json-repair 0.19.1__tar.gz → 0.20.0__tar.gz

{json_repair-0.19.1/src/json_repair.egg-info → json_repair-0.20.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.19.1
+Version: 0.20.0
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License

{json_repair-0.19.1 → json_repair-0.20.0}/pyproject.toml

@@ -3,7 +3,7 @@ requires = ["setuptools>=61.0"]
 build-backend = "setuptools.build_meta"
 [project]
 name = "json_repair"
-version = "0.19.1"
+version = "0.20.0"
 license = {file = "LICENSE"}
 authors = [
   { name="Stefano Baccianella", email="4247706+mangiucugna@users.noreply.github.com" },

{json_repair-0.19.1 → json_repair-0.20.0}/src/json_repair/json_repair.py

@@ -22,6 +22,7 @@ If something is wrong (a missing parantheses or quotes for example) it will use
 All supported use cases are in the unit tests
 """
 
+import os
 import json
 from typing import Any, Dict, List, Union, TextIO
 
@@ -31,7 +32,9 @@ class JSONParser:
         # The string to parse
         self.json_str = json_str
         # Alternatively, the file description with a json file in it
-        self.json_fd = json_fd
+        if json_fd:
+            # This is a trick we do to treat the file wrapper as an array
+            self.json_str = StringFileWrapper(json_fd)
         # Index is our iterator that will keep track of which character we are looking at right now
         self.index = 0
         # This is used in the object member parsing to manage the special cases of missing quotes in key or value
@@ -246,7 +249,8 @@ class JSONParser:
             rstring_delimiter = "”"
         elif char.isalpha():
             # This could be a <boolean> and not a string. Because (T)rue or (F)alse or (N)ull are valid
-            if char.lower() in ["t", "f", "n"]:
+            # But remember, object keys are only of type string
+            if char.lower() in ["t", "f", "n"] and self.get_context() != "object_key":
                 value = self.parse_boolean_or_null()
                 if value != "":
                     return value
@@ -263,7 +267,8 @@ class JSONParser:
                 self.index += 1
                 return self.parse_json()
             self.log(
-                "While parsing a string, we found no starting quote, ignoring", "info"
+                "While parsing a string, we found no starting quote. Will add the quote back",
+                "info",
             )
             missing_quotes = True
 
@@ -330,32 +335,50 @@ class JSONParser:
                     # Check if eventually there is a rstring delimiter, otherwise we bail
                     i = 1
                     next_c = self.get_char_at(i)
-                    while next_c and next_c != rstring_delimiter:
+                    check_comma_in_object_value = True
+                    while next_c and next_c not in [
+                        rstring_delimiter,
+                        lstring_delimiter,
+                    ]:
+                        # This is a bit of a weird workaround, essentially in object_value context we don't always break on commas
+                        # This is because the routine after will make sure to correct any bad guess and this solves a corner case
+                        if next_c.isalpha():
+                            check_comma_in_object_value = False
                         # If we are in an object context, let's check for the right delimiters
                         if (
-                            next_c == lstring_delimiter
-                            or ("object_key" in self.context and next_c == ":")
-                            or ("object_value" in self.context and next_c in ["}", ","])
+                            ("object_key" in self.context and next_c in [":", "}"])
+                            or ("object_value" in self.context and next_c == "}")
                             or ("array" in self.context and next_c in ["]", ","])
+                            or (
+                                check_comma_in_object_value
+                                and self.get_context() == "object_value"
+                                and next_c == ","
+                            )
                         ):
                             break
                         i += 1
                         next_c = self.get_char_at(i)
                     if next_c == rstring_delimiter:
-                        # But this might not be it! This could be just a missing comma
-                        # We need to check if we find a rstring_delimiter and a colon after
-                        i += 1
-                        next_c = self.get_char_at(i)
-                        while next_c and next_c != rstring_delimiter:
+                        if self.get_context() == "object_value":
+                            # But this might not be it! This could be just a missing comma
+                            # We found a delimiter and we need to check if this is a key
+                            # so find a rstring_delimiter and a colon after
                             i += 1
                             next_c = self.get_char_at(i)
-                        i += 1
-                        next_c = self.get_char_at(i)
-                        while next_c and next_c != ":":
-                            if next_c in [lstring_delimiter, rstring_delimiter, ","]:
-                                break
+                            while next_c and next_c != rstring_delimiter:
+                                i += 1
+                                next_c = self.get_char_at(i)
                             i += 1
                             next_c = self.get_char_at(i)
+                            while next_c and next_c != ":":
+                                if next_c in [
+                                    lstring_delimiter,
+                                    rstring_delimiter,
+                                    ",",
+                                ]:
+                                    break
+                                i += 1
+                                next_c = self.get_char_at(i)
                         # Only if we fail to find a ':' then we know this is misplaced quote
                         if next_c != ":":
                             self.log(
@@ -451,36 +474,22 @@ class JSONParser:
         try:
             return self.json_str[self.index + count]
         except IndexError:
-            if self.json_fd:
-                self.json_fd.seek(self.index + count)
-                char = self.json_fd.read(1)
-                if char == "":
-                    return False
-                return char
-            else:
-                return False
+            return False
 
