json-repair 0.39.0__py3-none-any.whl → 0.39.1__py3-none-any.whl

json_repair/json_parser.py

@@ -9,6 +9,7 @@ JSONReturnType = Union[Dict[str, Any], List[Any], str, float, int, bool, None]
 class JSONParser:
     # Constants
     STRING_DELIMITERS = ['"', "'", "“", "”"]
+    NUMBER_CHARS = set("0123456789-.eE/,")
 
     def __init__(
         self,
@@ -129,8 +130,6 @@ class JSONParser:
             # Context is used in the string parser to manage the lack of quotes
             self.context.set(ContextValues.OBJECT_KEY)
 
-            self.skip_whitespaces_at()
-
             # Save this index in case we need find a duplicate key
             rollback_index = self.index
 
@@ -219,18 +218,13 @@ class JSONParser:
                 char = self.get_char_at()
 
         # Especially at the end of an LLM generated json you might miss the last "]"
-        char = self.get_char_at()
         if char and char != "]":
             self.log(
-                "While parsing an array we missed the closing ], adding it back",
-            )
-            self.index -= 1
-            # Add the missing closing bracket
-            self.json_str = (
-                self.json_str[: self.index + 1] + "]" + self.json_str[self.index + 1 :]
+                "While parsing an array we missed the closing ], ignoring it",
             )
 
         self.index += 1
+
         self.context.reset()
         return arr
 
@@ -275,15 +269,11 @@ class JSONParser:
             self.log(
                 "While parsing a string, we found a literal instead of a quote",
             )
-            self.log(
-                "While parsing a string, we found no starting quote. Will add the quote back",
-            )
             missing_quotes = True
 
         if not missing_quotes:
             self.index += 1
 
-        self.skip_whitespaces_at()
         # There is sometimes a weird case of doubled quotes, we manage this also later in the while loop
         if self.get_char_at() in self.STRING_DELIMITERS:
             # If the next character is the same type of quote, then we manage it as double quotes
@@ -583,6 +573,13 @@ class JSONParser:
                     elif (
                         next_c == rstring_delimiter and self.get_char_at(i - 1) != "\\"
                     ):
+                        # Check if self.index:self.index+i is only whitespaces, break if that's the case
+                        if all(
+                            str(self.get_char_at(j)).isspace()
+                            for j in range(1, i)
+                            if self.get_char_at(j)
+                        ):
+                            break
                         if self.context.current == ContextValues.OBJECT_VALUE:
                             # But this might not be it! This could be just a missing comma
                             # We found a delimiter and we need to check if this is a key
@@ -610,26 +607,16 @@ class JSONParser:
                                 self.index += 1
                                 char = self.get_char_at()
                         elif self.context.current == ContextValues.ARRAY:
-                            # In array context this could be something like "lorem "ipsum" sic"
-                            # So let's check if we find a rstring_delimiter forward otherwise end early
-                            i = self.skip_to_character(rstring_delimiter, idx=i + 1)
-                            next_c = self.get_char_at(i)
-                            if next_c and next_c == rstring_delimiter:
-                                # Ok now if I find a comma or a closing ], that can be have also an optional rstring_delimiter before them
-                                # We can consider this a misplaced quote
-                                i += 1
-                                i = self.skip_whitespaces_at(
-                                    idx=i, move_main_index=False
-                                )
-                                next_c = self.get_char_at(i)
-                                if next_c and next_c in [",", "]"]:
-                                    self.log(
-                                        "While parsing a string, we a misplaced quote that would have closed the string but has a different meaning here, ignoring it",
-                                    )
-                                    unmatched_delimiter = not unmatched_delimiter
-                                    string_acc += str(char)
-                                    self.index += 1
-                                    char = self.get_char_at()
+                            # If we got up to here it means that this is a situation like this:
+                            # ["bla bla bla "puppy" bla bla bla "kitty" bla bla"]
+                            # So we need to ignore this quote
+                            self.log(
+                                "While parsing a string in Array context, we detected a quoted section that would have closed the string but has a different meaning here, ignoring it",
+                            )
+                            unmatched_delimiter = not unmatched_delimiter
+                            string_acc += str(char)
+                            self.index += 1
+                            char = self.get_char_at()
 
         if (
             char
@@ -663,10 +650,9 @@ class JSONParser:
     def parse_number(self) -> Union[float, int, str, JSONReturnType]:
         # <number> is a valid real number expressed in one of a number of given formats
         number_str = ""
-        number_chars = set("0123456789-.eE/,")
         char = self.get_char_at()
         is_array = self.context.current == ContextValues.ARRAY
-        while char and char in number_chars and (char != "," or not is_array):
+        while char and char in self.NUMBER_CHARS and (not is_array or char != ","):
             number_str += char
             self.index += 1
             char = self.get_char_at()

