- •Preface
- •Introduction
- •SWIG resources
- •About this manual
- •Prerequisites
- •Organization of this manual
- •How to avoid reading the manual
- •Credits
- •What’s new?
- •Bug reports
- •SWIG is free
- •Introduction
- •What is SWIG?
- •Life before SWIG
- •Life after SWIG
- •The SWIG package
- •A SWIG example
- •The swig command
- •Building a Perl5 module
- •Building a Python module
- •Shortcuts
- •Documentation generation
- •Building libraries and modules
- •C syntax, but not a C compiler
- •Non-intrusive interface building
- •Hands off code generation
- •Event driven C programming
- •Automatic documentation generation
- •Summary
- •SWIG for Windows and Macintosh
- •SWIG on Windows 95/NT
- •SWIG on the Power Macintosh
- •Cross platform woes
- •How to survive this manual
- •Scripting Languages
- •The two language view of the world
- •How does a scripting language talk to C?
- •Wrapper functions
- •Variable linking
- •Constants
- •Structures and classes
- •Shadow classes
- •Building scripting language extensions
- •Static linking
- •Shared libraries and dynamic loading
- •Linking with shared libraries
- •SWIG Basics
- •Running SWIG
- •Input format
- •SWIG Output
- •Comments
- •C Preprocessor directives
- •SWIG Directives
- •Simple C functions, variables, and constants
- •Integers
- •Floating Point
- •Character Strings
- •Variables
- •Constants
- •Pointers and complex objects
- •Simple pointers
- •Run time pointer type checking
- •Derived types, structs, and classes
- •Typedef
- •Getting down to business
- •Passing complex datatypes by value
- •Return by value
- •Linking to complex variables
- •Arrays
- •Creating read-only variables
- •Renaming declarations
- •Overriding call by reference
- •Default/optional arguments
- •Pointers to functions
- •Typedef and structures
- •Character strings and structures
- •Array members
- •C constructors and destructors
- •Adding member functions to C structures
- •Nested structures
- •Other things to note about structures
- •C++ support
- •Supported C++ features
- •C++ example
- •Constructors and destructors
- •Member functions
- •Static members
- •Member data
- •Protection
- •Enums and constants
- •References
- •Inheritance
- •Templates
- •Renaming
- •Adding new methods
- •SWIG, C++, and the Legislation of Morality
- •The future of C++ and SWIG
- •Objective-C
- •Objective-C Example
- •Constructors and destructors
- •Instance methods
- •Class methods
- •Member data
- •Protection
- •Inheritance
- •Referring to other classes
- •Categories
- •Implementations and Protocols
- •Renaming
- •Adding new methods
- •Other issues
- •Conditional compilation
- •The #if directive
- •Code Insertion
- •The output of SWIG
- •Code blocks
- •Inlined code blocks
- •Initialization blocks
- •Wrapper code blocks
- •A general interface building strategy
- •Preparing a C program for SWIG
- •What to do with main()
- •Working with the C preprocessor
- •How to cope with C++
- •How to avoid creating the interface from hell
- •Multiple files and the SWIG library
- •The %include directive
- •The %extern directive
- •The %import directive
- •The SWIG library
- •Library example
- •Creating Library Files
- •tclsh.i
- •malloc.i
- •Static initialization of multiple modules
- •More about the SWIG library
- •Documentation System
- •Introduction
- •How it works
- •Choosing a documentation format
- •Function usage and argument names
- •Titles, sections, and subsections
- •Formatting
- •Default Formatting
- •Comment Formatting variables
- •Sorting
- •Comment placement and formatting
- •Tabs and other annoyances
- •Ignoring comments
- •C Information
- •Adding Additional Text
- •Disabling all documentation
- •An Example
- •ASCII Documentation
- •HTML Documentation
- •LaTeX Documentation
- •C++ Support
- •The Final Word?
- •Pointers, Constraints, and Typemaps
- •Introduction
- •The SWIG Pointer Library
- •Pointer Library Functions
- •A simple example
- •Creating arrays
- •Packing a data structure
- •Introduction to typemaps
- •The idea (in a nutshell)
- •Using some typemaps
- •Managing input and output parameters
- •Input Methods
- •Output Methods
- •Input/Output Methods
- •Using different names
- •Applying constraints to input values
- •Simple constraint example
- •Constraint methods
- •Applying constraints to new datatypes
- •Writing new typemaps
- •Motivations for using typemaps
- •Managing special data-types with helper functions
- •A Typemap Implementation
- •What is a typemap?
