json-repair 0.38.0__tar.gz → 0.39.1__tar.gz

{json_repair-0.38.0/src/json_repair.egg-info → json_repair-0.39.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: json_repair
-Version: 0.38.0
+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-0.38.0 → json_repair-0.39.1}/README.md

@@ -158,12 +158,12 @@ 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
@@ -188,13 +188,13 @@ 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-0.38.0 → json_repair-0.39.1}/pyproject.toml

@@ -3,7 +3,7 @@ requires = ["setuptools>=61.0"]
 build-backend = "setuptools.build_meta"
 [project]
 name = "json_repair"
-version = "0.38.0"
+version = "0.39.1"
 license = {file = "LICENSE"}
 authors = [
   { name="Stefano Baccianella", email="4247706+mangiucugna@users.noreply.github.com" },

{json_repair-0.38.0 → json_repair-0.39.1}/src/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-0.38.0 → json_repair-0.39.1}/src/json_repair/json_repair.py

@@ -160,7 +160,7 @@ def cli(inline_args: Optional[List[str]] = None) -> int:
 
     Args:
         inline_args (Optional[List[str]]): List of command-line arguments for testing purposes. Defaults to None.
-            - filename (str): The JSON file to repair
+            - filename (str): The JSON file to repair. If omitted, the JSON is read from stdin.
             - -i, --inline (bool): Replace the file inline instead of returning the output to stdout.
             - -o, --output TARGET (str): If specified, the output will be written to TARGET filename instead of stdout.
             - --ensure_ascii (bool): Pass ensure_ascii=True to json.dumps(). Will pass False otherwise.
@@ -174,9 +174,15 @@ def cli(inline_args: Optional[List[str]] = None) -> int:
 
     Example:
         >>> cli(['example.json', '--indent', '4'])
+        >>> cat json.txt | json_repair
     """
     parser = argparse.ArgumentParser(description="Repair and parse JSON files.")
-    parser.add_argument("filename", help="The JSON file to repair")
+    # Make the filename argument optional; if omitted, we will read from stdin.
+    parser.add_argument(
+        "filename",
+        nargs="?",
+        help="The JSON file to repair (if omitted, reads from stdin)",
+    )
     parser.add_argument(
         "-i",
         "--inline",
@@ -204,9 +210,12 @@ def cli(inline_args: Optional[List[str]] = None) -> int:
     if inline_args is None:  # pragma: no cover
         args = parser.parse_args()
     else:
-        args = parser.parse_args(
-            inline_args
-        )  # This is needed so this function is testable
+        args = parser.parse_args(inline_args)
+
+    # Inline mode requires a filename, so error out if none was provided.
+    if args.inline and not args.filename:  # pragma: no cover
+        print("Error: Inline mode requires a filename", file=sys.stderr)
+        sys.exit(1)
 
     if args.inline and args.output:  # pragma: no cover
         print("Error: You cannot pass both --inline and --output", file=sys.stderr)
@@ -217,8 +226,12 @@ def cli(inline_args: Optional[List[str]] = None) -> int:
         ensure_ascii = True
 
     try:
-        result = from_file(args.filename)
-
+        # Use from_file if a filename is provided; otherwise read from stdin.
+        if args.filename:
+            result = from_file(args.filename)
+        else:
+            data = sys.stdin.read()
+            result = loads(data)
         if args.inline or args.output:
             with open(args.output or args.filename, mode="w") as fd:
                 json.dump(result, fd, indent=args.indent, ensure_ascii=ensure_ascii)

{json_repair-0.38.0 → json_repair-0.39.1/src/json_repair.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: json_repair
-Version: 0.38.0
+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-0.38.0 → json_repair-0.39.1}/tests/test_json_repair.py

@@ -3,6 +3,7 @@ from unittest.mock import patch
 import os.path
 import pathlib
 import tempfile
+import io
 
 def test_basic_types_valid():
     assert repair_json("True", return_objects=True) == ""
@@ -125,6 +126,8 @@ def test_array_edge_cases():
     assert repair_json('["lorem "ipsum" sic"]') == '["lorem \\"ipsum\\" sic"]'
     assert repair_json('{"key1": ["value1", "value2"}, "key2": ["value3", "value4"]}') == '{"key1": ["value1", "value2"], "key2": ["value3", "value4"]}'
     assert repair_json('[ "value", /* comment */ "value2" ]') == '["value", "value2"]'
+    assert repair_json('{"key": ["value" "value1" "value2"]}') == '{"key": ["value", "value1", "value2"]}'
+    assert repair_json('{"key": ["lorem "ipsum" dolor "sit" amet, "consectetur" ", "lorem "ipsum" dolor", "lorem"]}') == '{"key": ["lorem \\"ipsum\\" dolor \\"sit\\" amet, \\"consectetur\\" ", "lorem \\"ipsum\\" dolor", "lorem"]}'
 
