json-repair 0.21.0__tar.gz → 0.23.0__tar.gz

{json_repair-0.21.0/src/json_repair.egg-info → json_repair-0.23.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.21.0
+Version: 0.23.0
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License

{json_repair-0.21.0 → json_repair-0.23.0}/pyproject.toml

@@ -3,7 +3,7 @@ requires = ["setuptools>=61.0"]
 build-backend = "setuptools.build_meta"
 [project]
 name = "json_repair"
-version = "0.21.0"
+version = "0.23.0"
 license = {file = "LICENSE"}
 authors = [
   { name="Stefano Baccianella", email="4247706+mangiucugna@users.noreply.github.com" },

{json_repair-0.21.0 → json_repair-0.23.0}/src/json_repair/json_repair.py

@@ -1,7 +1,7 @@
 """
 This module will parse the JSON file following the BNF definition:
 
-    <json> ::= <primitive> | <container>
+    <json> ::= <container>
 
     <primitive> ::= <number> | <string> | <boolean>
     ; Where:
@@ -89,15 +89,32 @@ class JSONParser:
     def parse(
         self,
     ) -> Union[JSONReturnType, Tuple[JSONReturnType, List[Dict[str, str]]]]:
+        json = self.parse_json()
+        if self.index < len(self.json_str):
+            json = [json]
+            last_index = self.index
+            while self.index < len(self.json_str):
+                j = self.parse_json()
+                if j != "":
+                    json.append(j)
+                if self.index == last_index:
+                    self.index += 1
+                last_index = self.index
+            if len(json) == 1:
+                json = json[0]
+            elif len(json) == 0:
+                json = ""
         if self.logger.log_level == "none":
-            return self.parse_json()
+            return json
         else:
-            return self.parse_json(), self.logger.log
+            return json, self.logger.log
 
     def parse_json(
         self,
     ) -> JSONReturnType:
         char = self.get_char_at()
+        # This parser will ignore any basic element (string or number) that is not inside an array or object
+        is_in_context = len(self.context) > 0
         # False means that we are at the end of the string provided, is the base case for recursion
         if char is False:
             return ""
@@ -120,10 +137,10 @@ class JSONParser:
             )
             return ""
         # <string> starts with a quote
-        elif char in ['"', "'", "“"] or char.isalpha():
+        elif is_in_context and (char in ['"', "'", "“"] or char.isalpha()):
             return self.parse_string()
         # <number> starts with [0-9] or minus
-        elif char.isdigit() or char == "-" or char == ".":
+        elif is_in_context and (char.isdigit() or char == "-" or char == "."):
             return self.parse_number()
         # If everything else fails, we just ignore and move on
         else:
@@ -304,14 +321,6 @@ class JSONParser:
                 "While parsing a string, we found a literal instead of a quote",
                 "info",
             )
-            if self.get_context() == "":
-                # A string literal in the wild isn't a valid json and not something we can fix
-                self.log(
-                    "While parsing a string, we found a literal outside of context, ignoring it",
-                    "info",
-                )
-                self.index += 1
-                return self.parse_json()
             self.log(
                 "While parsing a string, we found no starting quote. Will add the quote back",
                 "info",
@@ -656,3 +665,6 @@ def from_file(
     fd.close()
 
     return jsonobj
+
+
+repair_json("[]{}")

{json_repair-0.21.0 → json_repair-0.23.0/src/json_repair.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.21.0
+Version: 0.23.0
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License

{json_repair-0.21.0 → json_repair-0.23.0}/tests/test_json_repair.py

@@ -1,9 +1,9 @@
 from src.json_repair.json_repair import from_file, repair_json, loads
 
 def test_basic_types_valid():
-    assert repair_json("True", return_objects=True) == True
-    assert repair_json("False", return_objects=True) == False
-    assert repair_json("Null", return_objects=True) == None
+    assert repair_json("True", return_objects=True) == ""
+    assert repair_json("False", return_objects=True) == ""
+    assert repair_json("Null", return_objects=True) == ""
     assert repair_json("1", return_objects=True) == 1
     assert repair_json("[]", return_objects=True) == []
     assert repair_json("[1, 2, 3, 4]", return_objects=True) == [1, 2, 3, 4]
@@ -114,7 +114,7 @@ def test_array_edge_cases():
 
     
 def test_escaping():
-    assert repair_json("'\"'") == '"\\\""'
+    assert repair_json("'\"'") == '""'
     assert repair_json("{\"key\": 'string\"\n\t\le'") == '{"key": "string\\"\\n\\tle"}'
     assert repair_json(r'{"real_content": "Some string: Some other string \t Some string <a href=\"https://domain.com\">Some link</a>"') == r'{"real_content": "Some string: Some other string \t Some string <a href=\"https://domain.com\">Some link</a>"}'
     assert repair_json('{"key_1\n": "value"}') == '{"key_1": "value"}'
@@ -164,9 +164,15 @@ def test_leading_trailing_characters():
     assert repair_json("""{    "a": "",    "b": [ { "c": 1} ] \n}```""") == '{"a": "", "b": [{"c": 1}]}'
     assert repair_json("Based on the information extracted, here is the filled JSON output: ```json { 'a': 'b' } ```") == '{"a": "b"}'
     assert repair_json("""
+                       The next 64 elements are:
                        ```json
                        { "key": "value" }
                        ```""") == '{"key": "value"}'
+def test_multiple_jsons():
+    assert repair_json("[]{}") == "[[], {}]"
+    assert repair_json("{}[]") == "[{}, []]"
+    assert repair_json('{"key":"value"}[1,2,3,True]') == '[{"key": "value"}, ["1,2,3", true]]'
+    assert repair_json('lorem ```json {"key":"value"} ``` ipsum ```json [1,2,3,True] ``` 42') == '[{"key": "value"}, ["1,2,3", true]]'
 
 def test_repair_json_with_objects():
     # Test with valid JSON strings

{json_repair-0.21.0 → json_repair-0.23.0}/tests/test_performance.py

@@ -19,7 +19,7 @@ def test_true_true_correct(benchmark):
   mean_time = benchmark.stats.get("median")
 
   # Define your time threshold in seconds
-  max_time = 13 / 10 ** 4  # 1.3 millisecond
+  max_time = 14 / 10 ** 4  # 1.4 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"
@@ -31,7 +31,7 @@ def test_true_true_incorrect(benchmark):
   mean_time = benchmark.stats.get("median")
 
   # Define your time threshold in seconds
-  max_time = 13 / 10 ** 4  # 1.3 millisecond
+  max_time = 14 / 10 ** 4  # 1.4 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"
@@ -53,7 +53,7 @@ def test_true_false_incorrect(benchmark):
   mean_time = benchmark.stats.get("median")
 
   # Define your time threshold in seconds
-  max_time = 13 / 10 ** 4  # 1.3 millisecond
+  max_time = 14 / 10 ** 4  # 1.4 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"