aids.py

import re
import subprocess

regex = """<script>var[\s]+profile[\s]*=[\s]*({[^å]+)(?=</script>)"""

test_str = ("""
<script>var profile = {"env":null,
				"id":"xxx","listingStatus":"xxx",
				"price":{"value":"xxxx","text":"xxxx"},
				"createdDate":{"value":"xxxx","text":"xxxx","ago":"xxxxx"},
				"title":"Xxxxx",
				"text":"Xxxxx",
				"matrixData":[{"label":"Mærke ","value":"Xxxxx"},
							{"label":"Model","value":"Xxxxx"},
							{"label":"Xxxxx"},
							{"label":"Stand","value":"Xxxxx"}],
				"images":[{"small":"Xxxxx","medium":"Xxxxx","large":"Xxxxx"},
						{"small":"Xxxxx","medium":"Xxxxx","large":"Xxxxx"},
						{"small":"Xxxxx","medium":"Xxxxx","large":"Xxxxx"},
						{"small":"Xxxxx","medium":"Xxxxx","large":"Xxxxx"},
						{"small":"Xxxxx","medium":"Xxxxx","large":"Xxxxx"}],
				"breadcrumb":{"level1":{"name":"Xxxxx",
							"id":"Xxxxx",
							"url":"Xxxxx"},
							"level2":{"name":"Xxxxx",
							"id":"Xxxxx",
							"url":"XXXX/"},
				"level3":{"name":"XXXX",
						"id":"XXXX",
						"url":"XXXX"}
				},
				"headMeta":{"title":"XXXX",
							"description":"XXXX",
							"url":"XXXX",
							"price":"XXXX",
							"image":"XXXX"
							},
				"profile":{"type":"XXXX",
							"typeText":"XXXX",
							"ownerId":"XXXX",
							"name":"XXXX",
							"address":{"address":"XXXX",
										"zip":"XXXX",
										"city":"XXXX",
										"country":"XXXX",
										"geoLocation":{"latitude":0,"longitude":0}
										},
								"nemIdValidated":null,
								"website":null,
								"registrationDate":{"value":"XXXX",
													"text":"XXXX",
													"ago":"XXXX"},
								"profilePicture":{"url":"XXXX",
													"name":"XXXX"},
								"cvr":null,
								"usPs":null,
								"openingHours":null,
								"soiLink":"XXXX",
								"contactOptions":[{"type":"XXXX",
												"data":[],
												"title":"XXXX"}],
												"mapLink":null,
												"isFollowedByUser":false},
								"links":[{"caption":"XXXX",
										"rel":"XXXX",
										"verb":"XXXX",
										"href":"XXXX",
										"gaAction":"XXXX"}],
								"dataLayer":{"p":{"t":"XXXX",
												"pl":"XXXX",
												"v":"XXXX"},
											"a":{"id":0,
												"t":"XXXX",
												"cdt":0,
												"age":0,
												"lpdt":0,
												"dl":0,
												"ic":0,
												"prc":{"amt":0,
														"cur":"xxx"},
												"ftr":["QuestionsAndAnswers",
														"VisibleContactInformation",
														"Shipping"],
												"u":{"li":false,
													"huid":"XXXX",
													"hue":"XXXX",
													"at":"XXXX"},
												"attr":{"Make":"XXXX",
														"Condition":"XXXX",
														"Model":"XXXX",
														"Memory":"XXXX",
														"SELLER_TYPE":"XXXX"},
												"syiId":"XXXX"},
								"c":{"l1":{"id":0,
											"n":"XXXX"},
									"l2":{"id":0,
											"n":"XXXX"},
									"l3":{"id":0,
											"n":"XXXX"},
									"c":{"id":0,"n":"XXXX"}},
								"l":{"c":{"id":0,"n":"XXXX"},
								"l1":{"id":0,"n":"XXXX"},
								"l2":{"id":0,"n":"XXXX"},
								"l3":{"id":0,"n":"XXXX"},
								"l4":{"id":0,"n":"XXXX"},
								"ltlng":"xxxx",
								"pcid":"0"}},
				"relatedListings":[{"links":[{"rel":"self",
											"verb":"GET",
											"href":"xxxx"},
											{"rel":"image","verb":"GET","href":"xxxx"}],
									"text":"xxxx",
									"price":{"value":0,
												"text":"2.139 kr."}},
											{"links":[{"rel":"self","verb":"GET","href":"xxxx"},
											{"rel":"image","verb":"GET","href":"xxxx"}],
											"text":"xxxx",
											"price":{"value":0,"text":"xxxx"}},
									{"links":[{"rel":"self","verb":"GET","href":"xxxx"},
												{"rel":"image","verb":"GET","href":"xxxx"}],
												"text":"xxxx","price":{"value":0,
																		"text":"xxxx"}}],
									"bilbasenRelatedListings":[],
									"popularSearches":[{"query":"xxxx","url":"xxxx","label":"xxxx"},
														{"query":"xxxx","url":"xxxx","label":"mobiltelefoner"},
														{"query":"xxxx","url":"xxxx","label":"mobiltelefoner"},
														{"query":"xxxx","url":"xxxx","label":"mobiltelefoner"},
														{"query":"xxxx","url":"xxxx","label":"mobiltelefoner"},
														],
									"sellersOtherItems":null,
									"adsense":{"format":"plas",
												"pageLevelParameters":{"testgl":"da",
												"textColorPalette":["#333333"],
												"priceCurrency":"xxxx",
												"hl":"xxxx",
												"adsafe":"medium",
												"pubId":"xxxx",
												"query":"xxxx",
												"channel":"xxxx", "target":"_blank","adLayout":"xxxx"}},
									"advertising":null,"isFavorite":false};
                                    </script>""")

def to_json(javascript):

    javascript = javascript.rstrip(' ;\n\r')

    try:
        cmd = ['node', '-e', """console.log(JSON.stringify(""" + javascript + """))"""]
        out = subprocess.check_output(cmd, stderr=subprocess.STDOUT, universal_newlines=True)
    except subprocess.CalledProcessError as exc:
        # dette var ikke korrekt javascript 
        return (False, exc.output)
    else:
        # dette var korrekt javascript
        return (True, out)

matches = re.finditer(regex, test_str, re.MULTILINE | re.DOTALL)

for matchNum, match in enumerate(matches, start=1):
    for groupNum in range(0, len(match.groups())):
        groupNum = groupNum + 1
        output = to_json(match.group(groupNum))
        if(output[0]):
            print(output[1])