json-repair 0.19.0__py3-none-any.whl → 0.19.2__py3-none-any.whl

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
 
@@ -309,6 +314,15 @@ 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
@@ -442,36 +456,22 @@ 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
@@ -493,22 +493,9 @@ 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,
@@ -584,3 +571,31 @@ 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-0.19.0.dist-info → json_repair-0.19.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: json_repair
-Version: 0.19.0
+Version: 0.19.2
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License

json_repair-0.19.2.dist-info/RECORD

@@ -0,0 +1,7 @@
+json_repair/__init__.py,sha256=IIzSm1DsCRrr8seF3UeMZXwxcq-tE3j-8d1WBxvEJvE,178
+json_repair/json_repair.py,sha256=qNkCy5by3jFhCAC3pDNKzDECOJr1Crh-P71j6N1CBBg,24405
+json_repair-0.19.2.dist-info/LICENSE,sha256=wrjQo8MhNrNCicXtMe3MHmS-fx8AmQk1ue8AQwiiFV8,1076
+json_repair-0.19.2.dist-info/METADATA,sha256=UdTpknO6GMk16WlPV-YxSFd_QbY25WiNkqj2IE3B_NA,7333
+json_repair-0.19.2.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+json_repair-0.19.2.dist-info/top_level.txt,sha256=7-VZwZN2CgB_n0NlSLk-rEUFh8ug21lESbsblOYuZqw,12
+json_repair-0.19.2.dist-info/RECORD,,

json_repair-0.19.0.dist-info/RECORD

@@ -1,7 +0,0 @@
-json_repair/__init__.py,sha256=IIzSm1DsCRrr8seF3UeMZXwxcq-tE3j-8d1WBxvEJvE,178
-json_repair/json_repair.py,sha256=ORf1wm6wTXriTJBtCJtoFU4rEw4daAoqV0ktdyhcOT0,23775
-json_repair-0.19.0.dist-info/LICENSE,sha256=wrjQo8MhNrNCicXtMe3MHmS-fx8AmQk1ue8AQwiiFV8,1076
-json_repair-0.19.0.dist-info/METADATA,sha256=Mj5le5SqwFzYmWGZWu5JbjZNqX4cYPP_h1XpQDYNeOI,7333
-json_repair-0.19.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
-json_repair-0.19.0.dist-info/top_level.txt,sha256=7-VZwZN2CgB_n0NlSLk-rEUFh8ug21lESbsblOYuZqw,12
-json_repair-0.19.0.dist-info/RECORD,,