From f31be6eed16daca947286a490d652fe0980796be Mon Sep 17 00:00:00 2001
From: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
Date: Tue, 23 Jan 2024 10:34:22 +0100
Subject: [PATCH] Minor improvements to type inference and readibility, not
 enough to warrant a new release

---
 src/json_repair/json_repair.py | 27 +++++++--------------------
 tests/test_performance.py      |  4 ++--
 2 files changed, 9 insertions(+), 22 deletions(-)

diff --git a/src/json_repair/json_repair.py b/src/json_repair/json_repair.py
index 34adde7..bbda6ac 100644
--- a/src/json_repair/json_repair.py
+++ b/src/json_repair/json_repair.py
@@ -207,17 +207,18 @@ def parse_string(self, use_single_quotes=False) -> str:
             if fixed_quotes:
                 if self.context == "object_key" and (char == ":" or char.isspace()):
                     break
-                elif self.context == "object_value" and (char == "," or char == "}"):
+                elif self.context == "object_value" and char in [",", "}"]:
                     break
             self.index += 1
             char = self.get_char_at()
             # ChatGPT sometimes forget to quote links in markdown like: { "content": "[LINK]("https://google.com")" }
             if (
                 char == string_terminator
-                and self.get_next_char() != ","
+                # Next character is not a comma
+                and self.get_char_at(1) != ","
                 and (
                     fix_broken_markdown_link
-                    or (self.get_prev_char(2) + self.get_prev_char()) == "]("
+                    or (self.get_char_at(-2) == "]" and self.get_char_at(-1)) == "("
                 )
             ):
                 fix_broken_markdown_link = not fix_broken_markdown_link
@@ -239,7 +240,7 @@ def parse_string(self, use_single_quotes=False) -> str:
 
         return self.json_str[start:end]
 
-    def parse_number(self) -> Union[float, int]:
+    def parse_number(self) -> Union[float, int, str]:
         # <number> is a valid real number expressed in one of a number of given formats
         number_str = ""
         number_chars = set("0123456789-.eE")
@@ -257,7 +258,7 @@ def parse_number(self) -> Union[float, int]:
             # This is a string then
             return self.parse_string()
 
-    def parse_boolean_or_null(self) -> Union[bool, None]:
+    def parse_boolean_or_null(self) -> Union[bool, str, None]:
         # <boolean> is one of the literal strings 'true', 'false', or 'null' (unquoted)
         boolean_map = {"true": (True, 4), "false": (False, 5), "null": (None, 4)}
         for key, (value, length) in boolean_map.items():
@@ -272,21 +273,7 @@ def insert_char_at(self, char: str) -> None:
         self.json_str = self.json_str[: self.index] + char + self.json_str[self.index :]
         self.index += 1
 
-    def get_char_at(self) -> Union[str, bool]:
-        # Why not use something simpler? Because we might be out of bounds and doing this check all the time is annoying
-        try:
-            return self.json_str[self.index]
-        except IndexError:
-            return False
-
-    def get_prev_char(self, count=1):
-        # Why not use something simpler? Because we might be out of bounds and doing this check all the time is annoying
-        try:
-            return self.json_str[self.index - count]
-        except IndexError:
-            return False
-
-    def get_next_char(self, count=1):
+    def get_char_at(self, count: int = 0) -> Union[str, bool]:
         # Why not use something simpler? Because we might be out of bounds and doing this check all the time is annoying
         try:
             return self.json_str[self.index + count]
diff --git a/tests/test_performance.py b/tests/test_performance.py
index 9534b9d..2e3ffb3 100644
--- a/tests/test_performance.py
+++ b/tests/test_performance.py
@@ -282,7 +282,7 @@ def test_true_true(benchmark):
   mean_time = benchmark.stats.get("median")
 
   # Define your time threshold in seconds (100ms in this case)
-  max_time = 1.1 / 10 ** 6  # 1.1 microsecond
+  max_time = 1.2 / 10 ** 6  # 1.2 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"
@@ -304,7 +304,7 @@ def test_false_true(benchmark):
   mean_time = benchmark.stats.get("median")
 
   # Define your time threshold in seconds (ms in this case)
-  max_time = 1 / 10 ** 3  # 1 millisecond
+  max_time = 1.2 / 10 ** 3  # 1.2 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"