     def skip_whitespaces_at(self) -> None:
         """
         This function quickly iterates on whitespaces, syntactic sugar to make the code more concise
         """
-        if self.json_fd:
-            char = self.get_char_at()
-            while char and char.isspace():
-                self.index += 1
-                char = self.get_char_at()
-        else:
-            # If this is not a file stream, we do this monster here to make this function much much faster
+        try:
+            char = self.json_str[self.index]
+        except IndexError:
+            return
+        while char.isspace():
+            self.index += 1
             try:
                 char = self.json_str[self.index]
             except IndexError:
                 return
-            while char.isspace():
-                self.index += 1
-                try:
-                    char = self.json_str[self.index]
-                except IndexError:
-                    return
 
     def set_context(self, value: str) -> None:
         # If a value is provided update the context variable and save in stack
@@ -502,22 +511,9 @@ class JSONParser:
     def log(self, text: str, level: str) -> None:
         if level == self.logger["log_level"]:
             context = ""
-            if self.json_fd:
-                self.json_fd.seek(self.index - self.logger["window"])
-                context = self.json_fd.read(self.logger["window"] * 2)
-                self.json_fd.seek(self.index)
-            else:
-                start = (
-                    self.index - self.logger["window"]
-                    if (self.index - self.logger["window"]) >= 0
-                    else 0
-                )
-                end = (
-                    self.index + self.logger["window"]
-                    if (self.index + self.logger["window"]) <= len(self.json_str)
-                    else len(self.json_str)
-                )
-                context = self.json_str[start:end]
+            start = max(self.index - self.logger["window"], 0)
+            end = min(self.index + self.logger["window"], len(self.json_str))
+            context = self.json_str[start:end]
             self.logger["log"].append(
                 {
                     "text": text,
@@ -593,3 +589,31 @@ def from_file(
     fd.close()
 
     return jsonobj
+
+
+class StringFileWrapper:
+    # This is a trick to simplify the code above, transform the filedescriptor handling into an array handling
+    def __init__(self, fd: TextIO) -> None:
+        self.fd = fd
+        self.length = None
+
+    def __getitem__(self, index: int) -> Any:
+        if isinstance(index, slice):
+            self.fd.seek(index.start)
+            value = self.fd.read(index.stop - index.start)
+            self.fd.seek(index.start)
+            return value
+        else:
+            self.fd.seek(index)
+            return self.fd.read(1)
+
+    def __len__(self) -> int:
+        if not self.length:
+            current_position = self.fd.tell()
+            self.fd.seek(0, os.SEEK_END)
+            self.length = self.fd.tell()
+            self.fd.seek(current_position)
+        return self.length
+
+    def __setitem__(self):
+        raise Exception("This is read-only!")

