Fix #154, sometimes when repairing a confusing delimiter the poor closing brace was involved. Try not do that anymore

Author: mangiucugna

.github/workflows/python-package.yml

@@ -17,7 +17,7 @@ jobs:
       fail-fast: false
       matrix:
       # https://github.com/actions/python-versions/blob/main/versions-manifest.json
-        python-version: ["3.10", "3.11", "3.12", "3.13", "3.14.0-rc.2"]
+        python-version: ["3.10", "3.11", "3.12", "3.13", "3.14"]
 
     steps:
     - uses: actions/checkout@v5

.pre-commit-config.yaml

@@ -8,7 +8,7 @@ repos:
     pass_filenames: false
 - repo: https://github.com/astral-sh/ruff-pre-commit
   # Ruff version.
-  rev: v0.13.3
+  rev: v0.14.1
   hooks:
     # Run the linter.
     - id: ruff-check
@@ -38,7 +38,7 @@ repos:
     pass_filenames: false
     types: [python]
 - repo: https://github.com/semgrep/pre-commit
-  rev: "v1.139.0"
+  rev: "v1.140.0"
   hooks:
     - id: semgrep
       args:

pyproject.toml

@@ -3,7 +3,7 @@ requires = ["setuptools>=61.0"]
 build-backend = "setuptools.build_meta"
 [project]
 name = "json_repair"
-version = "0.52.0"
+version = "0.52.1"
 license = "MIT"
 license-files = ["LICENSE"]
 authors = [

src/json_repair/json_parser.py

@@ -155,7 +155,7 @@ def skip_whitespaces_at(self, idx: int = 0, move_main_index=True) -> int:
                 return idx
         return idx
 
-    def skip_to_character(self, character: str | list, idx: int = 0) -> int:
+    def skip_to_character(self, character: str | list[str], idx: int = 0) -> int:
         """
         This function quickly iterates to find a character, syntactic sugar to make the code more concise
         """

src/json_repair/parse_string.py

@@ -1,13 +1,13 @@
 from typing import TYPE_CHECKING
 
-from .constants import STRING_DELIMITERS
+from .constants import STRING_DELIMITERS, JSONReturnType
 from .json_context import ContextValues
 
 if TYPE_CHECKING:
     from .json_parser import JSONParser
 
 
-def parse_string(self: "JSONParser") -> str | bool | None:
+def parse_string(self: "JSONParser") -> JSONReturnType:
     # <string> is a string of valid characters enclosed in quotes
     # i.e. { name: "John" }
     # Somehow all weird cases in an invalid JSON happen to be resolved in this function, so be careful here
@@ -201,6 +201,16 @@ def parse_string(self: "JSONParser") -> str | bool | None:
             if not self.get_char_at(i):
                 # No delimiter found
                 break
+        if self.context.current == ContextValues.OBJECT_VALUE and char == "}":
+            # We found the end of an object while parsing a value
+            # Check if the object is really over, to avoid doubling the closing brace
+            i = self.skip_whitespaces_at(idx=1, move_main_index=False)
+            next_c = self.get_char_at(i)
+            if not next_c:
+                self.log(
+                    "While parsing a string in object value context, we found a } that closes the object, stopping here",
+                )
+                break
         string_acc += char
         self.index += 1
         char = self.get_char_at()

tests/test_parse_object.py

@@ -23,6 +23,7 @@ def test_parse_object():
 def test_parse_object_edge_cases():
     assert repair_json("{foo: [}") == '{"foo": []}'
     assert repair_json('{"": "value"') == '{"": "value"}'
+    assert repair_json('{"key": "v"alue"}') == '{"key": "v\\"alue\\""}'
     assert repair_json('{"value_1": true, COMMENT "value_2": "data"}') == '{"value_1": true, "value_2": "data"}'
     assert (
         repair_json('{"value_1": true, SHOULD_NOT_EXIST "value_2": "data" AAAA }')