{json_repair-0.39.0.dist-info → json_repair-0.39.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: json_repair
-Version: 0.39.0
+Version: 0.39.1
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License
@@ -196,12 +196,12 @@ pipx install json-repair
 to know all options available:
 ```
 $ json_repair -h
-usage: json_repair [-h] [-i] [-o TARGET] [--ensure_ascii] [--indent INDENT] filename
+usage: json_repair [-h] [-i] [-o TARGET] [--ensure_ascii] [--indent INDENT] [filename]
 
 Repair and parse JSON files.
 
 positional arguments:
-  filename              The JSON file to repair
+  filename              The JSON file to repair (if omitted, reads from stdin)
 
 options:
   -h, --help            show this help message and exit
@@ -226,13 +226,13 @@ In this example, any version that starts with `0.` will be acceptable, allowing
 # How to cite
 If you are using this library in your academic work (as I know many folks are) please find the BibTex here:
 
-    @software{Baccianella_JSON_Repair_-_2024,
+    @software{Baccianella_JSON_Repair_-_2025,
         author = {Baccianella, Stefano},
-        month = aug,
+        month = feb,
         title = {{JSON Repair - A python module to repair invalid JSON, commonly used to parse the output of LLMs}},
         url = {https://github.com/mangiucugna/json_repair},
-        version = {0.28.3},
-        year = {2024}
+        version = {0.39.0},
+        year = {2025}
     }
 
 Thank you for citing my work and please send me a link to the paper if you can!

{json_repair-0.39.0.dist-info → json_repair-0.39.1.dist-info}/RECORD

@@ -1,13 +1,13 @@
 json_repair/__init__.py,sha256=c4L2kZrHvWEKfj_ODU2naliNuvU6FlFVxtF0hbLe6s8,178
 json_repair/__main__.py,sha256=EsJb-y89uZEvGQQg1GdIDWzfDwfOMvVekKEtdguQXCM,67
 json_repair/json_context.py,sha256=mm6dOyrPJ1sDskTORZSXCW7W9-5veMlUKqXQ3Hw3EG4,971
-json_repair/json_parser.py,sha256=BQsH8CRy59C2176bMwVerfqbHDXfLoEC1v5frmCiv7M,39020
+json_repair/json_parser.py,sha256=kt58S7pHxCOfqktzn48iMrvd3vi7HTfK6OD02PWwWcc,38189
 json_repair/json_repair.py,sha256=k-5HRRlCqrxNmJi0u1KE3IUeL4HXqi1XZ7oAL-NFDLo,10314
 json_repair/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 json_repair/string_file_wrapper.py,sha256=koZmdq2-Z5K7XF1bDqX6dEbNaVMJYcMTjq-aGe6NQvA,4526
-json_repair-0.39.0.dist-info/LICENSE,sha256=wrjQo8MhNrNCicXtMe3MHmS-fx8AmQk1ue8AQwiiFV8,1076
-json_repair-0.39.0.dist-info/METADATA,sha256=cArvqcMBL9FVCwnJGtsaeF7lXWjOFWG3_1OueGjOiRs,11794
-json_repair-0.39.0.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
-json_repair-0.39.0.dist-info/entry_points.txt,sha256=SNfge3zPSP-ASqriYU9r3NAPaXdseYr7ciPMKdV2uSw,57
-json_repair-0.39.0.dist-info/top_level.txt,sha256=7-VZwZN2CgB_n0NlSLk-rEUFh8ug21lESbsblOYuZqw,12
-json_repair-0.39.0.dist-info/RECORD,,
+json_repair-0.39.1.dist-info/LICENSE,sha256=wrjQo8MhNrNCicXtMe3MHmS-fx8AmQk1ue8AQwiiFV8,1076
+json_repair-0.39.1.dist-info/METADATA,sha256=T1k1afyqqWG-NXYHpgntQsPYpmIdkLg72eSo_iNwZZk,11827
+json_repair-0.39.1.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+json_repair-0.39.1.dist-info/entry_points.txt,sha256=SNfge3zPSP-ASqriYU9r3NAPaXdseYr7ciPMKdV2uSw,57
+json_repair-0.39.1.dist-info/top_level.txt,sha256=7-VZwZN2CgB_n0NlSLk-rEUFh8ug21lESbsblOYuZqw,12
+json_repair-0.39.1.dist-info/RECORD,,