{json_repair-0.19.1 → json_repair-0.20.0/src/json_repair.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.19.1
+Version: 0.20.0
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License

{json_repair-0.19.1 → json_repair-0.20.0}/tests/test_json_repair.py

@@ -1,27 +1,61 @@
 from src.json_repair.json_repair import from_file, repair_json, loads
 
 
-def test_repair_json():
-    # Test with valid JSON strings
+def test_valid_json():
     assert repair_json("[]") == "[]"
-    assert 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("{'key': 'string', 'key2': false, \"key3\": null, \"key4\": unquoted}") == '{"key": "string", "key2": false, "key3": null, "key4": "unquoted"}'
+    assert repair_json("[1, 2, 3, 4]") == "[1, 2, 3, 4]"
     assert (
         repair_json('{"name": "John", "age": 30, "city": "New York"}')
         == '{"name": "John", "age": 30, "city": "New York"}'
     )
-    assert repair_json("[1, 2, 3, 4]") == "[1, 2, 3, 4]"
     assert (
         repair_json('{"employees":["John", "Anna", "Peter"]} ')
         == '{"employees": ["John", "Anna", "Peter"]}'
     )
+    assert repair_json('{"key": "value:value"}') == '{"key": "value:value"}'
+    assert (
+        repair_json('{"text": "The quick brown fox,"}')
+        == '{"text": "The quick brown fox,"}'
+    )
+    assert (
+        repair_json('{"text": "The quick brown fox won\'t jump"}')
+        == '{"text": "The quick brown fox won\'t jump"}'
+    )
+    assert repair_json('{"key": ""') == '{"key": ""}'
+    assert (
+        repair_json('{"key1": {"key2": [1, 2, 3]}}') == '{"key1": {"key2": [1, 2, 3]}}'
+    )
+    assert (
+        repair_json('{"key": 12345678901234567890}') == '{"key": 12345678901234567890}'
+    )
+    assert repair_json('{"key": "value\u263A"}') == '{"key": "value\\u263a"}'
+    assert repair_json('{"key": "value\\nvalue"}') == '{"key": "value\\nvalue"}'
+
+def test_brackets_edge_cases():
+    assert repair_json("[{]") == "[]"
+    assert repair_json("   {  }   ") == "{}"
+    assert repair_json("[") == "[]"
+    assert repair_json("]") == '""'
+    assert repair_json("{") == "{}"
+    assert repair_json("}") == '""'
+    assert repair_json('{"') == '{}'
+    assert repair_json('["') == '[]'
+    assert repair_json('{foo: [}') == '{"foo": []}'
 