- •Creating a new typemap
- •Deleting a typemap
- •Copying a typemap
- •Typemap matching rules
- •Common typemap methods
- •Writing typemap code
- •Scope
- •Creating local variables
- •Special variables
- •Typemaps for handling arrays
- •Typemaps and the SWIG Library
- •Implementing constraints with typemaps
- •Typemap examples
- •How to break everything with a typemap
- •Typemaps and the future
- •Exception Handling
- •The %except directive
- •Handling exceptions in C code
- •Exception handling with longjmp()
- •Handling C++ exceptions
- •Using The SWIG exception library
- •Debugging and other interesting uses for %except
- •More Examples
- •SWIG and Perl5
- •Preliminaries
- •Running SWIG
- •Compiling a dynamic module
- •Building a dynamic module with MakeMaker
- •Building a static version of Perl
- •Compilation problems and compiling with C++
- •Building Perl Extensions under Windows 95/NT
- •Running SWIG from Developer Studio
- •Using NMAKE
- •Modules, packages, and classes
- •Basic Perl interface
- •Functions
- •Global variables
- •Constants
- •Pointers
- •Structures and C++ classes
- •A simple Perl example
- •Graphs
- •Sample Perl Script
- •Accessing arrays and other strange objects
- •Implementing methods in Perl
- •Shadow classes
- •Getting serious
- •Wrapping C libraries and other packages
- •Building a Perl5 interface to MATLAB
- •The MATLAB engine interface
- •Wrapping the MATLAB matrix functions
- •Putting it all together
- •Graphical Web-Statistics in Perl5
- •Handling output values (the easy way)
- •Exception handling
- •Remapping datatypes with typemaps
- •A simple typemap example
- •Perl5 typemaps
- •Typemap variables
- •Name based type conversion
- •Converting a Perl5 array to a char **
- •Using typemaps to return values
- •Accessing array structure members
- •Turning Perl references into C pointers
- •Useful functions
- •Standard typemaps
- •Pointer handling
- •Return values
- •The gory details on shadow classes
- •Module and package names
- •What gets created?
- •Object Ownership
- •Nested Objects
- •Shadow Functions
- •Inheritance
- •Iterators
- •Where to go from here?
- •SWIG and Python
- •Preliminaries
- •Running SWIG
- •Compiling a dynamic module
- •Using your module
- •Compilation problems and compiling with C++
- •Building Python Extensions under Windows 95/NT
- •Running SWIG from Developer Studio
- •Using NMAKE
- •The low-level Python/C interface
- •Modules
- •Functions
- •Variable Linking
- •Constants
- •Pointers
- •Structures
- •C++ Classes
- •Python shadow classes
- •A simple example
- •Why write shadow classes in Python?
- •Automated shadow class generation
- •Compiling modules with shadow classes
- •Where to go for more information
- •About the Examples
- •Solving a simple heat-equation
- •The C++ code
- •Making a quick and dirty Python module
- •Using our new module
- •Accessing array data
- •Use Python for control, C for performance
- •Getting even more serious about array access
- •Implementing special Python methods in C
- •Summary (so far)
- •Wrapping a C library
- •Preparing a module
- •Using the gd module
- •Building a simple 2D imaging class
- •A mathematical function plotter
- •Plotting an unstructured mesh
- •From C to SWIG to Python
- •Putting it all together
- •Merging modules
- •Using dynamic loading
- •Use static linking
- •Building large multi-module systems
- •A complete application
- •Exception handling
- •Remapping C datatypes with typemaps
- •What is a typemap?
- •Python typemaps
- •Typemap variables
- •Name based type conversion
- •Converting Python list to a char **
- •Using typemaps to return arguments
- •Mapping Python tuples into small arrays
- •Accessing array structure members
- •Useful Functions
- •Standard typemaps
- •Pointer handling
- •Implementing C callback functions in Python
- •Other odds and ends
- •Adding native Python functions to a SWIG module
- •The gory details of shadow classes
- •A simple shadow class
- •Module names
- •Two classes
- •The this pointer
- •Object ownership
- •Constructors and Destructors
- •Member data
- •Printing
- •Shadow Functions
- •Nested objects
- •Inheritance and shadow classes
- •Methods that return new objects
- •Performance concerns and hints
- •SWIG and Tcl
- •Preliminaries
- •Running SWIG
- •Additional SWIG options
- •Compiling a dynamic module (Unix)
- •Using a dynamic module
- •Static linking
- •Compilation problems
- •Using [incr Tcl] namespaces
- •Building Tcl/Tk Extensions under Windows 95/NT
- •Running SWIG from Developer Studio
- •Using NMAKE
- •Basic Tcl Interface
- •Functions
- •Global variables
- •Constants
- •Pointers
- •Structures
- •C++ Classes
- •The object oriented interface
- •Creating new objects
- •Invoking member functions
- •Deleting objects
- •Accessing member data
- •Changing member data
- •Relationship with pointers
- •About the examples
- •Binary trees in Tcl
- •Making a quick a dirty Tcl module
- •Building a C data structure in Tcl
- •Implementing methods in C
- •Building an object oriented C interface
- •Building C/C++ data structures with Tk
- •Accessing arrays
- •Building a simple OpenGL module
- •Wrapping gl.h
- •Wrapping glu.h
- •Wrapping the aux library
- •A few helper functions
- •An OpenGL package
- •Using the OpenGL module
- •Problems with the OpenGL interface
- •Exception handling
- •Typemaps
- •What is a typemap?
