json-repair 0.50.1__tar.gz → 0.52.0__tar.gz

{json_repair-0.50.1/src/json_repair.egg-info → json_repair-0.52.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: json_repair
-Version: 0.50.1
+Version: 0.52.0
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License-Expression: MIT
@@ -269,6 +269,7 @@ You will need owner access to this repository
 - Ruby: https://github.com/sashazykov/json-repair-rb
 - Rust: https://github.com/oramasearch/llm_json
 - R: https://github.com/cgxjdzz/jsonRepair
+- Java: https://github.com/du00cs/json-repairj
 ---
 ## Star History
 

{json_repair-0.50.1 → json_repair-0.52.0}/README.md

@@ -252,6 +252,7 @@ You will need owner access to this repository
 - Ruby: https://github.com/sashazykov/json-repair-rb
 - Rust: https://github.com/oramasearch/llm_json
 - R: https://github.com/cgxjdzz/jsonRepair
+- Java: https://github.com/du00cs/json-repairj
 ---
 ## Star History
 

{json_repair-0.50.1 → json_repair-0.52.0}/pyproject.toml

@@ -3,7 +3,7 @@ requires = ["setuptools>=61.0"]
 build-backend = "setuptools.build_meta"
 [project]
 name = "json_repair"
-version = "0.50.1"
+version = "0.52.0"
 license = "MIT"
 license-files = ["LICENSE"]
 authors = [
@@ -26,12 +26,17 @@ pythonpath = [
   "."
 ]
 [tool.coverage.run]
+source = ["src"]
 omit = [
     "*/.cursor/extensions/*",
     "*/pythonFiles/lib/python/*",
     "*/site-packages/*",
+    "src/json_repair/__main__.py",
 ]
 [tool.coverage.report]
+include = [
+    "src/json_repair/*",
+]
 exclude_also = [
     'def __repr__',
     'if self.debug:',
@@ -111,4 +116,4 @@ line-ending = "auto"
 
 [tool.ruff.lint.per-file-ignores]
 # Explicit re-exports is fine in __init__.py, still a code smell elsewhere.
-"__init__.py" = ["PLC0414"]
+"__init__.py" = ["PLC0414"]

{json_repair-0.50.1 → json_repair-0.52.0}/src/json_repair/parse_object.py

@@ -1,6 +1,6 @@
 from typing import TYPE_CHECKING
 
-from .constants import JSONReturnType
+from .constants import STRING_DELIMITERS, JSONReturnType
 from .json_context import ContextValues
 
 if TYPE_CHECKING:
@@ -10,6 +10,7 @@ if TYPE_CHECKING:
 def parse_object(self: "JSONParser") -> dict[str, JSONReturnType]:
     # <object> ::= '{' [ <member> *(', ' <member>) ] '}' ; A sequence of 'members'
     obj: dict[str, JSONReturnType] = {}
+    start_index = self.index
     # Stop when you either find the closing parentheses or you have iterated over the entire string
     while (self.get_char_at() or "}") != "}":
         # This is what we expect to find:
@@ -112,4 +113,31 @@ def parse_object(self: "JSONParser") -> dict[str, JSONReturnType]:
         self.skip_whitespaces_at()
 
     self.index += 1
+
+    # If the object is empty but also isn't just {}
+    if not obj and self.index - start_index > 2:
+        self.log("Parsed object is empty, we will try to parse this as an array instead")
+        self.index = start_index
+        return self.parse_array()
+
+    # Check if there are more key-value pairs after the closing brace
+    # This handles cases like '{"key": "value"}, "key2": "value2"}'
+    # But only if we're not in a nested context
+    if not self.context.empty:
+        return obj
+
+    self.skip_whitespaces_at()
+    if (self.get_char_at() or "") != ",":
+        return obj
+    self.index += 1
+    self.skip_whitespaces_at()
+    if (self.get_char_at() or "") not in STRING_DELIMITERS:
+        return obj
+    self.log(
+        "Found a comma and string delimiter after object closing brace, checking for additional key-value pairs",
+    )
+    additional_obj = self.parse_object()
+    if isinstance(additional_obj, dict):
+        obj.update(additional_obj)
+
     return obj

{json_repair-0.50.1 → json_repair-0.52.0/src/json_repair.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: json_repair
-Version: 0.50.1
+Version: 0.52.0
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License-Expression: MIT
@@ -269,6 +269,7 @@ You will need owner access to this repository
 - Ruby: https://github.com/sashazykov/json-repair-rb
 - Rust: https://github.com/oramasearch/llm_json
 - R: https://github.com/cgxjdzz/jsonRepair
+- Java: https://github.com/du00cs/json-repairj
 ---
 ## Star History
 

{json_repair-0.50.1 → json_repair-0.52.0}/tests/test_parse_array.py

@@ -35,7 +35,11 @@ def test_parse_array_edge_cases():
     )
     assert repair_json('{"k"e"y": "value"}') == '{"k\\"e\\"y": "value"}'
     assert repair_json('["key":"value"}]') == '[{"key": "value"}]'
-    assert repair_json('[{"key": "value", "key') == '[{"key": "value"}]'
+    assert repair_json('[{"key": "value", "key') == '[{"key": "value"}, ["key"]]'
+    assert repair_json("{'key1', 'key2'}") == '["key1", "key2"]'
+
+
+def test_parse_array_missing_quotes():
     assert repair_json('["value1" value2", "value3"]') == '["value1", "value2", "value3"]'
     assert (
         repair_json('{"bad_one":["Lorem Ipsum", "consectetur" comment" ], "good_one":[ "elit", "sed", "tempor"]}')

