json-repair 0.46.2__py3-none-any.whl → 0.47.1__py3-none-any.whl

json_repair/json_parser.py

@@ -365,13 +365,14 @@ class JSONParser:
                 )
                 break
             if (
-                (missing_quotes or not self.stream_stable)
+                not self.stream_stable
                 and self.context.current == ContextValues.OBJECT_VALUE
                 and char
                 in [
                     ",",
                     "}",
                 ]
+                and string_acc[-1] != rstring_delimiter
             ):
                 rstring_delimiter_missing = True
                 # check if this is a case in which the closing comma is NOT missing instead
@@ -434,9 +435,10 @@ class JSONParser:
                     )
                     break
             if (
-                (missing_quotes or not self.stream_stable)
+                not self.stream_stable
                 and char == "]"
                 and ContextValues.ARRAY in self.context.context
+                and string_acc[-1] != rstring_delimiter
             ):
                 # We found the end of an array and we are in array context
                 # So let's check if we find a rstring_delimiter forward otherwise end early
@@ -456,9 +458,17 @@ class JSONParser:
                 if char in [rstring_delimiter, "t", "n", "r", "b", "\\"]:
                     string_acc = string_acc[:-1]
                     escape_seqs = {"t": "\t", "n": "\n", "r": "\r", "b": "\b"}
-                    string_acc += escape_seqs.get(char, char) or char
+                    string_acc += escape_seqs.get(char, char)
                     self.index += 1
                     char = self.get_char_at()
+                    while char and string_acc[-1] == "\\" and char in [rstring_delimiter, "\\"]:
+                        # this is a bit of a special case, if I don't do this it will close the loop or create a train of \\
+                        # I don't love it though
+                        string_acc = string_acc[:-1]
+                        string_acc += char
+                        self.index += 1
+                        char = self.get_char_at()
+                    continue
                 elif char in ["u", "x"]:
                     # If we find a unicode escape sequence, normalize it
                     num_chars = 4 if char == "u" else 2
@@ -499,7 +509,7 @@ class JSONParser:
                     )
                     break
             # ChatGPT sometimes forget to quote stuff in html tags or markdown, so we do this whole thing here
-            if char == rstring_delimiter:
+            if char == rstring_delimiter and string_acc[-1] != "\\":
                 # Special case here, in case of double quotes one after another
                 if doubled_quotes and self.get_char_at(1) == rstring_delimiter:
                     self.log("While parsing a string, we found a doubled quote, ignoring it")

json_repair/json_repair.py

@@ -37,9 +37,9 @@ def repair_json(
     skip_json_loads: bool = False,
     logging: bool = False,
     json_fd: TextIO | None = None,
-    ensure_ascii: bool = True,
     chunk_length: int = 0,
     stream_stable: bool = False,
+    **json_dumps_args,
 ) -> str: ...
 
 
@@ -50,9 +50,9 @@ def repair_json(
     skip_json_loads: bool = False,
     logging: bool = False,
     json_fd: TextIO | None = None,
-    ensure_ascii: bool = True,
     chunk_length: int = 0,
     stream_stable: bool = False,
+    **json_dumps_args,
 ) -> JSONReturnType | tuple[JSONReturnType, list[dict[str, str]]]: ...
 
 
@@ -62,9 +62,9 @@ def repair_json(
     skip_json_loads: bool = False,
     logging: bool = False,
     json_fd: TextIO | None = None,
-    ensure_ascii: bool = True,
     chunk_length: int = 0,
     stream_stable: bool = False,
+    **json_dumps_args,
 ) -> JSONReturnType | tuple[JSONReturnType, list[dict[str, str]]]:
     """
     Given a json formatted string, it will try to decode it and, if it fails, it will try to fix it.
@@ -96,7 +96,7 @@ def repair_json(
     # Avoid returning only a pair of quotes if it's an empty string
     elif parsed_json == "":
         return ""
-    return json.dumps(parsed_json, ensure_ascii=ensure_ascii)
+    return json.dumps(parsed_json, **json_dumps_args)
 
 
 def loads(

{json_repair-0.46.2.dist-info → json_repair-0.47.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: json_repair
-Version: 0.46.2
+Version: 0.47.1
 Summary: A package to repair broken json strings
 Author-email: Stefano Baccianella <4247706+mangiucugna@users.noreply.github.com>
 License: MIT License
@@ -175,6 +175,10 @@ will return
 