- •Tcl typemaps
- •Typemap variables
- •Name based type conversion
- •Converting a Tcl list to a char **
- •Remapping constants
- •Returning values in arguments
- •Mapping C structures into Tcl Lists
- •Useful functions
- •Standard typemaps
- •Pointer handling
- •Writing a main program and Tcl_AppInit()
- •Creating a new package initialization library
- •Combining Tcl/Tk Extensions
- •Limitations to this approach
- •Dynamic loading
- •Turning a SWIG module into a Tcl Package.
- •Building new kinds of Tcl interfaces (in Tcl)
- •Shadow classes
- •Extending the Tcl Netscape Plugin
- •Using the plugin
- •Tcl8.0 features
- •Advanced Topics
- •Creating multi-module packages
- •Runtime support (and potential problems)
- •Why doesn’t C++ inheritance work between modules?
- •The SWIG runtime library
- •A few dynamic loading gotchas
- •Dynamic Loading of C++ modules
- •Inside the SWIG type-checker
- •Type equivalence
- •Type casting
- •Why a name based approach?
- •Performance of the type-checker
- •Extending SWIG
- •Introduction
- •Prerequisites
- •SWIG Organization
- •The organization of this chapter
- •Compiling a SWIG extension
- •Required C++ compiler
- •Writing a main program
- •Compiling
- •SWIG output
- •The Language class (simple version)
- •A tour of SWIG datatypes
- •The DataType class
- •Function Parameters
- •The String Class
- •Hash Tables
- •The WrapperFunction class
- •Typemaps (from C)
- •The typemap C API.
- •What happens on typemap lookup?
- •How many typemaps are there?
- •File management
- •Naming Services
- •Code Generation Functions
- •Writing a Real Language Module
- •Command Line Options and Basic Initialization
- •Starting the parser
- •Emitting headers and support code
- •Setting a module name
- •Final Initialization
- •Cleanup
- •Creating Commands
- •Creating a Wrapper Function
- •Manipulating Global Variables
- •Constants
- •A Quick Intermission
- •Writing the default typemaps
- •The SWIG library and installation issues
- •C++ Processing
- •How C++ processing works
- •Language extensions
- •Hints
- •Documentation Processing
- •Documentation entries
- •Creating a usage string
- •Writing a new documentation module
- •Using a new documentation module
- •Where to go for more information
- •The Future of SWIG
- •Index
SWIG Users Guide |
SWIG and Python |
170 |
|
|
|
% ld -shared example.o example_wrap.o -o examplecmodule.so
Notice the naming of ‘examplecmodule.so’ as opposed to ‘examplemodule.so’ that would have been created without shadow classes.
When using static linking, no changes need to be made to the compilation process.
Where to go for more information
Shadow classes turn out to be so useful that they are used almost all of the time with SWIG. All of the examples presented here will assume that shadow classes have been enabled. The precise implementation of shadow classes is described at the end of this chapter and is not necessary to effectively use SWIG.
About the Examples
The next few sections will go through a series of Python examples of varying complexity. These examples are designed to illustrate how SWIG can be used to integrate C/C++ and Python in a variety of ways. Some of the things that will be covered include :
•Controlling a simple C++ program with Python
•Wrapping a C library.
•Adding Python methods to existing C++ classes
•Accessing arrays and other common data structures.
•Building reusable components.
•Writing C/C++ callback functions in Python.
Solving a simple heat-equation
In this example, we will show how Python can be used to control a simple physics application-- in this case, some C++ code for solving a 2D heat equation. This example is probably overly simplistic, but hopefully it’s enough to give you some ideas.
