json-repair 0.27.0__tar.gz → 0.27.2__tar.gz

{json_repair-0.27.0/src/json_repair.egg-info → json_repair-0.27.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.27.0
+Version: 0.27.2
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License
@@ -38,7 +38,7 @@ License-File: LICENSE
 [![PyPI](https://img.shields.io/pypi/v/json-repair)](https://pypi.org/project/json-repair/)
 ![Python version](https://img.shields.io/badge/python-3.8+-important)
 [![PyPI downloads](https://img.shields.io/pypi/dm/json-repair)](https://pypi.org/project/json-repair/)
-
+[![Github Sponsors](https://img.shields.io/github/sponsors/mangiucugna)](https://github.com/sponsors/mangiucugna)
 
 This simple package can be used to fix an invalid json string. To know all cases in which this package will work, check out the unit test.
 

{json_repair-0.27.0 → json_repair-0.27.2}/README.md

@@ -1,7 +1,7 @@
 [![PyPI](https://img.shields.io/pypi/v/json-repair)](https://pypi.org/project/json-repair/)
 ![Python version](https://img.shields.io/badge/python-3.8+-important)
 [![PyPI downloads](https://img.shields.io/pypi/dm/json-repair)](https://pypi.org/project/json-repair/)
-
+[![Github Sponsors](https://img.shields.io/github/sponsors/mangiucugna)](https://github.com/sponsors/mangiucugna)
 
 This simple package can be used to fix an invalid json string. To know all cases in which this package will work, check out the unit test.
 

{json_repair-0.27.0 → json_repair-0.27.2}/pyproject.toml

@@ -3,7 +3,7 @@ requires = ["setuptools>=61.0"]
 build-backend = "setuptools.build_meta"
 [project]
 name = "json_repair"
-version = "0.27.0"
+version = "0.27.2"
 license = {file = "LICENSE"}
 authors = [
   { name="Stefano Baccianella", email="4247706+mangiucugna@users.noreply.github.com" },

{json_repair-0.27.0 → json_repair-0.27.2}/src/json_repair/json_repair.py

@@ -179,21 +179,12 @@ class JSONParser:
 
             # <member> starts with a <string>
             key = ""
-            while key == "" and self.get_char_at():
-                current_index = self.index
+            while self.get_char_at():
                 key = self.parse_string()
 
-                # This can happen sometimes like { "": "value" }
-                if key == "" and self.get_char_at() == ":":
-                    key = "empty_placeholder"
-                    self.log(
-                        "While parsing an object we found an empty key, replacing with empty_placeholder",
-                        "info",
-                    )
+                if key != "" or (key == "" and self.get_char_at() == ":"):
+                    # If the string is empty but there is a object divider, we are done here
                     break
-                elif key == "" and self.index == current_index:
-                    # Sometimes the string search might not move the index at all, that might lead us to an infinite loop
-                    self.index += 1
 
             self.skip_whitespaces_at()
 
@@ -226,13 +217,6 @@ class JSONParser:
             # Remove trailing spaces
             self.skip_whitespaces_at()
 
-        # Especially at the end of an LLM generated json you might miss the last "}"
-        if (self.get_char_at() or "}") != "}":
-            self.log(
-                "While parsing an object, we couldn't find the closing }, ignoring",
-                "info",
-            )
-
         self.index += 1
         return obj
 
@@ -261,13 +245,6 @@ class JSONParser:
             while char and (char.isspace() or char == ","):
                 self.index += 1
                 char = self.get_char_at()
-            # If this is the right value of an object and we are closing the object, it means the array is over
-            if self.get_context() == "object_value" and char == "}":
-                self.log(
-                    "While parsing an array inside an object, we got to the end without finding a ]. Stopped parsing",
-                    "info",
-                )
-                break
 
         # Especially at the end of an LLM generated json you might miss the last "]"
         char = self.get_char_at()
@@ -275,14 +252,6 @@ class JSONParser:
             self.log(
                 "While parsing an array we missed the closing ], adding it back", "info"
             )
-            # Sometimes when you fix a missing "]" you'll have a trailing "," there that makes the JSON invalid
-            if char == ",":
-                # Remove trailing "," before adding the "]"
-                self.log(
-                    "While parsing an array, found a trailing , before adding ]",
-                    "info",
-                )
-
             self.index -= 1
 
         self.index += 1
@@ -337,6 +306,11 @@ class JSONParser:
 
