我目前正在处理图像，其中一些是 DXT 压缩的，我需要一种使用 Python 解压缩和压缩这些文件的简单方法。不幸的是，我找不到任何可以为我做这件事的图书馆。
有谁知道一个好的 Python DXT 压缩库，或者压缩库的接口(interface)？
-- dav1d
编辑:
libsquish 是实现此目的的方法，但不幸的是 Python 绑定(bind)不起作用，所以这里是解决方案。
用 C++ 制作一个 squish-Wrapper，它导出内部访问 libsquish 的函数:
#include <squish.h>
typedef unsigned char u8;
extern "C" {
void CompressMasked( u8 const* rgba, int mask, void* block, int flags ) {
squish::CompressMasked(rgba, mask, block, flags);
}
void Compress( u8 const* rgba, void* block, int flags ) {
squish::Compress(rgba, block, flags);
}
void Decompress( u8* rgba, void const* block, int flags ) {
squish::Decompress(rgba, block, flags);
}
int GetStorageRequirements( int width, int height, int flags ) {
return squish::GetStorageRequirements(width, height, flags);
}
void CompressImage( u8 const* rgba, int width, int height, void* blocks, int flags ) {
squish::CompressImage(rgba, width, height, blocks, flags);
}
void DecompressImage( u8* rgba, int width, int height, void const* blocks, int flags ) {
squish::DecompressImage(rgba, width, height, blocks, flags);
}
}
创建一个动态库(在windows上是dll，在linux上是so，我称之为libsquishc.so)并用ctypes打开它。
我的做法(只导出我需要的功能):
from ctypes import CDLL, c_int, byref, create_string_buffer
import os.path
libsquish_path = os.path.join(os.path.split(os.path.abspath(__file__))[0], 'libsquishc.so')
libsquish = CDLL(libsquish_path)
DXT1 = 1 << 0
DXT3 = 1 << 1
DXT5 = 1 << 2
COLOR_ITERATIVE_CLUSTER_FIT = 1 << 8
COLOR_CLUSTER_FIT = 1 << 3
COLOR_RANGE_FIT = 1 << 4
WEIGHT_COLOR_BY_ALPHA = 1 << 7
GetStorageRequirements = libsquish.GetStorageRequirements
GetStorageRequirements.argtypes = [c_int, c_int, c_int]
GetStorageRequirements.restype = c_int
def compress_image(rgba, width, height, flags):
rgba = create_string_buffer(rgba)
c = GetStorageRequirements(width, height, flags)
buffer = create_string_buffer(c)
libsquish.Compress(byref(rgba), byref(buffer), c_int(flags))
return buffer.raw
def decompress_image(block, width, height, flags):
block = create_string_buffer(block)
c = width*height*4
rgba = create_string_buffer(c)
libsquish.DecompressImage(byref(rgba), c_int(width), c_int(height), byref(block), c_int(flags))
return rgba.raw
请您参考如下方法:
libSquish有a patch to add Python bindings .
编辑:安装程序好像是
- 下载squish-1.11.zip
- 解压并编译 - 应该生成一个 libsquish.a 文件
- 下载并安装Cython (听起来像是你这样做的)
- 创建一个临时目录并“应用”补丁——它会删除一堆新文件，这些文件是绑定(bind)代码
- 运行安装程序(sudo python setup.py install)
如果您这样做但仍然有错误，那么也许您应该 (a) 分享实际的错误消息，以便我们找出原因，或者 (b) 直接联系补丁作者 - mat (at) kivy.org
Edit2:编译错误很短，我将把它包括在这里:
running install
running build
running build_ext
skipping 'squish.c' Cython extension (up-to-date)
building 'squish' extension
gcc -pthread -fno-strict-aliasing -march=i686 -mtune=generic -O2 -pipe -fstack-protector --param=ssp-buffer-size=4 -D_FORTIFY_SOURCE=2 -DNDEBUG -march=i686 -mtune=generic -O2 -pipe -fstack-protector --param=ssp-buffer-size=4 -D_FORTIFY_SOURCE=2 -fPIC -I.. -I/usr/include/python2.7 -c squish.c -o build/temp.linux-i686-2.7/squish.o
In file included from squish.c:274:0:
/usr/include/squish.h:32:1: error: unknown type name 'namespace'
/usr/include/squish.h:32:18: error: expected '=', ',', ';', 'asm' or '__attribute__' before '{' token
squish.c: In function '__pyx_pf_6squish_compressImage':
squish.c:790:22: error: 'squish' undeclared (first use in this function)
squish.c:790:22: note: each undeclared identifier is reported only once for each function it appears in
squish.c:790:28: error: expected ';' before ':' token
squish.c:866:10: error: expected expression before ':' token
squish.c: In function '__pyx_pf_6squish_2decompressImage':
squish.c:1202:10: error: expected expression before ':' token
error: command 'gcc' failed with exit status 1
squish.h 的相关部分看起来像
#ifndef SQUISH_H
#define SQUISH_H
//! All squish API functions live in this namespace.
namespace squish {
// -----------------------------------------------------------------------------
它看起来像是在 namespace 关键字上窒息，我想说这意味着它在应该编译为 C++ 时编译为 C。