-    # Test with invalid JSON strings
+def test_general_edge_cases():
+    assert repair_json("\"") == '""'
+    assert repair_json("\n") == '""'
+    assert repair_json(" ") == '""'
+    assert repair_json("[[1\n\n]") == "[[1]]"
+    assert repair_json("string") == '""'
+    assert repair_json("stringbeforeobject {}") == '{}'
+
+def test_mixed_data_types(): 
+    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}'
+
+def test_missing_and_mixed_quotes():
+    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')
         == '{"name": "John", "age": 30, "city": "New York"}'
@@ -38,6 +72,13 @@ def test_repair_json():
         repair_json('{"name": John, "age": 30, "city": "New York"}')
         == '{"name": "John", "age": 30, "city": "New York"}'
     )
+    assert repair_json('{“slanted_delimiter”: "value"}') == '{"slanted_delimiter": "value"}'
+    assert (
+        repair_json('{"name": "John", "age": 30, "city": "New')
+        == '{"name": "John", "age": 30, "city": "New"}'
+    )
+
+def test_array_edge_cases():
     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]"
@@ -46,51 +87,28 @@ def test_repair_json():
         repair_json('{"employees":["John", "Anna",')
         == '{"employees": ["John", "Anna"]}'
     )
-
-    # Test with edge cases
-    assert repair_json(" ") == '""'
-    assert repair_json("[") == "[]"
-    assert repair_json("]") == '""'
-    assert repair_json("[[1\n\n]") == "[[1]]"
-    assert repair_json("{") == "{}"
-    assert repair_json("}") == '""'
-    assert repair_json("string") == '""'
-    assert repair_json("stringbeforeobject {}") == '{}'
-    assert repair_json('{"') == '{}'
-    assert repair_json('["') == '[]'
-    assert repair_json("'\"'") == '"\\\""'
-    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"}'
-    assert repair_json('{“slanted_delimiter”: "value"}') == '{"slanted_delimiter": "value"}'
-    assert (
-        repair_json('{"name": "John", "age": 30, "city": "New')
-        == '{"name": "John", "age": 30, "city": "New"}'
-    )
     assert (
         repair_json('{"employees":["John", "Anna", "Peter')
         == '{"employees": ["John", "Anna", "Peter"]}'
     )
-    assert (
-        repair_json('{"employees":["John", "Anna", "Peter"]}')
-        == '{"employees": ["John", "Anna", "Peter"]}'
-    )
-    assert (
-        repair_json('{"text": "The quick brown fox,"}')
-        == '{"text": "The quick brown fox,"}'
-    )
-    assert (
-        repair_json('{"text": "The quick brown fox won\'t jump"}')
-        == '{"text": "The quick brown fox won\'t jump"}'
-    )
+    assert repair_json('{"key1": {"key2": [1, 2, 3') == '{"key1": {"key2": [1, 2, 3]}}'
+
+    
+def test_escaping():
+    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"}'
+    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"}'
     }
-    # Test with garbage comments
     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"}'
@@ -98,35 +116,38 @@ def test_repair_json():
     assert {
         repair_json('{"": true, "key2": "value2"}') == '{"empty_placeholder": true, "key2": "value_2"}'
     }
-    # Test a nasty corner case
-    assert repair_json(' - { "test_key": ["test_value", "test_value2"] }') == '{"test_key": ["test_value", "test_value2"]}'
-
-    #Test markdown stupidities from ChatGPT
-    assert repair_json('{ "content": "[LINK]("https://google.com")" }') == '{"content": "[LINK](\\"https://google.com\\")"}'
-    assert repair_json('{ "content": "[LINK](" }') == '{"content": "[LINK]("}'
-    assert repair_json('{ "content": "[LINK](", "key": true }') == '{"content": "[LINK](", "key": true}'
-    assert repair_json("""
-                       ```json
-                       { "key": "value" }
-                       ```""") == '{"key": "value"}'
-    assert repair_json('````{ "key": "value" }```') == '{"key": "value"}'
-    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"}]}'
-    assert repair_json('{"key":""') == '{"key": ""}'
     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"}'
+    assert repair_json('''{ "a": "{ b": {} }" }''') == '{"a": "{ b"}'
+    assert repair_json("""{"b": "xxxxx" true}""") == '{"b": "xxxxx"}'
+    assert repair_json('{"key": "Lorem "ipsum" s,"}') == '{"key": "Lorem \\"ipsum\\" s,"}'
+
+def test_number_edge_cases():
+    assert repair_json(' - { "test_key": ["test_value", "test_value2"] }') == '{"test_key": ["test_value", "test_value2"]}'
     assert repair_json('{"key": 1/3}') == '{"key": "1/3"}'
     assert repair_json('{"key": .25}') == '{"key": 0.25}'
-    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('''{"number": 1,"reason": "According...""ans": "YES"}''') == '{"number": 1, "reason": "According...", "ans": "YES"}'
-    assert repair_json('{"key": 1/3, "foo": "bar"}') == '{"key": "1/3", "foo": "bar"}'
     assert repair_json('{"here": "now", "key": 1/3, "foo": "bar"}') == '{"here": "now", "key": "1/3", "foo": "bar"}'
     assert repair_json('{"key": 12345/67890}') == '{"key": "12345/67890"}'
     assert repair_json('[105,12') == '["105,12"]'
     assert repair_json('{"key", 105,12,') == '{"key": "105,12"}'
+    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"}'
+
+def test_markdown():
+    assert repair_json('{ "content": "[LINK]("https://google.com")" }') == '{"content": "[LINK](\\"https://google.com\\")"}'
+    assert repair_json('{ "content": "[LINK](" }') == '{"content": "[LINK]("}'
+    assert repair_json('{ "content": "[LINK](", "key": true }') == '{"content": "[LINK](", "key": true}'
 