         # There is sometimes a weird case of doubled quotes, we manage this also later in the while loop
         if self.get_char_at() == lstring_delimiter:
+            # If it's an empty key, this was easy
+            if self.get_context() == "object_key" and self.get_char_at(1) == ":":
+                self.index += 1
+                return ""
+
             # This is a valid exception only if it's closed by a double delimiter again
             i = 1
             next_c = self.get_char_at(i)
@@ -412,7 +386,7 @@ class JSONParser:
             string_acc += char
             self.index += 1
             char = self.get_char_at()
-            if len(string_acc) > 1 and string_acc[-1] == "\\":
+            if len(string_acc) > 0 and string_acc[-1] == "\\":
                 # This is a special case, if people use real strings this might happen
                 self.log("Found a stray escape sequence, normalizing it", "info")
                 string_acc = string_acc[:-1]
@@ -429,6 +403,7 @@ class JSONParser:
                         "While parsing a string, we found a doubled quote, ignoring it",
                         "info",
                     )
+                    self.index += 1
                 elif missing_quotes and self.get_context() == "object_value":
                     # In case of missing starting quote I need to check if the delimeter is the end or the beginning of a key
                     i = 1
@@ -575,22 +550,18 @@ class JSONParser:
             # The number ends with a non valid character for a number/currency, rolling back one
             number_str = number_str[:-1]
             self.index -= 1
-        if number_str:
-            try:
-                if "," in number_str:
-                    return str(number_str)
-                if "." in number_str or "e" in number_str or "E" in number_str:
-                    return float(number_str)
-                elif number_str == "-":
-                    # If there is a stray "-" this will throw an exception, throw away this character
-                    return self.parse_json()
-                else:
-                    return int(number_str)
-            except ValueError:
-                return number_str
-        else:
-            # If nothing works, let's skip and keep parsing
-            return self.parse_json()
+        try:
+            if "," in number_str:
+                return str(number_str)
+            if "." in number_str or "e" in number_str or "E" in number_str:
+                return float(number_str)
+            elif number_str == "-":
+                # If there is a stray "-" this will throw an exception, throw away this character
+                return self.parse_json()
+            else:
+                return int(number_str)
+        except ValueError:
+            return number_str
 
     def parse_boolean_or_null(self) -> Union[bool, str, None]:
         # <boolean> is one of the literal strings 'true', 'false', or 'null' (unquoted)

{json_repair-0.27.0 → json_repair-0.27.2/src/json_repair.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.27.0
+Version: 0.27.2
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License
@@ -38,7 +38,7 @@ License-File: LICENSE
 [![PyPI](https://img.shields.io/pypi/v/json-repair)](https://pypi.org/project/json-repair/)
 ![Python version](https://img.shields.io/badge/python-3.8+-important)
 [![PyPI downloads](https://img.shields.io/pypi/dm/json-repair)](https://pypi.org/project/json-repair/)
-
+[![Github Sponsors](https://img.shields.io/github/sponsors/mangiucugna)](https://github.com/sponsors/mangiucugna)
 
 This simple package can be used to fix an invalid json string. To know all cases in which this package will work, check out the unit test.
 

{json_repair-0.27.0 → json_repair-0.27.2}/tests/test_json_repair.py

@@ -97,6 +97,7 @@ def test_missing_and_mixed_quotes():
     assert repair_json('[{"key": "value", COMMENT "notes": "lorem "ipsum", sic."}]') == '[{"key": "value", "notes": "lorem \\"ipsum\\", sic."}]'
     assert repair_json('{"key": ""value"}') == '{"key": "value"}'
     assert repair_json('{"key": "value", 5: "value"}') == '{"key": "value", "5": "value"}'
+    assert repair_json('{"foo": "\\"bar\\""') == '{"foo": "\\"bar\\""}'
 
 def test_array_edge_cases():
     assert repair_json("[1, 2, 3,") == "[1, 2, 3]"
@@ -124,20 +125,11 @@ def test_escaping():
     assert repair_json('{"key\t_": "value"}') == '{"key\\t_": "value"}'
     
     
-def test_object_edge_cases():    
-    assert {
-        repair_json('{"value_1": "value_2": "data"}') == '{"value_1": "value_2", "data": ""}'
-    }
-    assert {
-        repair_json('{"value_1": true, COMMENT "value_2": "data"}') == '{"value_1": "value_2", "": "data"}'
-    }
+def test_object_edge_cases():
+    assert repair_json('{       ') == '{}'
+    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 }') == '{"value_1": true, "value_2": "data"}'
-    assert {
-        repair_json('{"" : true, "key2": "value2"}') == '{" ": true, "key2": "value_2"}'
-    }
-    assert {
-        repair_json('{"": true, "key2": "value2"}') == '{"empty_placeholder": true, "key2": "value_2"}'
-    }
+    assert repair_json('{"" : true, "key2": "value2"}') == '{"": true, "key2": "value2"}'
     assert repair_json('{""answer"":[{""traits"":''Female aged 60+'',""answer1"":""5""}]}') == '{"answer": [{"traits": "Female aged 60+", "answer1": "5"}]}'
     assert repair_json('{ "words": abcdef", "numbers": 12345", "words2": ghijkl" }') == '{"words": "abcdef", "numbers": 12345, "words2": "ghijkl"}'
     assert repair_json('''{"number": 1,"reason": "According...""ans": "YES"}''') == '{"number": 1, "reason": "According...", "ans": "YES"}'
@@ -158,6 +150,7 @@ def test_number_edge_cases():
     assert repair_json('{"key": 1/3, "foo": "bar"}') == '{"key": "1/3", "foo": "bar"}'
     assert repair_json('{"key": 10-20}') == '{"key": "10-20"}'
     assert repair_json('{"key": 1.1.1}') == '{"key": "1.1.1"}'
+    assert repair_json('[- ') == '[]'
 
 def test_markdown():
     assert repair_json('{ "content": "[LINK]("https://google.com")" }') == '{"content": "[LINK](\\"https://google.com\\")"}'