0

我正在构建一个在线服务,可以制作网页截图并将其像图像一样返回给用户。这个怎么运作:

1)我为 X 服务器创建虚拟帧缓冲区 - 进一步不要创建真正的窗口

2)在环境变量中注意要使用什么显示

3)创建子进程(否则前一点不会生效)

4) 在子进程中创建 webkit.WebView(),“显示”窗口并加载网页

5)当我收到有关页面已完全加载的通知时-我制作屏幕截图并将其保存到文件中(仅在项目的这一步中,也-如何将其返回到用户的浏览器?我知道Content type: image/png,但更进一步-gtk。 gdk.Pixbuf.save(stdout_file_name)?)

所以!问题!如果我从控制台运行它python parent.cgi- 一切都很完美,但如果我在 Web 浏览器中打开parent.cgi(服务器运行apache2) - 然后在真正的浏览器页面中尝试无限加载,并且在服务器上的进程中我看到Xvfb出现并加上三个(显然是和正确)python过程:

python return_picture.cgi
python /home/argon/www/wool/cgi-bin/parent.cgi
bash/sh -c python return_picture.cgi

编码:

父.cgi:

#! /usr/bin/env python
# -*- coding: utf-8 -*-

import gtk,webkit,gobject,sys,os,time,subprocess,logging
import cgitb,cgi
import signal
cgitb.enable()

logging.basicConfig(format = u'%(levelname)-8s [%(asctime)s] %(message)s', level = logging.INFO, filename = u'mylog.log')
logging.critical('ONE MORE TIME')
print "Content-Type: text/html;charset=utf-8"
print '' #it is importantly required

class writer(object):
    def write(self, data):
        logging.critical(data)

sys.stdout = writer()
sys.stderr = writer()
class XServer():
    def __init__(self, silence_xvfb=True, display='1', screen='0', xvfb_timeout=3):
        self.pidfile = '/tmp/.X%s-lock' % display
        redirect = '> /dev/null 2>&1'
        redirect = ''
        if not silence_xvfb:
            redirect = ''
        cmd = ' '.join(['Xvfb', ':'+display, '-screen', screen, '1600x1200x24', redirect])
        if(os.path.isfile(self.pidfile)):
            self._kill_xvfb()
        #os.system(cmd+' &')
        subprocess.Popen(cmd+' &', shell=True) #works througth filenodes thats why it is impossible to redirect to log - overriding of file.write() does not make sense
        print 'XVFB STARTED'
        self.xvfb = True
        start = time.time()
        while(True):
            diff = time.time() - start
            if(diff > xvfb_timeout):
                raise SystemError("Timed-Out waiting for Xvfb to start - {0} sec".format(xvfb_timeout))
            if(os.path.isfile(self.pidfile)):
                break
            else:
                time.sleep(0.05)

        os.putenv('DISPLAY', ':%s' % display)

    def _kill_xvfb(self):
        pid = int(open(self.pidfile).read().strip())
        os.kill(pid, signal.SIGINT)
        print 'KILLED'

    def __del__(self):
        # Kill the frame buffer
        if(self.xvfb):
            self._kill_xvfb()

xserver = XServer()

logging.debug('lets create child')
child = subprocess.Popen("python return_picture.cgi",shell=True,stdout=subprocess.PIPE)
s=child.stdout.readline()
print 'there again'

return_picture.cgi:

#! /usr/bin/env python
# -*- coding: utf-8 -*-

from blessings import Terminal
import gtk,webkit,gobject,sys,os,time,signal,logging
import cgitb,cgi

cgitb.enable()
t = Terminal()
logging.basicConfig(format = u'%(levelname)-8s [%(asctime)s] %(message)s', level = logging.DEBUG, filename = u'mylog.log')

class writer(object):
    def write(self, data):
        logging.critical(data)

sys.stdout = writer()
sys.stderr = writer()
logging.debug('IN CHILD')
#print "Content-Type: image/png"
#print "Content-Type: text/html;charset=utf-8"
#print '' #it is importantly required
#print 'hello'
web=webkit.WebView()
win=gtk.Window()
index = 0
finished = False
def finished_cb(web_view,sig,res):
    global finished
    global index
    index += 1
    #print index,': ',
    status = web_view.get_property("load-status").value_name
    logging.debug(str(index)+': '+status)
    if "FINISH" in status and not finished:
        finished = True
        gobject.timeout_add(500,drawWindow)
        print 'timeout'
        return

sig2= "resource-load-finished"
web.connect(sig2, finished_cb)
url = 'http://google.com/'
web.open(url)

win.add(web)

def drawWindow():
    width, height = win.get_size()
    pixbuf = gtk.gdk.Pixbuf(gtk.gdk.COLORSPACE_RGB, False, 8, width, height)

    screenshot = pixbuf.get_from_drawable(win.window, win.get_colormap(), 
                                          0, 0, 0, 0, width, height)
    ret = screenshot.subpixbuf(0,0,300,20)
    ret = screenshot
    stdot_filename = os.readlink('/proc/self/fd/0') 
    print stdot_filename
    #screenshot.save(stdot_filename, 'png')

    gtk.main_quit()
    screenshot.save('screenshot.png', 'png')
    print 'screenshot saved'
win.show_all()
gtk.main()
4

1 回答 1

1

您应该将屏幕截图请求与其生成分离。

让网页插入要截取屏幕截图的 URL 到队列中。有一个进程将项目从该队列中拉出并运行屏幕截图生成脚本。

这样,Web 浏览器就不会等待屏幕截图进程运行(无论如何都可能会失败),您可以轻松处理重复请求,并且您不会压倒您的服务器启动更多的 WebKit 实例,而不是同时运行.

这就是我在http://bookmarkly.com上的做法

于 2012-08-24T21:24:34.610 回答