json-repair 0.15.4__tar.gz → 0.15.6__tar.gz

{json_repair-0.15.4/src/json_repair.egg-info → json_repair-0.15.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.15.4
+Version: 0.15.6
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License

{json_repair-0.15.4 → json_repair-0.15.6}/pyproject.toml

@@ -3,7 +3,7 @@ requires = ["setuptools>=61.0"]
 build-backend = "setuptools.build_meta"
 [project]
 name = "json_repair"
-version = "0.15.4"
+version = "0.15.6"
 license = {file = "LICENSE"}
 authors = [
   { name="Stefano Baccianella", email="4247706+mangiucugna@users.noreply.github.com" },

{json_repair-0.15.4 → json_repair-0.15.6}/src/json_repair/json_repair.py

@@ -174,6 +174,7 @@ class JSONParser:
         arr = []
         # Stop when you either find the closing parentheses or you have iterated over the entire string
         while (self.get_char_at() or "]") != "]":
+            self.skip_whitespaces_at()
             value = self.parse_json()
 
             # It is possible that parse_json() returns nothing valid, so we stop
@@ -218,6 +219,7 @@ class JSONParser:
 
         # Flag to manage corner cases related to missing starting quote
         fixed_quotes = False
+        doubled_quotes = False
         lstring_delimiter = rstring_delimiter = '"'
         if isinstance(string_quotes, list):
             lstring_delimiter = string_quotes[0]
@@ -226,17 +228,20 @@ class JSONParser:
             lstring_delimiter = rstring_delimiter = string_quotes
         # There is sometimes a weird case of doubled quotes, we manage this also later in the while loop
         if self.get_char_at(1) == lstring_delimiter:
-            next_next_char = self.get_char_at(2) or ""
-            if not next_next_char.isspace() and next_next_char not in [
-                ",",
-                ":",
-                "]",
-                "}",
-            ]:
+            # This is a valid exception only if it's closed by a double delimiter again
+            i = 2
+            next_c = self.get_char_at(i)
+            while next_c and next_c != rstring_delimiter:
+                i += 1
+                next_c = self.get_char_at(i)
+            # Now check that the next character is also a delimiter to ensure that we have "".....""
+            # In that case we ignore this rstring delimiter
+            if next_c and (self.get_char_at(i + 1) or "") == rstring_delimiter:
                 self.log(
-                    "While parsing a string, we found a doubled quote, ignoring it",
+                    "While parsing a string, we found a valid starting doubled quote, ignoring it",
                     "info",
                 )
+                doubled_quotes = True
                 self.index += 1
         char = self.get_char_at()
         if char != lstring_delimiter:
@@ -277,13 +282,9 @@ class JSONParser:
                     self.remove_char_at(-1)
                     self.index -= 1
             # ChatGPT sometimes forget to quote stuff in html tags or markdown, so we do this whole thing here
-            if (
-                char == rstring_delimiter
-                # Next character is not a delimiter
-                and self.get_char_at(1) not in [",", ":", "]", "}"]
-            ):
+            if char == rstring_delimiter:
                 # Special case here, in case of double quotes one after another
-                if self.get_char_at(1) == rstring_delimiter:
+                if doubled_quotes and self.get_char_at(1) == rstring_delimiter:
                     self.log(
                         "While parsing a string, we found a doubled quote, ignoring it",
                         "info",
@@ -292,13 +293,20 @@ class JSONParser:
                     self.remove_char_at()
                 else:
                     # Check if eventually there is a rstring delimiter, otherwise we bail
-                    i = 2
+                    i = 1
+                    context = self.get_context()
                     next_c = self.get_char_at(i)
                     while next_c and next_c != rstring_delimiter:
+                        # If we are in an object context, let's check for the right delimiters
+                        if (
+                            (context == "object_key" and next_c == ":")
+                            or (context == "object_value" and next_c in ["}", ","])
+                            or (context == "" and next_c in ["]", ","])
+                        ):
+                            break
                         i += 1
                         next_c = self.get_char_at(i)
-                    # In that case we ignore this rstring delimiter
-                    if next_c:
+                    if next_c == rstring_delimiter:
                         self.log(
                             "While parsing a string, we a misplaced quote that would have closed the string but has a different meaning here, ignoring it",
                             "info",
@@ -414,7 +422,7 @@ class JSONParser:
 
     def get_context(self) -> str:
         try:
-            return self.context[0]
+            return self.context[-1]
         except Exception:
             return ""
 

{json_repair-0.15.4 → json_repair-0.15.6/src/json_repair.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.15.4
+Version: 0.15.6
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License

{json_repair-0.15.4 → json_repair-0.15.6}/tests/test_json_repair.py

@@ -107,7 +107,6 @@ def test_repair_json():
     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"}]}'
-    assert repair_json('{""answer":[{""traits":""Female aged 60+",""answer1":""5"}]}') == '{"answer": [{"traits": "Female aged 60+", "answer1": "5"}]}'
     assert repair_json('{"key":""') == '{"key": ""}'
     assert repair_json('{ "words": abcdef", "numbers": 12345", "words2": ghijkl" }') == '{"words": "abcdef", "numbers": 12345, "words2": "ghijkl"}'
     assert repair_json('{"key": 1/3}') == '{"key": "1/3"}'
@@ -191,6 +190,18 @@ def test_repair_json_with_objects():
 }
 ''', True) == {"resourceType": "Bundle", "id": "1", "type": "collection", "entry": [{"resource": {"resourceType": "Patient", "id": "1", "name": [{"use": "official", "family": "Corwin", "given": ["Keisha", "Sunny"], "prefix": ["Mrs."]}, {"use": "maiden", "family": "Goodwin", "given": ["Keisha", "Sunny"], "prefix": ["Mrs."]}]}}]}
     assert repair_json('{\n"html": "<h3 id="aaa">Waarom meer dan 200 Technical Experts - "Passie voor techniek"?</h3>"}', True) == {'html': '<h3 id="aaa">Waarom meer dan 200 Technical Experts - "Passie voor techniek"?</h3>'}
+    assert repair_json("""
+        [
+            {
+                "foo": "Foo bar baz",
+                "tag": "#foo-bar-baz"
+            },
+            {
+                "foo": "foo bar "foobar" foo bar baz.",
+                "tag": "#foo-bar-foobar"
+            }
+        ]
+        """, True) == [{"foo": "Foo bar baz", "tag": "#foo-bar-baz"},{"foo": "foo bar \"foobar\" foo bar baz.", "tag": "#foo-bar-foobar" }]
 
 
 def test_repair_json_corner_cases_generate_by_gpt():