+def test_leading_trailing_characters():
+    assert repair_json('````{ "key": "value" }```') == '{"key": "value"}'
+    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("""
+                       ```json
+                       { "key": "value" }
+                       ```""") == '{"key": "value"}'
 
 def test_repair_json_with_objects():
     # Test with valid JSON strings
@@ -142,47 +163,6 @@ def test_repair_json_with_objects():
     assert repair_json('{"employees":["John", "Anna", "Peter"]} ', return_objects=True) == {
         "employees": ["John", "Anna", "Peter"]
     }
-
-    # Test with invalid JSON strings
-    assert repair_json('{"name": "John", "age": 30, "city": "New York', return_objects=True) == {
-        "name": "John",
-        "age": 30,
-        "city": "New York",
-    }
-    assert repair_json('{"name": "John", "age": 30, city: "New York"}', return_objects=True) == {
-        "name": "John",
-        "age": 30,
-        "city": "New York",
-    }
-    assert repair_json('{"name": "John", "age": 30, "city": New York}', return_objects=True) == {
-        "name": "John",
-        "age": 30,
-        "city": "New York",
-    }
-    assert repair_json('{"employees":["John", "Anna",', return_objects=True) == {
-        "employees": ["John", "Anna"]
-    }
-
-    # Test with edge cases
-    assert repair_json(" ", return_objects=True) == ""
-    assert repair_json("[", return_objects=True) == []
-    assert repair_json("{", return_objects=True) == {}
-    assert repair_json('{"key": "value:value"}', return_objects=True) == {"key": "value:value"}
-    assert repair_json("{'key': 'string', 'key2': false, \"key3\": null, \"key4\": unquoted}", return_objects=True) == {"key": "string", "key2": False, "key3": None, "key4": "unquoted"}
-    assert repair_json('{"name": "John", "age": 30, "city": "New', return_objects=True) == {
-        "name": "John",
-        "age": 30,
-        "city": "New",
-    }
-    assert repair_json('{"employees":["John", "Anna", "Peter', return_objects=True) == {
-        "employees": ["John", "Anna", "Peter"]
-    }
-    
-    #Test with garbage comments
-    assert repair_json('{"value_1": true, SHOULD_NOT_EXIST "value_2": "data" AAAA }', return_objects=True) == {'value_1': True, 'value_2': 'data'}
-
-    #Test markdown stupidities from ChatGPT
-    assert repair_json('{ "content": "[LINK]("https://google.com")" }', return_objects=True) == { "content": "[LINK](\"https://google.com\")"}
     assert repair_json('''
 {
   "resourceType": "Bundle",
@@ -216,73 +196,6 @@ def test_repair_json_with_objects():
         ]
         """, return_objects=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():
-    # Test with nested JSON
-    assert (
-        repair_json('{"key1": {"key2": [1, 2, 3]}}') == '{"key1": {"key2": [1, 2, 3]}}'
-    )
-    assert repair_json('{"key1": {"key2": [1, 2, 3') == '{"key1": {"key2": [1, 2, 3]}}'
-
-    # Test with empty keys
-    assert repair_json('{"": "value"}') == '{"": "value"}'
-
-    # Test with Unicode characters
-    assert repair_json('{"key": "value\u263A"}') == '{"key": "value\\u263a"}'
-
-    # Test with special characters
-    assert repair_json('{"key": "value\\nvalue"}') == '{"key": "value\\nvalue"}'
-
-    # Test with large numbers
-    assert (
-        repair_json('{"key": 12345678901234567890}') == '{"key": 12345678901234567890}'
-    )
-
-    # Test with whitespace
-    assert repair_json(' { "key" : "value" } ') == '{"key": "value"}'
-
-    # Test with null values
-    assert repair_json('{"key": null}') == '{"key": null}'
-
-    # Test with numeric-like values
-    assert repair_json('{"key": 10-20}') == '{"key": "10-20"}'
-    assert repair_json('{"key": 1.1.1}') == '{"key": "1.1.1"}'
-
-
-def test_repair_json_corner_cases_generate_by_gpt_with_objects():
-    # Test with nested JSON
-    assert repair_json('{"key1": {"key2": [1, 2, 3]}}', return_objects=True) == {
-        "key1": {"key2": [1, 2, 3]}
-    }
-    assert repair_json('{"key1": {"key2": [1, 2, 3', return_objects=True) == {
-        "key1": {"key2": [1, 2, 3]}
-    }
-
-    # Test with empty keys
-    assert repair_json('{"": "value"}', return_objects=True) == {"": "value"}
-
-    # Test with Unicode characters
-    assert repair_json('{"key": "value\u263A"}', return_objects=True) == {"key": "value☺"}
-
-    # Test with special characters
-    assert repair_json('{"key": "value\\nvalue"}', return_objects=True) == {"key": "value\nvalue"}
-
-    # Test with large numbers
-    assert repair_json('{"key": 12345678901234567890}', return_objects=True) == {
-        "key": 12345678901234567890
-    }
-
-    # Test with whitespace
-    assert repair_json(' { "key" : "value" } ', return_objects=True) == {"key": "value"}
-
-    # Test with null values
-    assert repair_json('{"key": null}', return_objects=True) == {"key": None}
-
-    # Test with numeric-like values
-    assert repair_json('{"key": 10-20}', return_objects=True) == {"key": "10-20"}
-    assert repair_json('{"key": 1.1.1}', return_objects=True) == {"key": "1.1.1"}
-
-
 def test_repair_json_skip_json_loads():
     assert repair_json('{"key": true, "key2": false, "key3": null}', skip_json_loads=True) == '{"key": true, "key2": false, "key3": null}'
     assert repair_json('{"key": true, "key2": false, "key3": null}', return_objects=True, skip_json_loads=True) == {"key": True, "key2": False, "key3": None}
@@ -291,9 +204,20 @@ def test_repair_json_skip_json_loads():
 