 def test_escaping():
     assert repair_json("'\"'") == '""'
@@ -149,7 +152,7 @@ def test_object_edge_cases():
     assert repair_json('{"lorem": ipsum, sic, datum.",}') == '{"lorem": "ipsum, sic, datum."}'
     assert repair_json('{"lorem": sic tamet. "ipsum": sic tamet, quick brown fox. "sic": ipsum}') == '{"lorem": "sic tamet.", "ipsum": "sic tamet", "sic": "ipsum"}'
     assert repair_json('{"lorem_ipsum": "sic tamet, quick brown fox. }') == '{"lorem_ipsum": "sic tamet, quick brown fox."}'
-    assert repair_json('{"key":value, " key2":"value2" }') == '{"key": "value", "key2": "value2"}'
+    assert repair_json('{"key":value, " key2":"value2" }') == '{"key": "value", " key2": "value2"}'
     assert repair_json('{"key":value "key2":"value2" }') == '{"key": "value", "key2": "value2"}'
     assert repair_json("{'text': 'words{words in brackets}more words'}") == '{"text": "words{words in brackets}more words"}'
     assert repair_json('{text:words{words in brackets}}') == '{"text": "words{words in brackets}"}'
@@ -264,8 +267,8 @@ def test_repair_json_from_file():
         # Write content to the temporary file
         with os.fdopen(temp_fd, 'w') as tmp:
             tmp.write("{key:value}")
-        assert from_file(filename=temp_path, logging=True) == ({'key': 'value'}, [{'text': 'While parsing a string, we found a literal instead of a quote', 'context': '{key:value}'}, {'text': 'While parsing a string, we found no starting quote. Will add the quote back', 'context': '{key:value}'}, {'context': '{key:value}', 'text': 'While parsing a string missing the left delimiter in object key context, we found a :, stopping here',}, {'text': 'While parsing a string, we missed the closing quote, ignoring', 'context': '{key:value}'}, {'text': 'While parsing a string, we found a literal instead of a quote', 'context': '{key:value}'}, {'text': 'While parsing a string, we found no starting quote. Will add the quote back', 'context': '{key:value}'}, {'context': '{key:value}', 'text': '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'}, {'text': 'While parsing a string, we missed the closing quote, ignoring', 'context': '{key:value}'}])
-        assert from_file(filename=temp_path, logging=True, chunk_length=2) == ({'key': 'value'}, [{'text': 'While parsing a string, we found a literal instead of a quote', 'context': '{key:value}'}, {'text': 'While parsing a string, we found no starting quote. Will add the quote back', 'context': '{key:value}'}, {'context': '{key:value}', 'text': 'While parsing a string missing the left delimiter in object key context, we found a :, stopping here',}, {'text': 'While parsing a string, we missed the closing quote, ignoring', 'context': '{key:value}'}, {'text': 'While parsing a string, we found a literal instead of a quote', 'context': '{key:value}'}, {'text': 'While parsing a string, we found no starting quote. Will add the quote back', 'context': '{key:value}'}, {'context': '{key:value}', 'text': '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'}, {'text': 'While parsing a string, we missed the closing quote, ignoring', 'context': '{key:value}'}])
+        assert from_file(filename=temp_path, logging=True) == ({'key': 'value'}, [{'text': 'While parsing a string, we found a literal instead of a quote', 'context': '{key:value}'}, {'context': '{key:value}', 'text': 'While parsing a string missing the left delimiter in object key context, we found a :, stopping here',}, {'text': 'While parsing a string, we missed the closing quote, ignoring', 'context': '{key:value}'}, {'text': 'While parsing a string, we found a literal instead of a quote', 'context': '{key:value}'}, {'context': '{key:value}', 'text': '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'}, {'text': 'While parsing a string, we missed the closing quote, ignoring', 'context': '{key:value}'}])
+        assert from_file(filename=temp_path, logging=True, chunk_length=2) == ({'key': 'value'}, [{'text': 'While parsing a string, we found a literal instead of a quote', 'context': '{key:value}'}, {'context': '{key:value}', 'text': 'While parsing a string missing the left delimiter in object key context, we found a :, stopping here',}, {'text': 'While parsing a string, we missed the closing quote, ignoring', 'context': '{key:value}'}, {'text': 'While parsing a string, we found a literal instead of a quote', 'context': '{key:value}'}, {'context': '{key:value}', 'text': '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'}, {'text': 'While parsing a string, we missed the closing quote, ignoring', 'context': '{key:value}'}])
     finally:
         # Clean up - delete the temporary file
         os.remove(temp_path)
@@ -316,4 +319,14 @@ def test_cli(capsys):
     finally:
         # Clean up - delete the temporary file
         os.remove(temp_path)
-        os.remove(tempout_path)
+        os.remove(tempout_path)
+    
+    # Prepare a JSON string that needs to be repaired.
+    test_input = "{key:value"
+    # Expected output when running cli with --indent 0.
+    expected_output = '{\n"key": "value"\n}\n'
+    # Patch sys.stdin so that cli() reads from it instead of a file.
+    with patch('sys.stdin', new=io.StringIO(test_input)):
+        cli(inline_args=['--indent', 0])
+    captured = capsys.readouterr()
+    assert captured.out == expected_output