98 lines
3.0 KiB
Python
98 lines
3.0 KiB
Python
import orjson
|
|
from json_repair import repair_json
|
|
|
|
|
|
def _find_unclosed(json_str):
|
|
"""
|
|
Identifies the unclosed braces and brackets in the JSON string.
|
|
|
|
Args:
|
|
json_str (str): The JSON string to analyze.
|
|
|
|
Returns:
|
|
list: A list of unclosed elements in the order they were opened.
|
|
"""
|
|
unclosed = []
|
|
inside_string = False
|
|
escape_next = False
|
|
|
|
for char in json_str:
|
|
if inside_string:
|
|
if escape_next:
|
|
escape_next = False
|
|
elif char == "\\":
|
|
escape_next = True
|
|
elif char == '"':
|
|
inside_string = False
|
|
else:
|
|
if char == '"':
|
|
inside_string = True
|
|
elif char in "{[":
|
|
unclosed.append(char)
|
|
elif char in "}]":
|
|
if unclosed and ((char == "}" and unclosed[-1] == "{") or (char == "]" and unclosed[-1] == "[")):
|
|
unclosed.pop()
|
|
|
|
return unclosed
|
|
|
|
|
|
# The following code is used to fix a broken JSON string.
|
|
# From HippoRAG2 (GitHub: OSU-NLP-Group/HippoRAG)
|
|
def fix_broken_generated_json(json_str: str) -> str:
|
|
"""
|
|
Fixes a malformed JSON string by:
|
|
- Removing the last comma and any trailing content.
|
|
- Iterating over the JSON string once to determine and fix unclosed braces or brackets.
|
|
- Ensuring braces and brackets inside string literals are not considered.
|
|
|
|
If the original json_str string can be successfully loaded by orjson.loads(), will directly return it without any modification.
|
|
|
|
Args:
|
|
json_str (str): The malformed JSON string to be fixed.
|
|
|
|
Returns:
|
|
str: The corrected JSON string.
|
|
"""
|
|
|
|
try:
|
|
# Try to load the JSON to see if it is valid
|
|
orjson.loads(json_str)
|
|
return json_str # Return as-is if valid
|
|
except orjson.JSONDecodeError: ...
|
|
|
|
# Step 1: Remove trailing content after the last comma.
|
|
last_comma_index = json_str.rfind(",")
|
|
if last_comma_index != -1:
|
|
json_str = json_str[:last_comma_index]
|
|
|
|
# Step 2: Identify unclosed braces and brackets.
|
|
unclosed_elements = _find_unclosed(json_str)
|
|
|
|
# Step 3: Append the necessary closing elements in reverse order of opening.
|
|
closing_map = {"{": "}", "[": "]"}
|
|
for open_char in reversed(unclosed_elements):
|
|
json_str += closing_map[open_char]
|
|
|
|
return json_str
|
|
|
|
|
|
def new_fix_broken_generated_json(json_str: str) -> str:
|
|
"""
|
|
使用 json-repair 库修复格式错误的 JSON 字符串。
|
|
|
|
如果原始 json_str 字符串可以被 orjson.loads() 成功加载，则直接返回而不进行任何修改。
|
|
|
|
参数:
|
|
json_str (str): 需要修复的格式错误的 JSON 字符串。
|
|
|
|
返回:
|
|
str: 修复后的 JSON 字符串。
|
|
"""
|
|
try:
|
|
# 尝试加载 JSON 以查看其是否有效
|
|
orjson.loads(json_str)
|
|
return json_str # 如果有效则按原样返回
|
|
except orjson.JSONDecodeError:
|
|
# 如果无效，则尝试修复它
|
|
return repair_json(json_str)
|