将抓取的项目作为json传递给PHP脚本

Question

So I want to pass my scraped items to a php script, I have in my pipeline:

class TalkPhpPipeline(object):

    def process_item(self, item, spider):
        toPHP = json.dumps(dict(item))

        os.system( '/usr/bin/php script.php %s' % toPHP)
        return item

and in my script.php

<? require_once('wp-config.php');
$string = $argv[1];
echo 'PHP see this ';
var_dump($string);

The spider talks to script.php just fine, however my script.php only sees a srting of lenght 7 like this {title:, if I instead replace toPHP with item['title'], then the php side sees [usomewhere only we know], what I really want to pass to php is a string like: {'title': [u'somewhere only we know']} or just {'title': ['somewhere only we know']}, of course with all other fields that my item contains, how do I do that?

Thanks,

Answer 1

The shell is messing with the quotes in the command line. So use a different function (such as subprocess.call) that allows you to specify individual arguments:

subprocess.call(['/usr/bin/php', 'script.php', json.dumps(dict(item))])

Note that unless the JSON is very short, you may be better off just passing it to PHP in a different way, such as through a pipe, because operating systems have command line length limits.

In Python, you could use the subprocess.Popen class to pipe the JSON to the PHP script, which could retrieve the JSON using fgets(STDIN).