From 11287d721c4505a63f84995bec0b1cf058b45dfd Mon Sep 17 00:00:00 2001
From: Bhavya Srivastava <bhavya17037@iiitd.ac.in>
Date: Fri, 29 May 2020 17:38:21 +0530
Subject: [PATCH] Solves issue #45 Add jsonStringFixer.py, which has a function
 to add quotes around any required text in a json string. Used this in
 jsonld.py to handle invalid jsonld string.

---
 extruct/jsonStringFixer.py | 101 +++++++++++++++++++++++++++++++++++++
 extruct/jsonld.py          |   2 +
 2 files changed, 103 insertions(+)
 create mode 100644 extruct/jsonStringFixer.py

diff --git a/extruct/jsonStringFixer.py b/extruct/jsonStringFixer.py
new file mode 100644
index 00000000..a9a7624c
--- /dev/null
+++ b/extruct/jsonStringFixer.py
@@ -0,0 +1,101 @@
+def addMissingQuotesInJsonString(s):
+
+	s = s.split("\n")
+	finalstring=''
+	for i in range(len(s)):
+		s1 = s[i]
+		s2 = ''
+		startflag=0
+		j=0
+		while j<len(s1):
+			if startflag==0:
+				while True:
+					if j==len(s1):
+						break
+					if (ord(s1[j])>=48 and ord(s1[j])<=57) or (ord(s1[j])>=65 and ord(s1[j])<=90) or (ord(s1[j])>=97 and ord(s1[j])<=122):
+						s2=s2+'"'+s1[j]
+						j=j+1
+						startflag=1
+						break
+					elif s1[j]=='"':
+						s2=s2+s1[j]
+						j=j+1
+						startflag=1
+						break
+					else:
+						s2=s2+s1[j]
+						j=j+1
+
+
+			elif startflag==1:
+				while True:
+					if j==len(s1):
+						s2=s2+'"'
+						startflag=0
+						break
+					if s1[j]=="]" or s1[j]==";" or s1[j]=="," or (s1[j]==":" and s1[j+1]!="/") or s1[j]=="}" :
+						s2=s2+'"'+s1[j]
+						j=j+1
+						startflag=0
+						break
+					elif s1[j]=='"':
+						s2=s2+s1[j]
+						j=j+1
+						startflag=0
+						break
+					else:
+						s2=s2+s1[j]
+						j=j+1
+		if startflag==1 and j==len(s1):
+			s2=s2+'"'
+			startflag=0
+
+		finalstring = finalstring + s2 +'\n'
+
+	s = finalstring.split("\n")
+	finalstring=''
+	for i in range(len(s)):
+		s1 = s[i]
+		s2 = ''
+		colonflag=0
+		j=0
+		while j<len(s1):
+			if colonflag==0:
+				while True:
+					if j==len(s1):
+						break
+					elif s1[j]==":":
+						s2=s2+s1[j]
+						j=j+1
+						colonflag=1
+						break
+					else:
+						s2=s2+s1[j]
+						j=j+1
+
+
+			elif colonflag==1:
+				while True:
+					if j==len(s1):
+						s2 =s2+'null'
+						colonflag=0
+						break
+					if s1[j]=="]" or s1[j]==";" or s1[j]=="," :
+						s2=s2+'null'+s1[j]
+						j=j+1
+						colonflag=0
+						break
+					elif s1[j]!=" ":
+						s2=s2+s1[j]
+						j=j+1
+						colonflag=0
+						break
+					else:
+						s2=s2+s1[j]
+						j=j+1
+		if colonflag==1 and j==len(s1):
+			s2 =s2+'null'
+			colonflag=0
+
+		finalstring = finalstring + s2 +'\n'
+	return finalstring
\ No newline at end of file
diff --git a/extruct/jsonld.py b/extruct/jsonld.py
index f11580eb..25e704c2 100644
--- a/extruct/jsonld.py
+++ b/extruct/jsonld.py
@@ -7,6 +7,7 @@
 import re
 
 import lxml.etree
+from jsonStringFixer import addMissingQuotesInJsonString
 
 from extruct.utils import parse_html
 
@@ -33,6 +34,7 @@ def _extract_items(self, node):
             # TODO: `strict=False` can be configurable if needed
             data = json.loads(script, strict=False)
         except ValueError:
+            script = addMissingQuotesInJsonString(script)
             # sometimes JSON-decoding errors are due to leading HTML or JavaScript comments
             data = json.loads(
                 HTML_OR_JS_COMMENTLINE.sub('', script), strict=False)