     {"test_chinese_ascii": "统一码"}
 
+### JSON dumps parameters
+
+More in general, `repair_json` will accept all parameters that `json.dumps` accepts and just pass them through (for example indent)
+
 ### Performance considerations
 If you find this library too slow because is using `json.loads()` you can skip that by passing `skip_json_loads=True` to `repair_json`. Like:
 

{json_repair-0.46.2.dist-info → json_repair-0.47.1.dist-info}/RECORD

@@ -1,14 +1,14 @@
 json_repair/__init__.py,sha256=6FDD6dEVM5Pb5o4Zodgw4ex30Hzy-YvNRy0vts9SQ4I,118
 json_repair/__main__.py,sha256=EsJb-y89uZEvGQQg1GdIDWzfDwfOMvVekKEtdguQXCM,67
 json_repair/json_context.py,sha256=WsMOjqpGSr6aaDONcrk8UFtTurzWon2Qq9AoBBYseoI,934
-json_repair/json_parser.py,sha256=B-DgJfyQOMHQ3F0RIBnltUGnGw0DFM-J7xOcLmCylVs,39744
-json_repair/json_repair.py,sha256=pyH5fCkS1lyNPVjkqXerQ91lBz3eTHDPgV1QtnvJm-Y,11243
+json_repair/json_parser.py,sha256=YBi07AfBGoZ54locsc6j1Y7WfdretFzmt0wXDEWwRo8,40321
+json_repair/json_repair.py,sha256=0qL2LuzlNJa3VnEqYNaJyZNAL2w18oAt2YvA-TlMxmY,11211
 json_repair/object_comparer.py,sha256=LlIF0MisRglzC-CiG5AxAEDCBWBHeJd-6uXYx0uRmCk,1175
 json_repair/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 json_repair/string_file_wrapper.py,sha256=tGkWBEUPE-CZPf4uSM5NE9oSDTpskX0myJiXsl-gbds,4333
-json_repair-0.46.2.dist-info/licenses/LICENSE,sha256=wrjQo8MhNrNCicXtMe3MHmS-fx8AmQk1ue8AQwiiFV8,1076
-json_repair-0.46.2.dist-info/METADATA,sha256=-EKRFk4rzF6I4EqFqEVfXJn7aPFrgFzdf1oCZfWgYLE,12208
-json_repair-0.46.2.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-json_repair-0.46.2.dist-info/entry_points.txt,sha256=SNfge3zPSP-ASqriYU9r3NAPaXdseYr7ciPMKdV2uSw,57
-json_repair-0.46.2.dist-info/top_level.txt,sha256=7-VZwZN2CgB_n0NlSLk-rEUFh8ug21lESbsblOYuZqw,12
-json_repair-0.46.2.dist-info/RECORD,,
+json_repair-0.47.1.dist-info/licenses/LICENSE,sha256=wrjQo8MhNrNCicXtMe3MHmS-fx8AmQk1ue8AQwiiFV8,1076
+json_repair-0.47.1.dist-info/METADATA,sha256=AgW6QikkfasSzZk-AplVvSkixKfeh4aEUx1UPFFHWmA,12368
+json_repair-0.47.1.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+json_repair-0.47.1.dist-info/entry_points.txt,sha256=SNfge3zPSP-ASqriYU9r3NAPaXdseYr7ciPMKdV2uSw,57
+json_repair-0.47.1.dist-info/top_level.txt,sha256=7-VZwZN2CgB_n0NlSLk-rEUFh8ug21lESbsblOYuZqw,12
+json_repair-0.47.1.dist-info/RECORD,,