 
 def test_repair_json_from_file():
-
     import os.path
     import pathlib
     path = pathlib.Path(__file__).parent.resolve()
 
     assert(from_file(os.path.join(path,"invalid.json"))) == '[{"_id": "655b66256574f09bdae8abe8", "index": 0, "guid": "31082ae3-b0f3-4406-90f4-cc450bd4379d", "isActive": false, "balance": "$2,562.78", "picture": "http://placehold.it/32x32", "age": 32, "eyeColor": "brown", "name": "Glover Rivas", "gender": "male", "company": "EMPIRICA", "email": "gloverrivas@empirica.com", "phone": "+1 (842) 507-3063", "address": "536 Montague Terrace, Jenkinsville, Kentucky, 2235", "about": "Mollit consectetur excepteur voluptate tempor dolore ullamco enim irure ullamco non enim officia. Voluptate occaecat proident laboris ea Lorem cupidatat reprehenderit nisi nisi aliqua. Amet nulla ipsum deserunt excepteur amet ad aute aute ex. Et enim minim sit veniam est quis dolor nisi sunt quis eiusmod in. Amet eiusmod cillum sunt occaecat dolor laboris voluptate in eiusmod irure aliqua duis.", "registered": "2023-11-18T09:32:36 -01:00", "latitude": 36.26102, "longitude": -91.304608, "tags": ["non", "tempor", "do", "ullamco", "dolore", "sunt", "ipsum"], "friends": [{"id": 0, "name": "Cara Shepherd"}, {"id": 1, "name": "Mason Farley"}, {"id": 2, "name": "Harriet Cochran"}], "greeting": "Hello, Glover Rivas! You have 7 unread messages.", "favoriteFruit": "strawberry"}, {"_id": "655b662585364bc57278bb6f", "index": 1, "guid": "0dea7a3a-f812-4dde-b78d-7a9b58e5da05", "isActive": true, "balance": "$1,359.48", "picture": "http://placehold.it/32x32", "age": 38, "eyeColor": "brown", "name": "Brandi Moreno", "gender": "female", "company": "MARQET", "email": "brandimoreno@marqet.com", "phone": "+1 (850) 434-2077", "address": "537 Doone Court, Waiohinu, Michigan, 3215", "about": "Irure proident adipisicing do Lorem do incididunt in laborum in eiusmod eiusmod ad elit proident. Eiusmod dolor ex magna magna occaecat. Nulla deserunt velit ex exercitation et irure sunt. Cupidatat ut excepteur ea quis labore sint cupidatat incididunt amet eu consectetur cillum ipsum proident. Occaecat exercitation aute laborum dolor proident reprehenderit laborum in voluptate culpa. Exercitation nulla adipisicing culpa aute est deserunt ea nisi deserunt consequat occaecat ut et non. Incididunt ex exercitation dolor dolor anim cillum dolore.", "registered": "2015-09-03T11:47:15 -02:00", "latitude": -19.768953, "longitude": 8.948458, "tags": ["laboris", "occaecat", "laborum", "laborum", "ex", "cillum", "occaecat"], "friends": [{"id": 0, "name": "Erna Kelly"}, {"id": 1, "name": "Black Mays"}, {"id": 2, "name": "Davis Buck"}], "greeting": "Hello, Brandi Moreno! You have 1 unread messages.", "favoriteFruit": "apple"}, {"_id": "655b6625870da431bcf5e0c2", "index": 2, "guid": "b17f6e3f-c898-4334-abbf-05cf222f143b", "isActive": false, "balance": "$1,493.77", "picture": "http://placehold.it/32x32", "age": 20, "eyeColor": "brown", "name": "Moody Meadows", "gender": "male", "company": "OPTIQUE", "email": "moodymeadows@optique.com", "phone": "+1 (993) 566-3041", "address": "766 Osborn Street, Bath, Maine, 7666", "about": "Non commodo excepteur nostrud qui adipisicing aliquip dolor minim nulla culpa proident. In ad cupidatat ea mollit ex est do deserunt proident nostrud. Cillum id id eiusmod amet exercitation nostrud cillum sunt deserunt dolore deserunt eiusmod mollit. Ut ex tempor ad laboris voluptate labore id officia fugiat exercitation amet.", "registered": "2015-01-16T02:48:28 -01:00", "latitude": -25.847327, "longitude": 63.95991, "tags": ["aute", "commodo", "adipisicing", "nostrud", "duis", "mollit", "ut"], "friends": [{"id": 0, "name": "Lacey Cash"}, {"id": 1, "name": "Gabrielle Harmon"}, {"id": 2, "name": "Ellis Lambert"}], "greeting": "Hello, Moody Meadows! You have 4 unread messages.", "favoriteFruit": "strawberry"}, {"_id": "655b6625f3e1bf422220854e", "index": 3, "guid": "92229883-2bfd-4974-a08c-1b506b372e46", "isActive": false, "balance": "$2,215.34", "picture": "http://placehold.it/32x32", "age": 22, "eyeColor": "brown", "name": "Heath Nguyen", "gender": "male", "company": "BLEENDOT", "email": "heathnguyen@bleendot.com", "phone": "+1 (989) 512-2797", "address": "135 Milton Street, Graniteville, Nebraska, 276", "about": "Consequat aliquip irure Lorem cupidatat nulla magna ullamco nulla voluptate adipisicing anim consectetur tempor aliquip. Magna aliqua nulla eu tempor esse proident. Proident fugiat ad ex Lorem reprehenderit dolor aliquip labore labore aliquip. Deserunt aute enim ea minim officia anim culpa sint commodo. Cillum consectetur excepteur aliqua exercitation Lorem veniam voluptate.", "registered": "2016-07-06T01:31:07 -02:00", "latitude": -60.997048, "longitude": -102.397885, "tags": ["do", "ad", "consequat", "irure", "tempor", "elit", "minim"], "friends": [{"id": 0, "name": "Walker Hernandez"}, {"id": 1, "name": "Maria Lane"}, {"id": 2, "name": "Mcknight Barron"}], "greeting": "Hello, Heath Nguyen! You have 4 unread messages.", "favoriteFruit": "apple"}, {"_id": "655b6625519a5b5e4b6742bf", "index": 4, "guid": "c5dc685f-6d0d-4173-b4cf-f5df29a1e8ef", "isActive": true, "balance": "$1,358.90", "picture": "http://placehold.it/32x32", "age": 33, "eyeColor": "brown", "name": "Deidre Duke", "gender": "female", "company": "OATFARM", "email": "deidreduke@oatfarm.com", "phone": "+1 (875) 587-3256", "address": "487 Schaefer Street, Wattsville, West Virginia, 4506", "about": "Laboris eu nulla esse magna sit eu deserunt non est aliqua exercitation commodo. Ad occaecat qui qui laborum dolore anim Lorem. Est qui occaecat irure enim deserunt enim aliqua ex deserunt incididunt esse. Quis in minim laboris proident non mollit. Magna ea do labore commodo. Et elit esse esse occaecat officia ipsum nisi.", "registered": "2021-09-12T04:17:08 -02:00", "latitude": 68.609781, "longitude": -87.509134, "tags": ["mollit", "cupidatat", "irure", "sit", "consequat", "anim", "fugiat"], "friends": [{"id": 0, "name": "Bean Paul"}, {"id": 1, "name": "Cochran Hubbard"}, {"id": 2, "name": "Rodgers Atkinson"}], "greeting": "Hello, Deidre Duke! You have 6 unread