json-repair 0.18.0__tar.gz → 0.19.1__tar.gz

{json_repair-0.18.0/src/json_repair.egg-info → json_repair-0.19.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.18.0
+Version: 0.19.1
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License

{json_repair-0.18.0 → json_repair-0.19.1}/pyproject.toml

@@ -3,7 +3,7 @@ requires = ["setuptools>=61.0"]
 build-backend = "setuptools.build_meta"
 [project]
 name = "json_repair"
-version = "0.18.0"
+version = "0.19.1"
 license = {file = "LICENSE"}
 authors = [
   { name="Stefano Baccianella", email="4247706+mangiucugna@users.noreply.github.com" },

{json_repair-0.18.0 → json_repair-0.19.1}/src/json_repair/json_repair.py

@@ -178,7 +178,12 @@ class JSONParser:
             if not value:
                 break
 
-            arr.append(value)
+            if value == "..." and self.get_char_at(-1) == ".":
+                self.log(
+                    "While parsing an array, found a stray '...'; ignoring it", "info"
+                )
+            else:
+                arr.append(value)
 
             # skip over whitespace after a value but before closing ]
             char = self.get_char_at()
@@ -304,6 +309,15 @@ class JSONParser:
             string_acc += char
             self.index += 1
             char = self.get_char_at()
+            if len(string_acc) > 1 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]
+                if char in [rstring_delimiter, "t", "n", "r", "b", "\\"]:
+                    escape_seqs = {"t": "\t", "n": "\n", "r": "\r", "b": "\b"}
+                    string_acc += escape_seqs.get(char, char)
+                    self.index += 1
+                    char = self.get_char_at()
             # ChatGPT sometimes forget to quote stuff in html tags or markdown, so we do this whole thing here
             if char == rstring_delimiter:
                 # Special case here, in case of double quotes one after another

{json_repair-0.18.0 → json_repair-0.19.1/src/json_repair.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.18.0
+Version: 0.19.1
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License

{json_repair-0.18.0 → json_repair-0.19.1}/tests/test_json_repair.py

@@ -39,6 +39,9 @@ def test_repair_json():
         == '{"name": "John", "age": 30, "city": "New York"}'
     )
     assert repair_json("[1, 2, 3,") == "[1, 2, 3]"
+    assert repair_json("[1, 2, 3, ...]") == "[1, 2, 3]"
+    assert repair_json("[1, 2, ... , 3]") == "[1, 2, 3]"
+    assert repair_json("[1, 2, '...', 3]") == '[1, 2, "...", 3]'
     assert (
         repair_json('{"employees":["John", "Anna",')
         == '{"employees": ["John", "Anna"]}'
@@ -56,7 +59,7 @@ def test_repair_json():
     assert repair_json('{"') == '{}'
     assert repair_json('["') == '[]'
     assert repair_json("'\"'") == '"\\\""'
-    assert repair_json("{\"key\": 'string\"\n\t\le'") == '{"key": "string\\"\\n\\t\\\\le"}'
+    assert repair_json("{\"key\": 'string\"\n\t\le'") == '{"key": "string\\"\\n\\tle"}'
     assert repair_json('{foo: [}') == '{"foo": []}'
     assert repair_json('''{ "a": "{ b": {} }" }''') == '{"a": "{ b"}'
     assert repair_json('{"key": "value:value"}') == '{"key": "value:value"}'
@@ -107,7 +110,7 @@ def test_repair_json():
                        { "key": "value" }
                        ```""") == '{"key": "value"}'
     assert repair_json('````{ "key": "value" }```') == '{"key": "value"}'
-    assert repair_json(r'{"real_content": "Some string: Some other string Some string <a href=\"https://domain.com\">Some link</a>"') == r'{"real_content": "Some string: Some other string Some string <a href=\\\"https://domain.com\\\">Some link</a>"}'
+    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"}'
     assert repair_json('{"key\t_": "value"}') == '{"key\\t_": "value"}'
     assert repair_json('{""answer"":[{""traits"":''Female aged 60+'',""answer1"":""5""}]}') == '{"answer": [{"traits": "Female aged 60+", "answer1": "5"}]}'
@@ -156,7 +159,6 @@ def test_repair_json_with_objects():
         "age": 30,
         "city": "New York",
     }
-    assert repair_json("[1, 2, 3,", return_objects=True) == [1, 2, 3]
     assert repair_json('{"employees":["John", "Anna",', return_objects=True) == {
         "employees": ["John", "Anna"]
     }