json-repair 0.13.0__py3-none-any.whl → 0.14.0__py3-none-any.whl

json_repair/json_repair.py

@@ -34,6 +34,7 @@ class JSONParser:
         self.index = 0
         # This is used in the object member parsing to manage the special cases of missing quotes in key or value
         self.context = ""
+        self.context_stack = []
 
     def parse(self) -> Union[Dict[str, Any], List[Any], str, float, int, bool, None]:
         return self.parse_json()
@@ -97,7 +98,7 @@ class JSONParser:
 
             # We are now searching for they string key
             # Context is used in the string parser to manage the lack of quotes
-            self.context = "object_key"
+            self.update_context("object_key")
 
             self.skip_whitespaces_at()
 
@@ -119,12 +120,13 @@ class JSONParser:
             if (self.get_char_at() or "") != ":":
                 self.insert_char_at(":")
             self.index += 1
-            self.context = "object_value"
+            self.update_context("")
+            self.update_context("object_value")
             # The value can be any valid json
             value = self.parse_json()
 
             # Reset context since our job is done
-            self.context = ""
+            self.update_context("")
             obj[key] = value
 
             if (self.get_char_at() or "") in [",", "'", '"']:
@@ -157,6 +159,9 @@ 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.context == "object_value" and char == "}":
+                break
 
         # Especially at the end of an LLM generated json you might miss the last "]"
         char = self.get_char_at()
@@ -166,6 +171,7 @@ class JSONParser:
                 # Remove trailing "," before adding the "]"
                 self.remove_char_at()
             self.insert_char_at("]")
+            self.index -= 1
 
         self.index += 1
         return arr
@@ -177,15 +183,14 @@ class JSONParser:
 
         # Flag to manage corner cases related to missing starting quote
         fixed_quotes = False
-        double_delimiter = False
         lstring_delimiter = rstring_delimiter = '"'
         if isinstance(string_quotes, list):
             lstring_delimiter = string_quotes[0]
             rstring_delimiter = string_quotes[1]
         elif isinstance(string_quotes, str):
             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:
-            double_delimiter = True
             self.index += 1
         char = self.get_char_at()
         if char != lstring_delimiter:
@@ -204,7 +209,6 @@ class JSONParser:
         # * It iterated over the entire sequence
         # * If we are fixing missing quotes in an object, when it finds the special terminators
         char = self.get_char_at()
-        fix_broken_markdown_link = False
         while char and char != rstring_delimiter:
             if fixed_quotes:
                 if self.context == "object_key" and (char == ":" or char.isspace()):
@@ -221,19 +225,27 @@ class JSONParser:
                 else:
                     self.remove_char_at(-1)
                     self.index -= 1
-            # ChatGPT sometimes forget to quote links in markdown like: { "content": "[LINK]("https://google.com")" }
+            # 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 comma
-                and self.get_char_at(1) != ","
-                and (
-                    fix_broken_markdown_link
-                    or (self.get_char_at(-2) == "]" and self.get_char_at(-1)) == "("
-                )
+                # Next character is not a delimiter
+                and self.get_char_at(1) not in [",", ":", "]", "}"]
             ):
-                fix_broken_markdown_link = not fix_broken_markdown_link
-                self.index += 1
-                char = self.get_char_at()
+                # Special case here, in case of double quotes one after another
+                if self.get_char_at(1) == rstring_delimiter:
+                    # self destruct this character
+                    self.remove_char_at()
+                else:
+                    # Check if eventually there is a rstring delimiter, otherwise we bail
+                    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)
+                    # In that case we ignore this rstring delimiter
+                    if next_c:
+                        self.index += 1
+                        char = self.get_char_at()
 
         if char and fixed_quotes and self.context == "object_key" and char.isspace():
             self.skip_whitespaces_at()
@@ -247,10 +259,8 @@ class JSONParser:
             self.insert_char_at(rstring_delimiter)
         else:
             self.index += 1
-            if double_delimiter and self.get_char_at() == rstring_delimiter:
-                self.index += 1
 
-        return self.json_str[start:end]
+        return self.json_str[start:end].rstrip()
 
     def parse_number(self) -> Union[float, int, str]:
         # <number> is a valid real number expressed in one of a number of given formats
@@ -319,6 +329,19 @@ class JSONParser:
             except IndexError:
                 return
 
+    def update_context(self, value: str) -> None:
+        # If a value is provided update the context variable and save in stack
+        if value:
+            if self.context:
+                self.context_stack.append(self.context)
+            self.context = value
+        # Otherwise pop and update the context, or empty if the stack is empty
+        else:
+            try:
+                self.context = self.context_stack.pop()
+            except Exception:
+                self.context = ""
+
 
 def repair_json(
     json_str: str, return_objects: bool = False, skip_json_loads: bool = False

{json_repair-0.13.0.dist-info → json_repair-0.14.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.13.0
+Version: 0.14.0
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License
@@ -150,7 +150,10 @@ You will need owner access to this repository
 - Run `python -m build`
 - Create a new release in Github, making sure to tag all the issues solved and contributors. Create the new tag, same as the one in the build configuration
 - Once the release is created, a new Github Actions workflow will start to publish on Pypi, make sure it didn't fail
-
+---
+# Repair JSON in other programming languages
+- Typescript: https://github.com/josdejong/jsonrepair
+- Go: https://github.com/RealAlexandreAI/json-repair
 ---
 # Bonus Content
 If you need some good Custom Instructions (System Message) to improve your chatbot responses try https://gist.github.com/mangiucugna/7ec015c4266df11be8aa510be0110fe4

json_repair-0.14.0.dist-info/RECORD

@@ -0,0 +1,7 @@
+json_repair/__init__.py,sha256=AlNie5y6BZBioGi5fzTAUvum_y0U5aL5aNsuQ_68LQc,175
+json_repair/json_repair.py,sha256=8B5HfWoLlUUtRYq1cnbajOxWiMSD9nxNW2cRFPjFfVE,15817
+json_repair-0.14.0.dist-info/LICENSE,sha256=wrjQo8MhNrNCicXtMe3MHmS-fx8AmQk1ue8AQwiiFV8,1076
+json_repair-0.14.0.dist-info/METADATA,sha256=82KldmuVFLXbNy6SXar9MsulkcBUM1K8RX13pNysHQU,7355
+json_repair-0.14.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+json_repair-0.14.0.dist-info/top_level.txt,sha256=7-VZwZN2CgB_n0NlSLk-rEUFh8ug21lESbsblOYuZqw,12
+json_repair-0.14.0.dist-info/RECORD,,

json_repair-0.13.0.dist-info/RECORD

@@ -1,7 +0,0 @@
-json_repair/__init__.py,sha256=AlNie5y6BZBioGi5fzTAUvum_y0U5aL5aNsuQ_68LQc,175
-json_repair/json_repair.py,sha256=BHDcvo8C-vc3QO4dhgD5cP7IZVkGWGDbv8hNmQH1B-I,14653
-json_repair-0.13.0.dist-info/LICENSE,sha256=wrjQo8MhNrNCicXtMe3MHmS-fx8AmQk1ue8AQwiiFV8,1076
-json_repair-0.13.0.dist-info/METADATA,sha256=E6v_JcBwXe4oMJhtxgF-9-24kG2j5LiH-ivqufNHFpA,7200
-json_repair-0.13.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
-json_repair-0.13.0.dist-info/top_level.txt,sha256=7-VZwZN2CgB_n0NlSLk-rEUFh8ug21lESbsblOYuZqw,12
-json_repair-0.13.0.dist-info/RECORD,,