{json_repair-0.50.1 → json_repair-0.52.0}/tests/test_parse_object.py

@@ -22,7 +22,6 @@ def test_parse_object():
 
 def test_parse_object_edge_cases():
     assert repair_json("{foo: [}") == '{"foo": []}'
-    assert repair_json("{       ") == "{}"
     assert repair_json('{"": "value"') == '{"": "value"}'
     assert repair_json('{"value_1": true, COMMENT "value_2": "data"}') == '{"value_1": true, "value_2": "data"}'
     assert (
@@ -83,3 +82,17 @@ def test_parse_object_edge_cases():
         == '{"key": "{\\"key\\":[\\"value\\"],\\"key2\\":\\"value2\\"}"}'
     )
     assert repair_json('{"key": , "key2": "value2"}') == '{"key": "", "key2": "value2"}'
+
+
+def test_parse_object_merge_at_the_end():
+    assert repair_json('{"key": "value"}, "key2": "value2"}') == '{"key": "value", "key2": "value2"}'
+    assert repair_json('{"key": "value"}, "key2": }') == '{"key": "value", "key2": ""}'
+    assert repair_json('{"key": "value"}, []') == '[{"key": "value"}, []]'
+    assert repair_json('{"key": "value"}, ["abc"]') == '[{"key": "value"}, ["abc"]]'
+    assert repair_json('{"key": "value"}, {}') == '[{"key": "value"}, {}]'
+    assert repair_json('{"key": "value"}, "" : "value2"}') == '{"key": "value", "": "value2"}'
+    assert repair_json('{"key": "value"}, "key2" "value2"}') == '{"key": "value", "key2": "value2"}'
+    assert (
+        repair_json('{"key1": "value1"}, "key2": "value2", "key3": "value3"}')
+        == '{"key1": "value1", "key2": "value2", "key3": "value3"}'
+    )

{json_repair-0.50.1 → json_repair-0.52.0}/tests/test_performance.py

@@ -1,9 +1,12 @@
 import os.path
 import pathlib
 
+import pytest
+
 from src.json_repair import repair_json
 
 path = pathlib.Path(__file__).parent.resolve()
+CI = os.getenv("CI") is not None
 
 with open(os.path.join(path, "valid.json")) as fd:
     correct_json = fd.read()
@@ -12,6 +15,7 @@ with open(os.path.join(path, "invalid.json")) as fd:
     incorrect_json = fd.read()
 
 
+@pytest.mark.skipif(CI, reason="Performance tests are skipped in CI")
 def test_true_true_correct(benchmark):
     benchmark(repair_json, correct_json, return_objects=True, skip_json_loads=True)
 
@@ -25,6 +29,7 @@ def test_true_true_correct(benchmark):
     assert mean_time < max_time, f"Benchmark exceeded threshold: {mean_time:.3f}s > {max_time:.3f}s"
 
 
+@pytest.mark.skipif(CI, reason="Performance tests are skipped in CI")
 def test_true_true_incorrect(benchmark):
     benchmark(repair_json, incorrect_json, return_objects=True, skip_json_loads=True)
 
@@ -38,6 +43,7 @@ def test_true_true_incorrect(benchmark):
     assert mean_time < max_time, f"Benchmark exceeded threshold: {mean_time:.3f}s > {max_time:.3f}s"
 
 
+@pytest.mark.skipif(CI, reason="Performance tests are skipped in CI")
 def test_true_false_correct(benchmark):
     benchmark(repair_json, correct_json, return_objects=True, skip_json_loads=False)
     # Retrieve the median execution time
@@ -50,6 +56,7 @@ def test_true_false_correct(benchmark):
     assert mean_time < max_time, f"Benchmark exceeded threshold: {mean_time:.3f}s > {max_time:.3f}s"
 
 
+@pytest.mark.skipif(CI, reason="Performance tests are skipped in CI")
 def test_true_false_incorrect(benchmark):
     benchmark(repair_json, incorrect_json, return_objects=True, skip_json_loads=False)
     # Retrieve the median execution time
@@ -62,6 +69,7 @@ def test_true_false_incorrect(benchmark):
     assert mean_time < max_time, f"Benchmark exceeded threshold: {mean_time:.3f}s > {max_time:.3f}s"
 
 
+@pytest.mark.skipif(CI, reason="Performance tests are skipped in CI")
 def test_false_true_correct(benchmark):
     benchmark(repair_json, correct_json, return_objects=False, skip_json_loads=True)
     # Retrieve the median execution time
@@ -74,6 +82,7 @@ def test_false_true_correct(benchmark):
     assert mean_time < max_time, f"Benchmark exceeded threshold: {mean_time:.3f}s > {max_time:.3f}s"
 
 
+@pytest.mark.skipif(CI, reason="Performance tests are skipped in CI")
 def test_false_true_incorrect(benchmark):
     benchmark(repair_json, incorrect_json, return_objects=False, skip_json_loads=True)
     # Retrieve the median execution time
@@ -86,6 +95,7 @@ def test_false_true_incorrect(benchmark):
     assert mean_time < max_time, f"Benchmark exceeded threshold: {mean_time:.3f}s > {max_time:.3f}s"
 
 
+@pytest.mark.skipif(CI, reason="Performance tests are skipped in CI")
 def test_false_false_correct(benchmark):
     benchmark(repair_json, correct_json, return_objects=False, skip_json_loads=False)
     # Retrieve the median execution time
@@ -98,6 +108,7 @@ def test_false_false_correct(benchmark):
     assert mean_time < max_time, f"Benchmark exceeded threshold: {mean_time:.3f}s > {max_time:.3f}s"
 
 
+@pytest.mark.skipif(CI, reason="Performance tests are skipped in CI")
 def test_false_false_incorrect(benchmark):
     benchmark(repair_json, incorrect_json, return_objects=False, skip_json_loads=False)
     # Retrieve the median execution time