UTC (about one second every 18 months). UTC is not a continuous scale ! When a leap second is introduced by International Earth Rotation Service, this is published in advance and the legal time sequence is as follows: 23:59:59 followed one second later by 23:59:60 followed one second later by 00:00:00. At the time of this writing (1999-01-05) the difference between TAI and UTC was 32 seconds, and the last leap second was introduced in 1998-12-31.
These calendars allow to represent any date from the mist of the past to the fog of the future, but they are not convenient for computation. Another time scale is possible: counting only the days from a reference. Such a time scale was introduced by Joseph-Juste Scaliger (Josephus Justus Scaliger) in 1583. He decided to use "-4713-01-01T12:00:00" as a reference date because it was at the same time a Monday, first of January of a leap year, there was an exact number of 19 years Meton cycle between this date and year 1 (for Easter computation), and it was at the beginning of a 15 years Roman indiction cycle. The day number counted from this reference is traditionally called Julian day, but it has really nothing to do with the Julian calendar.
Grace stores dates internally as reals numbers counted from a reference date. The default reference date is the one chosen by Scaliger, it is a classical reference for astronomical events. It can modified for a single session using the Edit->Preferences popup of the GUI. If you often work with a specific reference date you can set it for every sessions with a REFERENCE DATE command in your configuration file (see Default template).
The following date formats are supported (hour, minutes and seconds are always optional):
One should be aware that Grace does not allow to put a space in one data column as spaces are used to separate fields. You should always use another separator (:/.- or better T) between date and time in data files. The GUI, the batch language and the command line flags do not have this limitation, you can use spaces there without any problem. The T separator comes from the ISO8601 standard. Grace support its use also in european and us formats.
You can also provide a hint about the format ("ISO8601",
"european", "us") using the -datehint command line flag or the
ref name="Edit->Preferences" id="preferences"> popup of the GUI.
The formats are tried in the following order: first the hint
given by the user, then iso, european and us (there is no
ambiguity between calendar formats and numerical formats and
therefore no order is specified for them). The separators
between various fields can be any characters in the set: " :/.-T"
(one or more spaces act as one separator, other characters can
not be repeated, the T separator is allowed only between date and time,
mainly for iso8601), so the string "1999-12 31:23/59" is allowed
(but not recommended). The '-' character is used both as a
separator (it is traditionally used in iso8601 format) and as
the unary minus (for dates in the far past or for numerical
dates). By default years are left untouched, so 99 is a date far
away in the past. This behavior can be changed with the
Edit->preferences popup, or with the
DATE WRAP on
and DATE WRAP YEAR year
commands. Suppose for example that the wrap year is chosen as
1950, if the year is between 0 and 99 and is written with two or
less digits, it is mapped to the present era as follows:
range [00 ; 49] is mapped to [2000 ; 2049]
range [50 ; 99] is mapped to [1950 ; 1999]
with a wrap year set to 1970, the mapping would have been:
range [00 ; 69] is mapped to [2000 ; 2069]
range [70 ; 99] is mapped to [1970 ; 1999]
this is reasonably Y2K compliant and is consistent with current use. Specifying year 1 is still possible using more than two digits as follows: "0001-03-04" is unambiguously March the 4th, year 1. The inverse transform is applied for dates written by Grace, for example as tick labels. Using two digits only for years is not recommended, we introduce a wrap year + 100 bug here so this feature should be removed at some point in the future ...
The date scanner can be used either for Denys's and Gregorian calendars. Inexistent dates are detected, they include year 0, dates between 1582-10-05 and 1582-10-14, February 29th of non leap years, months below 1 or above 12, ... the scanner does not take into account leap seconds: you can think it works only in International Atomic Time (TAI) and not in Coordinated Unified Time (UTC). If you find yourself in a situation were you need UTC, a very precise scale, and should take into account leap seconds ... you should convert your data yourself (for example using International Atomic Time). But if you bother with that you probably already know what to do.
This is a very brief guide describing problems and workarounds for reading in project files saved with Xmgr. You should read the docs or just play with Grace to test new features and controls.