PHP JSON String escape Double Quotes inside value?

Question

I'm getting a JSON string from an api like this :

{ "title":"Example string's with "special" characters" }

which is not json decodable by using json_decode (its output is null).

so I want to change it to something json decodable like :

{ "title":"Example string's with \"special\" characters" }

or

{ "title":"Example string's with 'special' characters" }

in order to make json_decode function work, what should I do ?

Answer 1

Since yesterday I was trying to solve this tricky problem, and after a lot of hair pulling I came up with this solution.

First let us clarify our assumptions.

• json string should be right formated.
• The keys and values are quoted with double quotation.

Analyzing the problem:

We know that json keys formated like this (,"keyString":) and json value is (:"valueString",)

keyString: is any sequence of characters except (:").

valueString: is any sequence of characters except (",).

Our goal is to escape quotations inside valueString, to achive that we need to separate keyStrings and valueStrings.

• But we have also a valid json format like this ("keyString":digit,) this will cause a problem because it breaks out assumption that says values always ended with (",)
• Another problem is having empty values like ("keyString":" ")

Now after analyzing the problem we can say

1. json keyString has (,") before it and (":) after it.
2. json valueString can have (:") before and (",) after OR (:) before and digit as a value then (,) after OR (:) followed by (" ") then (,)

The solution: Using this facts the code will be

function escapeJsonValues($json_str){
  $patern = '~(?:,\s*"((?:.(?!"\s*:))+.)"\s*(?=\:))(?:\:\s*(?:(\d+)|("\s*")|(?:"((?!\s*")(?:.(?!"\s*,))+.)")))~';
  //remove { }
  $json_str = rtrim(trim(trim($json_str),'{'),'}');
  if(strlen($json_str)<5) {
    //not valid json string;
    return null;
  }
  //put , at the start nad the end of the string
  $json_str = ($json_str[strlen($json_str)-1] ===',') ?','.$json_str :','.$json_str.',';
  //strip all new lines from the string
  $json_str=preg_replace('~[\r\n\t]~','',$json_str); 

  preg_match_all($patern, $json_str, $matches);
  $json='{';
  for($i=0;$i<count($matches[0]);$i++){

        $json.='"'.$matches[1][$i].'":';
        //value is digit
        if(strlen($matches[2][$i])>0){
            $json.=$matches[2][$i].',';
        } 
        //no value
        elseif (strlen($matches[3][$i])>0) {
            $json.='"",';
        }
        //text, and now we can see if there is quotations it will be related to the text not json
        //so we can add slashes safely
        //also foreword slashes should be escaped 
        else{
            $json.='"'.str_replace(['\\','"' ],['/','\"'],$matches[4][$i]).'",';
        }
  }
  return trim(rtrim($json,','),',').'}';
}

Note: The code realizes white spaces.