json-repair 0.16.0__py3-none-any.whl → 0.16.2__py3-none-any.whl

json_repair/json_repair.py

@@ -11,7 +11,7 @@ This module will parse the JSON file following the BNF definition:
 
     <container> ::= <object> | <array>
     <array> ::= '[' [ <json> *(', ' <json>) ] ']' ; A sequence of JSON values separated by commas
-    <object> ::= '{' [ <member> *(', ' <member>) ] '}' ; A sequence of 'members'
+    <object> ::= '{' [ <string> *(', ' <member>) ] '}' ; A sequence of 'members'
     <member> ::= <string> ': ' <json> ; A pair consisting of a name, and a JSON value
 
 If something is wrong (a missing parantheses or quotes for example) it will use a few simple heuristics to fix the JSON string:
@@ -55,16 +55,18 @@ class JSONParser:
         if char is False:
             return ""
         # <object> starts with '{'
-        elif char == "{":
+        # but an object key must be a string
+        elif self.get_context() != "object_key" and char == "{":
             self.index += 1
             return self.parse_object()
         # <array> starts with '['
-        elif char == "[":
+        # but an object key must be a string
+        elif self.get_context() != "object_key" and char == "[":
             self.index += 1
             return self.parse_array()
         # there can be an edge case in which a key is empty and at the end of an object
         # like "key": }. We return an empty string here to close the object properly
-        elif char == "}":
+        elif self.get_context() != "object_key" and char == "}":
             self.log(
                 "At the end of an object we found a key with missing value, skipping",
                 "info",
@@ -78,10 +80,20 @@ class JSONParser:
         elif char == "“":
             return self.parse_string(string_quotes=["“", "”"])
         # <number> starts with [0-9] or minus
-        elif self.get_context() != "" and char.isdigit() or char == "-" or char == ".":
+        elif (
+            self.get_context() != ""
+            and self.get_context() != "object_key"
+            and char.isdigit()
+            or char == "-"
+            or char == "."
+        ):
             return self.parse_number()
         # <boolean> could be (T)rue or (F)alse or (N)ull
-        elif self.get_context() != "" and char.lower() in ["t", "f", "n"]:
+        elif (
+            self.get_context() != ""
+            and self.get_context() != "object_key"
+            and char.lower() in ["t", "f", "n"]
+        ):
             return self.parse_boolean_or_null()
         # This might be a <string> that is missing the starting '"'
         elif self.get_context() != "" and char.isalpha():
@@ -302,7 +314,8 @@ class JSONParser:
                     while next_c and next_c != rstring_delimiter:
                         # If we are in an object context, let's check for the right delimiters
                         if (
-                            ("object_key" in self.context and next_c == ":")
+                            next_c == lstring_delimiter
+                            or ("object_key" in self.context and next_c == ":")
                             or ("object_value" in self.context and next_c in ["}", ","])
                             or ("array" in self.context and next_c in ["]", ","])
                         ):
@@ -310,12 +323,28 @@ class JSONParser:
                         i += 1
                         next_c = self.get_char_at(i)
                     if next_c == rstring_delimiter:
-                        self.log(
-                            "While parsing a string, we a misplaced quote that would have closed the string but has a different meaning here, ignoring it",
-                            "info",
-                        )
-                        self.index += 1
-                        char = self.get_char_at()
+                        # But this might not be it! This could be just a missing comma
+                        # We need to check if we find a rstring_delimiter and a colon after
+                        i += 1
+                        next_c = self.get_char_at(i)
+                        while next_c and next_c != rstring_delimiter:
+                            i += 1
+                            next_c = self.get_char_at(i)
+                        i += 1
+                        next_c = self.get_char_at(i)
+                        while next_c and next_c != ":":
+                            if next_c in [lstring_delimiter, rstring_delimiter, ","]:
+                                break
+                            i += 1
+                            next_c = self.get_char_at(i)
+                        # Only if we fail to find a ':' then we know this is misplaced quote
+                        if next_c != ":":
+                            self.log(
+                                "While parsing a string, we a misplaced quote that would have closed the string but has a different meaning here, ignoring it",
+                                "info",
+                            )
+                            self.index += 1
+                            char = self.get_char_at()
 
         if (
             char

{json_repair-0.16.0.dist-info → json_repair-0.16.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.16.0
+Version: 0.16.2
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License

json_repair-0.16.2.dist-info/RECORD

@@ -0,0 +1,7 @@
+json_repair/__init__.py,sha256=AlNie5y6BZBioGi5fzTAUvum_y0U5aL5aNsuQ_68LQc,175
+json_repair/json_repair.py,sha256=Z1BiZlCBWDGiZeARAMcQ-PYRJE5PHFeTDGLLTEVg4fs,21822
+json_repair-0.16.2.dist-info/LICENSE,sha256=wrjQo8MhNrNCicXtMe3MHmS-fx8AmQk1ue8AQwiiFV8,1076
+json_repair-0.16.2.dist-info/METADATA,sha256=8CXuXrM_3G5ti6wSef9waYRW3_ilW5lhQ6KIhrFRG80,7355
+json_repair-0.16.2.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+json_repair-0.16.2.dist-info/top_level.txt,sha256=7-VZwZN2CgB_n0NlSLk-rEUFh8ug21lESbsblOYuZqw,12
+json_repair-0.16.2.dist-info/RECORD,,

json_repair-0.16.0.dist-info/RECORD

@@ -1,7 +0,0 @@
-json_repair/__init__.py,sha256=AlNie5y6BZBioGi5fzTAUvum_y0U5aL5aNsuQ_68LQc,175
-json_repair/json_repair.py,sha256=XahLp82VVwg8KgyywNxMFBHIPIbvPDp6uWMP1VD_40w,20418
-json_repair-0.16.0.dist-info/LICENSE,sha256=wrjQo8MhNrNCicXtMe3MHmS-fx8AmQk1ue8AQwiiFV8,1076
-json_repair-0.16.0.dist-info/METADATA,sha256=-OLuKPGwu4enrP7kP2947Gyz9l7JRVpGTRH1b1MB2ZY,7355
-json_repair-0.16.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
-json_repair-0.16.0.dist-info/top_level.txt,sha256=7-VZwZN2CgB_n0NlSLk-rEUFh8ug21lESbsblOYuZqw,12
-json_repair-0.16.0.dist-info/RECORD,,