PHP：将javascript对象字符串转换为php数组

Question

I get one page html source via phpQuery, and then get below string code from script tag in head via php regex:

var BASE_DATA = {
userInfo: {
  id: 0,
  userName: 'no-needed',
  avatarUrl: 'no-needed',
  isPgc: false,
  isOwner: false
},
headerInfo: {
  id: 0,
  isPgc: false,
  userName: 'no-needed',
  avatarUrl: 'no-needed',
  isHomePage: false,
  crumbTag: 'no-needed',
  hasBar: true
},
articleInfo: 
{
  title: 'needed',
  content: 'needed',
  groupId: 'needed',
  itemId: 'needed',
  type: 1,
  subInfo: {
    isOriginal: false,
    source: 'needed',
    time: 'needed'
  },
  tagInfo: {
    tags: [{"name":"no-needed 1"},{"name":"no-needed 2"},{"name":"no-needed 3"}],
    groupId: 'no-needed',
    itemId: 'no-needed',
    repin: 0,
  },
  has_extern_link: 0,
  coverImg: 'no-needed'
},
commentInfo:
{
  groupId: 'no-needed',
  itemId: 'no-needed',
  comments_count: 151,
  ban_comment: 0
},};

I want to convert this string to php array, like:

$base_data = array(
'articleInfo' => array(
    'title' => 'needed',
    'content' => 'needed',
    'groupId' => 'needed',
    'itemId' => 'needed',
    'subInfo' => array(
        'source' => 'needed',
        'time' => 'needed',
    ),
));

or

$base_data = array(
'title' => 'needed',
'content' => 'needed',
'groupId' => 'needed',
'itemId' => 'needed',
'subInfo' => array(
    'source' => 'needed',
    'time' => 'needed',
),);

I already tried with many ways, like: json_decode, get the content from the braces via php regex and the function preg_match_all.But all of them run not well.

I tried two ways:

the first way:

$json = str_ireplace(array('var BASE_DATA =', '};'), array('', '}'), $js);
json_decode($json, true);

the second way:

preg_match_all('/\{([^}]+)\}/', $js, $matches);
print_r($matches[1]);

or

preg_match_all('/articleInfo:\s*\{([^}]+)\}/', $script_text, $matches);
print_r($matches[1][0]);

It seems to close to finish, but it still looks no well, I have to parser string in articleInfo part.... that is why I posted this post.

I even wanted to use V8 JavaScript engine, but.....

do you anyone know the better way to finish it please ?

Answer 1

I had to reformat your JSON which was not valid (checked on https://jsonlint.com/ ).

I voluntarily used multiple str_replace() so you better understand the process, however you can optimize the code below by making multiple replacements at the same time within the same str_replace().

This works:

<?php

$to_decode = "var BASE_DATA = {
userInfo: {
  id: 0,
  userName: 'no-needed',
  avatarUrl: 'no-needed',
  isPgc: false,
  isOwner: false
},
headerInfo: {
  id: 0,
  isPgc: false,
  userName: 'no-needed',
  avatarUrl: 'no-needed',
  isHomePage: false,
  crumbTag: 'no-needed',
  hasBar: true
},
articleInfo: 
{
  title: 'needed',
  content: 'needed',
  groupId: 'needed',
  itemId: 'needed',
  type: 1,
  subInfo: {
    isOriginal: false,
    source: 'needed',
    time: 'needed'
  },
  tagInfo: {
    tags: [{\"name\":\"no-needed 1\"},{\"name\":\"no-needed 2\"},{\"name\":\"no-needed 3\"}],
    groupId: 'no-needed',
    itemId: 'no-needed',
    repin: 0,
  },
  has_extern_link: 0,
  coverImg: 'no-needed'
},
commentInfo:
{
  groupId: 'no-needed',
  itemId: 'no-needed',
  comments_count: 151,
  ban_comment: 0
},};";

/* Clean JSON and encapsulate in brackets */
$to_decode = str_replace('var BASE_DATA = {', '', $to_decode);
$to_decode = '{'.substr($to_decode, 0, -3).'}';

/* Remove spaces, tabs, new lines, etc. */
$to_decode = str_replace(' ', '', $to_decode);
$to_decode = str_replace("\n", '', $to_decode);
$to_decode = str_replace("\t", '', $to_decode);
$to_decode = str_replace("\r", '', $to_decode);

/* Encapsulate keys with quotes */
$to_decode = preg_replace('/([a-z_]+)\:/ui', '"{$1}":', $to_decode);
$to_decode = str_replace('"{', '"', $to_decode);
$to_decode = str_replace('}"', '"', $to_decode);
$to_decode = str_replace('\'', '"', $to_decode);

/* Remove unecessary trailing commas */
$to_decode = str_replace(',}', '}', $to_decode);

echo '<pre>';
var_dump(json_decode($to_decode));

Result using print_r :

(I added true/false for clarity, these will only show using var_dump() otherwise)

stdClass Object
(
    [userInfo] => stdClass Object
        (
            [id] => 0
            [userName] => no-needed
            [avatarUrl] => no-needed
            [isPgc] => false
            [isOwner] => false
        )

    [headerInfo] => stdClass Object
        (
            [id] => 0
            [isPgc] => false
            [userName] => no-needed
            [avatarUrl] => no-needed
            [isHomePage] => false
            [crumbTag] => no-needed
            [hasBar] => true
        )

    [articleInfo] => stdClass Object
        (
            [title] => needed
            [content] => needed
            [groupId] => needed
            [itemId] => needed
            [type] => 1
            [subInfo] => stdClass Object
                (
                    [isOriginal] => false
                    [source] => needed
                    [time] => needed
                )

            [tagInfo] => stdClass Object
                (
                    [tags] => Array
                        (
                            [0] => stdClass Object
                                (
                                    [name] => no-needed1
                                )

                            [1] => stdClass Object
                                (
                                    [name] => no-needed2
                                )

                            [2] => stdClass Object
                                (
                                    [name] => no-needed3
                                )

                        )

                    [groupId] => no-needed
                    [itemId] => no-needed
                    [repin] => 0
                )

            [has_extern_link] => 0
            [coverImg] => no-needed
        )

    [commentInfo] => stdClass Object
        (
            [groupId] => no-needed
            [itemId] => no-needed
            [comments_count] => 151
            [ban_comment] => 0
        )

)

Answer 2

thank @Bruno Leveque for your idea.

I edited your code like below so that it run well:

1. I changed $to_decode = str_replace(' ', '', $to_decode); to $to_decode = preg_replace('/[\\n| |\\s]{2,}/',' ',$to_decode); , that means all 1+ space will be changed to 1 space. because sometimes we need space, like: content: '

2. I added $to_decode = str_replace("'", '"', $to_decode); before your comment code /* Encapsulate keys with quotes */

3. changed $to_decode = preg_replace('/([a-z_]+)\\:/ui', '"{$1}":', $to_decode); to $to_decode = preg_replace('/([a-z_]+)\\: /ui', '"$1":', $to_decode); (one more space there); and commented //$to_decode = str_replace('"{', '"', $to_decode); and //$to_decode = str_replace('}"', '"', $to_decode);

4. added one more code: $to_decode = str_replace(", }", '}', $to_decode);

so my final code is :

because @Bruno Leveque does not know the exactly content of "needed" and "no-needed", so thank you for idea.

seems no perfect way....