json-repair 0.24.0__py3-none-any.whl → 0.25.1__py3-none-any.whl

json_repair/json_repair.py

@@ -119,40 +119,38 @@ class JSONParser:
     def parse_json(
         self,
     ) -> JSONReturnType:
-        char = self.get_char_at()
-        # This parser will ignore any basic element (string or number) that is not inside an array or object
-        is_in_context = len(self.context) > 0
-        # False means that we are at the end of the string provided, is the base case for recursion
-        if char is False:
-            return ""
-        # <object> starts with '{'
-        # but an object key must be a string
-        elif char == "{":
-            self.index += 1
-            return self.parse_object()
-        # <array> starts with '['
-        # but an object key must be a string
-        elif 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 == "}":
-            self.log(
-                "At the end of an object we found a key with missing value, skipping",
-                "info",
-            )
-            return ""
-        # <string> starts with a quote
-        elif is_in_context and (char in ['"', "'", "“"] or char.isalpha()):
-            return self.parse_string()
-        # <number> starts with [0-9] or minus
-        elif is_in_context and (char.isdigit() or char == "-" or char == "."):
-            return self.parse_number()
-        # If everything else fails, we just ignore and move on
-        else:
-            self.index += 1
-            return self.parse_json()
+        while True:
+            char = self.get_char_at()
+            # This parser will ignore any basic element (string or number) that is not inside an array or object
+            is_in_context = len(self.context) > 0
+            # False means that we are at the end of the string provided
+            if char is False:
+                return ""
+            # <object> starts with '{'
+            elif char == "{":
+                self.index += 1
+                return self.parse_object()
+            # <array> starts with '['
+            elif 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 == "}":
+                self.log(
+                    "At the end of an object we found a key with missing value, skipping",
+                    "info",
+                )
+                return ""
+            # <string> starts with a quote
+            elif is_in_context and (char in ['"', "'", "“"] or char.isalpha()):
+                return self.parse_string()
+            # <number> starts with [0-9] or minus
+            elif is_in_context and (char.isdigit() or char == "-" or char == "."):
+                return self.parse_number()
+            # If everything else fails, we just ignore and move on
+            else:
+                self.index += 1
 
     def parse_object(self) -> Dict[str, Any]:
         # <object> ::= '{' [ <member> *(', ' <member>) ] '}' ; A sequence of 'members'
@@ -370,6 +368,10 @@ class JSONParser:
                 if self.get_context() == "object_key" and (
                     char == ":" or char.isspace()
                 ):
+                    self.log(
+                        "While parsing a string missing the left delimiter in object key context, we found a :, stopping here",
+                        "info",
+                    )
                     break
                 elif self.get_context() == "object_value" and char in [",", "}"]:
                     rstring_delimiter_missing = True
@@ -389,6 +391,10 @@ class JSONParser:
                         if next_c and next_c in [",", "}"]:
                             rstring_delimiter_missing = False
                     if rstring_delimiter_missing:
+                        self.log(
+                            "While parsing a string missing the left delimiter in object value context, we found a , or } and we couldn't determine that a right delimiter was present. Stopping here",
+                            "info",
+                        )
                         break
             string_acc += char
             self.index += 1
@@ -410,6 +416,33 @@ class JSONParser:
                         "While parsing a string, we found a doubled quote, ignoring it",
                         "info",
                     )
+                elif missing_quotes and self.get_context() == "object_value":
+                    # In case of missing starting quote I need to check if the delimeter is the end or the beginning of a key
+                    i = 1
+                    next_c = self.get_char_at(i)
+                    while next_c and next_c not in [
+                        rstring_delimiter,
+                        lstring_delimiter,
+                    ]:
+                        i += 1
+                        next_c = self.get_char_at(i)
+                    if next_c:
+                        # We found a quote, now let's make sure there's a ":" following
+                        i += 1
+                        next_c = self.get_char_at(i)
+                        # found a delimiter, now we need to check that is followed strictly by a comma or brace
+                        while next_c and next_c.isspace():
+                            i += 1
+                            next_c = self.get_char_at(i)
+                        if next_c and next_c == ":":
+                            # Reset the cursor
+                            self.index -= 1
+                            char = self.get_char_at()
+                            self.log(
+                                "In a string with missing quotes and object value context, I found a delimeter but it turns out it was the beginning on the next key. Stopping here.",
+                                "info",
+                            )
+                            break
                 else:
                     # Check if eventually there is a rstring delimiter, otherwise we bail
                     i = 1

{json_repair-0.24.0.dist-info → json_repair-0.25.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.24.0
+Version: 0.25.1
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License

json_repair-0.25.1.dist-info/RECORD

@@ -0,0 +1,7 @@
+json_repair/__init__.py,sha256=IIzSm1DsCRrr8seF3UeMZXwxcq-tE3j-8d1WBxvEJvE,178
+json_repair/json_repair.py,sha256=PbfthTey1p04_9Q3QRGm8We1ueFVYwO5kQUShpfR10Y,30961
+json_repair-0.25.1.dist-info/LICENSE,sha256=wrjQo8MhNrNCicXtMe3MHmS-fx8AmQk1ue8AQwiiFV8,1076
+json_repair-0.25.1.dist-info/METADATA,sha256=QjbXE09AIfd70scALY8DPIBxg0TYTBlKg3KuJ7MAMRw,7596
+json_repair-0.25.1.dist-info/WHEEL,sha256=cpQTJ5IWu9CdaPViMhC9YzF8gZuS5-vlfoFihTBC86A,91
+json_repair-0.25.1.dist-info/top_level.txt,sha256=7-VZwZN2CgB_n0NlSLk-rEUFh8ug21lESbsblOYuZqw,12
+json_repair-0.25.1.dist-info/RECORD,,

{json_repair-0.24.0.dist-info → json_repair-0.25.1.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: bdist_wheel (0.43.0)
+Generator: setuptools (70.1.0)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

json_repair-0.24.0.dist-info/RECORD

@@ -1,7 +0,0 @@
-json_repair/__init__.py,sha256=IIzSm1DsCRrr8seF3UeMZXwxcq-tE3j-8d1WBxvEJvE,178
-json_repair/json_repair.py,sha256=jM509L0rg8AiWksnYdOqueopP8fmh0MJxpxx7LvaoiM,28954
-json_repair-0.24.0.dist-info/LICENSE,sha256=wrjQo8MhNrNCicXtMe3MHmS-fx8AmQk1ue8AQwiiFV8,1076
-json_repair-0.24.0.dist-info/METADATA,sha256=kSxFAdkH_qxJMX85DKW605BOZrbh8VCWzsAvP_X80cM,7596
-json_repair-0.24.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
-json_repair-0.24.0.dist-info/top_level.txt,sha256=7-VZwZN2CgB_n0NlSLk-rEUFh8ug21lESbsblOYuZqw,12
-json_repair-0.24.0.dist-info/RECORD,,