messages.", "favoriteFruit": "apple"}, {"_id": "655b6625a19b3f7e5f82f0ea", "index": 5, "guid": "75f3c264-baa1-47a0-b21c-4edac23d9935", "isActive": true, "balance": "$3,554.36", "picture": "http://placehold.it/32x32", "age": 26, "eyeColor": "blue", "name": "Lydia Holland", "gender": "female", "company": "ESCENTA", "email": "lydiaholland@escenta.com", "phone": "+1 (927) 482-3436", "address": "554 Rockaway Parkway, Kohatk, Montana, 6316", "about": "Consectetur ea est labore commodo laborum mollit pariatur non enim. Est dolore et non laboris tempor. Ea incididunt ut adipisicing cillum labore officia tempor eiusmod commodo. Cillum fugiat ex consectetur ut nostrud anim nostrud exercitation ut duis in ea. Eu et id fugiat est duis eiusmod ullamco quis officia minim sint ea nisi in.", "registered": "2018-03-13T01:48:56 -01:00", "latitude": -88.495799, "longitude": 71.840667, "tags": ["veniam", "minim", "consequat", "consequat", "incididunt", "consequat", "elit"], "friends": [{"id": 0, "name": "Debra Massey"}, {"id": 1, "name": "Weiss Savage"}, {"id": 2, "name": "Shannon Guerra"}], "greeting": "Hello, Lydia Holland! You have 5 unread messages.", "favoriteFruit": "banana"}]'
+
+    import tempfile
+    # Create a temporary file
+    temp_fd, temp_path = tempfile.mkstemp(suffix=".json")
+    try:
+    # Write content to the temporary file
+        with os.fdopen(temp_fd, 'w') as tmp:
+            tmp.write("{key:value}")
+        assert(from_file(temp_path, logging=True)) == ({'key': 'value'}, [{'text': 'While parsing a string, we found a literal instead of a quote', 'context': '{key:value}'}, {'text': 'While parsing a string, we found no starting quote. Will add the quote back', 'context': '{key:value}'}, {'text': 'While parsing a string, we missed the closing quote, ignoring', 'context': '{key:value}'}, {'text': 'While parsing a string, we found a literal instead of a quote', 'context': '{key:value}'}, {'text': 'While parsing a string, we found no starting quote. Will add the quote back', 'context': '{key:value}'}, {'text': 'While parsing a string, we missed the closing quote, ignoring', 'context': '{key:value}'}])
+    finally:
+        # Clean up - delete the temporary file
+        os.remove(temp_path)

{json_repair-0.19.1 → json_repair-0.20.0}/tests/test_performance.py

@@ -42,7 +42,7 @@ def test_true_false_correct(benchmark):
   mean_time = benchmark.stats.get("median")
 
   # Define your time threshold in seconds
-  max_time = 20 * (1 / 10 ** 6)  # 20 microsecond
+  max_time = 30 * (1 / 10 ** 6)  # 30 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"
@@ -64,7 +64,7 @@ def test_false_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"
@@ -75,7 +75,7 @@ def test_false_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"
@@ -86,7 +86,7 @@ def test_false_false_correct(benchmark):
   mean_time = benchmark.stats.get("median")
 
   # Define your time threshold in seconds
-  max_time = 50 / 10 ** 6  # 50 microsecond
+  max_time = 60 / 10 ** 6  # 60 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"