Split Out Shortcode Bracket from JSON Array Bracket With Compatibility

Question

Generally, I have a hard time with regex patterns beyond basic matching. I have a simplistic shortcode parser similar to how Wordpress has so an example is:

~PLUGIN::name_of_plugin["param1"="value1","param2"="value2"]~

I have this splitting and working, I just use explode() on the [ and trim the ] to get the parameters portion. Now, I want to pass a nested array so I am trying to do a JSON object into those brackets instead, so the string to parse would look like this:

$str    =   '~PLUGIN::name_of_plugin[{
            "category":"whatever",
            "test":[{
                    "name":"Title Something",
                    "desc":"123123-A",
                    "result":"Confirmation",
                    "conforms":"true",
                    "mass_spec":"true"
                }]
        }]~';

After a series of different pattern attempts with varying degrees of success, this is what I came up with that I would consider useable:

preg_match('/^\~([a-z]+::)([a-z\_\-]+)([^\~]+)\~/i',$str,$match);
print_r($match);

It matches to this (I can trim key 3 for what I need):

Array
(
    [0] => ~PLUGIN::name_of_plugin[{
                "category":"whatever",
                "test":[{
                        "name":"Title Something",
                        "desc":"123123-A",
                        "result":"Confirmation",
                        "conforms":"true",
                        "mass_spec":"true"
                    }]
            }]~
    [1] => PLUGIN::
    [2] => name_of_plugin
    [3] => [{
                "category":"whatever",
                "test":[{
                    "name":"Title Something",
                    "desc":"123123-A",
                    "result":"Confirmation",
                    "conforms":"true",
                    "mass_spec":"true"
                }]
            }]
)

The problem there is, if the shortcode doesn't have parameters like:

$str    =   '~PLUGIN::name_test~';

It splits out to this (notice the n in key 3):

Array
(
    [0] => ~PLUGIN::name_of_plugin~
    [1] => PLUGIN::
    [2] => name_of_plugi
    [3] => n
)

Is there some sort of forward-looking thing I should be doing that would make this split out to:

Array
(
    [0] => ~PLUGIN::name_of_plugin~
    [1] => PLUGIN::
    [2] => name_of_plugin
)

but also will split the parameters block out when it exists? I am trying not to do some sort of hack where I implode() key 2 and 3 or something.

Answer 1

It seems that this works by just making the last group optional. I only added ? before the last ~.

Also, ~ and _ are not a special characters, so no need to escape them.

I also added the anchor for end of line $. It's usually a good idea if you want to make sure it captures the whole string.

^~([a-z]+::)([a-z_\-]+)([^~]+)?~$

See it work: regex101