json-repair 0.6.1__tar.gz → 0.6.2__tar.gz

{json_repair-0.6.1/src/json_repair.egg-info → json_repair-0.6.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.6.1
+Version: 0.6.2
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License

{json_repair-0.6.1 → json_repair-0.6.2}/pyproject.toml

@@ -3,7 +3,7 @@ requires = ["setuptools>=61.0"]
 build-backend = "setuptools.build_meta"
 [project]
 name = "json_repair"
-version = "0.6.1"
+version = "0.6.2"
 license = {file = "LICENSE"}
 authors = [
   { name="Stefano Baccianella", email="4247706+mangiucugna@users.noreply.github.com" },

{json_repair-0.6.1 → json_repair-0.6.2}/src/json_repair/json_repair.py

@@ -207,17 +207,18 @@ class JSONParser:
             if fixed_quotes:
                 if self.context == "object_key" and (char == ":" or char.isspace()):
                     break
-                elif self.context == "object_value" and (char == "," or char == "}"):
+                elif self.context == "object_value" and char in [",", "}"]:
                     break
             self.index += 1
             char = self.get_char_at()
             # ChatGPT sometimes forget to quote links in markdown like: { "content": "[LINK]("https://google.com")" }
             if (
                 char == string_terminator
-                and self.get_next_char() != ","
+                # Next character is not a comma
+                and self.get_char_at(1) != ","
                 and (
                     fix_broken_markdown_link
-                    or (self.get_prev_char(2) + self.get_prev_char()) == "]("
+                    or (self.get_char_at(-2) == "]" and self.get_char_at(-1)) == "("
                 )
             ):
                 fix_broken_markdown_link = not fix_broken_markdown_link
@@ -239,7 +240,7 @@ class JSONParser:
 
         return self.json_str[start:end]
 
-    def parse_number(self) -> Union[float, int]:
+    def parse_number(self) -> Union[float, int, str]:
         # <number> is a valid real number expressed in one of a number of given formats
         number_str = ""
         number_chars = set("0123456789-.eE")
@@ -257,7 +258,7 @@ class JSONParser:
             # This is a string then
             return self.parse_string()
 
-    def parse_boolean_or_null(self) -> Union[bool, None]:
+    def parse_boolean_or_null(self) -> Union[bool, str, None]:
         # <boolean> is one of the literal strings 'true', 'false', or 'null' (unquoted)
         boolean_map = {"true": (True, 4), "false": (False, 5), "null": (None, 4)}
         for key, (value, length) in boolean_map.items():
@@ -272,21 +273,7 @@ class JSONParser:
         self.json_str = self.json_str[: self.index] + char + self.json_str[self.index :]
         self.index += 1
 
-    def get_char_at(self) -> Union[str, bool]:
-        # Why not use something simpler? Because we might be out of bounds and doing this check all the time is annoying
-        try:
-            return self.json_str[self.index]
-        except IndexError:
-            return False
-
-    def get_prev_char(self, count=1):
-        # Why not use something simpler? Because we might be out of bounds and doing this check all the time is annoying
-        try:
-            return self.json_str[self.index - count]
-        except IndexError:
-            return False
-
-    def get_next_char(self, count=1):
+    def get_char_at(self, count: int = 0) -> Union[str, bool]:
         # Why not use something simpler? Because we might be out of bounds and doing this check all the time is annoying
         try:
             return self.json_str[self.index + count]
@@ -320,8 +307,7 @@ def repair_json(
     It will return the fixed string by default.
     When `return_objects=True` is passed, it will return the decoded data structure instead.
     """
-    json_str = re.sub(r"^\s+", "", json_str)
-    json_str = re.sub(r"\s+$", "", json_str)
+    json_str = json_str.strip()
     json_str = re.sub(r"/\*.*?\*/", "", json_str)
     parser = JSONParser(json_str)
     if skip_json_loads:

{json_repair-0.6.1 → json_repair-0.6.2/src/json_repair.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.6.1
+Version: 0.6.2
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License

{json_repair-0.6.1 → json_repair-0.6.2}/tests/test_json_repair.py

@@ -7,8 +7,9 @@ def test_repair_json():
     assert repair_json("{}") == "{}"
     assert repair_json("\"") == '""'
     assert repair_json("\n") == '""'
-    assert repair_json('{"key": true, "key2": false, "key3": null}') == '{"key": true, "key2": false, "key3": null}'
-    assert repair_json('{"key": TRUE, "key2": FALSE, "key3": Null}') == '{"key": true, "key2": false, "key3": null}'
+    assert repair_json("   /* COMMENT */    ") == '""'
+    assert repair_json('  {"key": true, "key2": false, "key3": null}') == '{"key": true, "key2": false, "key3": null}'
+    assert repair_json('{"key": TRUE, "key2": FALSE, "key3": Null}   ') == '{"key": true, "key2": false, "key3": null}'
     assert repair_json("{'key': 'string', 'key2': false, \"key3\": null, \"key4\": unquoted}") == '{"key": "string", "key2": false, "key3": null, "key4": "unquoted"}'
     assert (
         repair_json('{"name": "John", "age": 30, "city": "New York"}')

{json_repair-0.6.1 → json_repair-0.6.2}/tests/test_performance.py

@@ -282,7 +282,7 @@ def test_true_true(benchmark):
   mean_time = benchmark.stats.get("median")
 
   # Define your time threshold in seconds (100ms in this case)
-  max_time = 1.1 / 10 ** 6  # 1.1 microsecond
+  max_time = 0.7 / 10 ** 6  # 0.7 microsecond
 
   # Assert that the average time is below the threshold
   assert mean_time < max_time, f"Benchmark exceeded threshold: {mean_time:.3f}s > {max_time:.3f}s"
@@ -293,7 +293,7 @@ def test_true_false(benchmark):
   mean_time = benchmark.stats.get("median")
 
   # Define your time threshold in seconds (100ms in this case)
-  max_time = 180 * (1 / 10 ** 6)  # 180 microsecond
+  max_time = 24 * (1 / 10 ** 6)  # 24 microsecond
 
   # Assert that the average time is below the threshold
   assert mean_time < max_time, f"Benchmark exceeded threshold: {mean_time:.3f}s > {max_time:.3f}s"
@@ -304,7 +304,7 @@ def test_false_true(benchmark):
   mean_time = benchmark.stats.get("median")
 
   # Define your time threshold in seconds (ms in this case)
-  max_time = 1 / 10 ** 3  # 1 millisecond
+  max_time = 900 / 10 ** 4  # 0.9 millisecond
 
   # Assert that the average time is below the threshold
   assert mean_time < max_time, f"Benchmark exceeded threshold: {mean_time:.3f}s > {max_time:.3f}s"
@@ -315,7 +315,7 @@ def test_false_false(benchmark):
   mean_time = benchmark.stats.get("median")
 
   # Define your time threshold in seconds (100ms in this case)
-  max_time = 210 * (1 / 10 ** 6)  # 210 microsecond
+  max_time = 55 / 10 ** 6  # 56 microsecond
 
   # Assert that the average time is below the threshold
   assert mean_time < max_time, f"Benchmark exceeded threshold: {mean_time:.3f}s > {max_time:.3f}s"