The C++ code
Our simple application consists of the following two files :
//File : pde.h
//Header file for Heat equation solver
#include <math.h> #include <stdio.h>
//A simple 2D Grid structure
//A simple structure for holding a 2D grid of values struct Grid2d {
Grid2d(int ni, int nj); ~Grid2d();
double **data;
int |
xpoints; |
int |
ypoints; |
}; |
|
Version 1.1, June 24, 1997
SWIG Users Guide |
SWIG and Python |
171 |
|
|
|
// Simple class for solving a heat equation */
class Heat2d { |
|
|
private: |
|
|
Grid2d |
*work; |
// Temporary grid, needed for solver |
double |
h,k; |
// Grid spacing |
public: |
|
|
Heat2d(int ni, int nj); |
|
|
~Heat2d(); |
|
|
Grid2d |
*grid; |
// Data |
double |
dt; |
// Timestep |
double |
time; |
// Elapsed time |
void |
solve(int nsteps); |
// Run for nsteps |
void |
set_temp(double temp); |
// Set temperature |
}; |
|
|
The supporting C++ code implements a simple partial differential equation solver and some operations on the grid data structure. The precise implementation isn’t important here, but all of the code can be found in the “Examples/python/manual” directory of the SWIG distribution.
Making a quick and dirty Python module
Given our simple application, making a Python module is easy. Simply use the following SWIG interface file :
// File : pde.i %module pde
%{
#include "pde.h" %}
%include pde.h
Since pde.h is fairly simple, we can simply include it directly into our interface file using %include. However, we also need to make sure we also include it in the %{,%} block--other- wise we’ll get a huge number of compiler errors when we compile the resulting wrapper file.
To build the module simply run SWIG with the following options
swig -python -shadow pde.i
and compile using the techniques described in the beginning of this chapter.
Using our new module
We are now ready to use our new module. To do this, we can simply write a Python script like this :
# A fairly uninteresting example
from pde import *
h = Heat2d(50,50) |
# Creates a new “problem” |
Version 1.1, June 24, 1997
SWIG Users Guide |
SWIG and Python |
172 |
|
|
|
h.set_temp(1.0) print "Dt = ", h.dt
# Solve something
for i in range(0,25): h.solve(100)
print "time = ", h.time
When run, we get rather exciting output such as the following :
Dt = 2.5e-05 time = 0.0025 time = 0.005 time = 0.0075
...
time = 0.06 time = 0.0625
(okay, it’s not that exciting--well, maybe it is if you don’t get out much).
While this has only been a simple example it is important to note that we could have just as easily written the same thing in C++. For example :
// Python example written in C++
#include "pde.h" #include <stdio.h>
int main(int argc, char **argv) {
Heat2d *h;
h = new Heat2d(50,50); printf("Dt = %g\n", h->dt);
h->set_temp(1.0);
for (int i = 0; i < 25; i++) { h->solve(100);
printf("time = %g\n", h->time);
}
}
For the most part, the code looks identical (although the Python version is simpler). As for performance, the Python version runs less than 1% slower than the C++ version on my machine. Given that most of the computational work is written in C++, there is very little performance penalty for writing the outer loop of our calculation in Python in this case.
Unfortunately, our Python version suffers a number of drawbacks. Most notably, there is no way for us to access any of the grid data (which is easily accomplished in C++). However, there are ways to fix this :
Version 1.1, June 24, 1997
SWIG Users Guide |
SWIG and Python |
173 |
|
|
|
Accessing array data
Let’s modify our heat equation problem so that we can access grid data directly from Python. This can be done by modifying our interface file as follows :
%module pde %{
#include "pde.h" %}
%include pde.h
// Add a few "helper" functions to extract grid data %inline %{
double Grid2d_get(Grid2d *g, int i, int j) { return g->data[i][j];
}
void Grid2d_set(Grid2d *g, int i, int j, double val) { g->data[i][j] = val;
}
%}
Rather than modifying our C++ code, it is easy enough to supply a few accessor functions directly in our interface file. These function may only be used from Python so this approach makes sense and it helps us keep our C++ code free from unnecessary clutter. The %inline directive is a convenient method for adding helper functions since the functions you declare show up in the interface automatically.
We can now use our accessor functions to write a more sophisticated Python script :
# An example using our set/get functions |
|
from pde import * |
|
# Set up an initial condition |
|
def initcond(h): |
|
h.set_temp(0.0) |
|
nx = h.grid.xpoints |
|
for i in range(0,nx): |
|
Grid2d_set(h.grid,i,0,1.0) |
# Set grid values |
# Dump out to a file def dump(h,filename):
f = open(filename,"w") nx = h.grid.xpoints ny = h.grid.ypoints for i in range(0,nx):
for j in range(0,ny):
f.write(str(Grid2d_get(h.grid,i,j))+"\n") # Get grid value
f.close()
# Set up a problem and run it
h = Heat2d(50,50) initcond(h) fileno = 1
for i in range(0,25):
Version 1.1, June 24, 1997