为何Python无法解析含转义引号的合法JSON？

以下是测试用的日志字符串：

TEST_LINE = """Oct 21 22:39:28 GMT [TRACE] (Carlos-288) org.some.awesome.LoggerFramework RID=8e9076-4dd9-ec96-8f35-bde193498f: {
    "service": "MyService",
    "operation": "queryShowSize",
    "requestID": "8e9076-4dd9-ec96-8f35-bde193498f",
    "timestamp": 1634815968000,
    "parameters": [
        {
            "__type": "org.some.awsome.code.service#queryShowSizeRequest",
            "externalID": {
                "__type": "org.some.awsome.code.common#CustomerID",
                "value": "48317"
            },
            "CountryID": {
                "__type": "org.some.awsome.code.common#CountryID",
                "value": "125"
            },
            "operationOriginalDate": 1.63462085667E9,
            "operationType": "MeasureWithToes",
            "measureInstrumentIdentifier": "595909-48d2-6115-85e8-b3aa7b"
        }
    ],
    "output": {
        "__type": "org.some.awsome.code.common#queryShowSizeReply",
        "shoeSize": {
            "value": "$ion_1_0 'org.some.awsome.model.processing.ShoeMeasurementBI@1.0'::'org.some.awsome.model.processing.FeetScience@1.0'::{customer_id:\"983017317\",measureInstrumentIdentifierTilda:\"595909-48d2-6115-85e8-b3aa7b\",foot_owner:\"Oedipus\",toe_code:\"LR2X10\",account_number_token:\"1234-2838316-1298470\",token_status:VALID,country_code:GRC,measure_store_format:METRIC}"
        }
    }
}
"""

我用正则提取出其中的JSON部分，验证为合法JSON，但Python的JSON模块解析失败，报错信息如下：

Exception has occurred: JSONDecodeError
Expecting ',' delimiter: line 25 column 156 (char 992)
  File "/Users/decoder/Downloads/json-problem.py", line 44, in read_json
    d = json.loads(line)
        ^^^^^^^^^^^^^^^^
  File "/Users/decoder/Downloads/json-problem.py", line 48, in <module>
    print(read_json(TEST_LINE))
          ^^^^^^^^^^^^^^^^^^^^
json.decoder.JSONDecodeError: Expecting ',' delimiter: line 25 column 156 (char 992)

错误指向output.shoeSize.value中的第一个\"，该嵌入的类JSON结构应作为普通字符串处理，且引号已转义，不应提前终止字符串。我的解析代码如下：

import re
import json

FIND_JSON = re.compile(
    r"\w{3} \d{2} (\d{2}[: ]){3}GMT \[[^]]+\] \([^)]+\) "
    r"org.some.awesome.LoggerFramework RID=[^:]+: "
)

def read_json(line: str) -> str | None:
    if not (m := FIND_JSON.match(line)):
        return None
    line = line[m.end(0) :]
    d = json.loads(line)
    return d


print(read_json(TEST_LINE))

尝试过raw_decode()方法，同样失败。

更新1：有评论指出存在未转义的双引号，但我看到的都是冒号后紧跟反斜杠加双引号，且验证合法，怀疑是复制粘贴时的格式转换问题。
更新2：补充了可复现问题的完整代码。

问题出在Python字符串的转义处理逻辑上：在Python的三重引号字符串中，\"会被自动解析为单个双引号，而非保留\和"的组合。也就是说，当你定义TEST_LINE时，里面的\"已经被转成了"，导致提取后的JSON字符串出现未转义的双引号，触发了解析错误。

举个例子，value字段的内容在Python实际存储时变成了：

$ion_1_0 'org.some.awsome.model.processing.ShoeMeasurementBI@1.0'::'org.some.awsome.model.processing.FeetScience@1.0'::{customer_id:"983017317",...}

这里的"983017317"没有被转义，直接终止了外层字符串，引发报错。

有两种解决方法：

方法1：使用原始字符串定义TEST_LINE #

在三重引号前加r，让Python不处理字符串内的转义字符：

TEST_LINE = r"""Oct 21 22:39:28 GMT [TRACE] (Carlos-288) org.some.awesome.LoggerFramework RID=8e9076-4dd9-ec96-8f35-bde193498f: {
    "service": "MyService",
    "operation": "queryShowSize",
    "requestID": "8e9076-4dd9-ec96-8f35-bde193498f",
    "timestamp": 1634815968000,
    "parameters": [
        {
            "__type": "org.some.awsome.code.service#queryShowSizeRequest",
            "externalID": {
                "__type": "org.some.awsome.code.common#CustomerID",
                "value": "48317"
            },
            "CountryID": {
                "__type": "org.some.awsome.code.common#CountryID",
                "value": "125"
            },
            "operationOriginalDate": 1.63462085667E9,
            "operationType": "MeasureWithToes",
            "measureInstrumentIdentifier": "595909-48d2-6115-85e8-b3aa7b"
        }
    ],
    "output": {
        "__type": "org.some.awsome.code.common#queryShowSizeReply",
        "shoeSize": {
            "value": "$ion_1_0 'org.some.awsome.model.processing.ShoeMeasurementBI@1.0'::'org.some.awsome.model.processing.FeetScience@1.0'::{customer_id:\"983017317\",measureInstrumentIdentifierTilda:\"595909-48d2-6115-85e8-b3aa7b\",foot_owner:\"Oedipus\",toe_code:\"LR2X10\",account_number_token:\"1234-2838316-1298470\",token_status:VALID,country_code:GRC,measure_store_format:METRIC}"
        }
    }
}
"""

方法2：对转义符进行二次转义 #

把字符串中的\"替换为\\"，让Python解析后保留\"的组合：

TEST_LINE = """Oct 21 22:39:28 GMT [TRACE] (Carlos-288) org.some.awesome.LoggerFramework RID=8e9076-4dd9-ec96-8f35-bde193498f: {
    "service": "MyService",
    "operation": "queryShowSize",
    "requestID": "8e9076-4dd9-ec96-8f35-bde193498f",
    "timestamp": 1634815968000,
    "parameters": [
        {
            "__type": "org.some.awsome.code.service#queryShowSizeRequest",
            "externalID": {
                "__type": "org.some.awsome.code.common#CustomerID",
                "value": "48317"
            },
            "CountryID": {
                "__type": "org.some.awsome.code.common#CountryID",
                "value": "125"
            },
            "operationOriginalDate": 1.63462085667E9,
            "operationType": "MeasureWithToes",
            "measureInstrumentIdentifier": "595909-48d2-6115-85e8-b3aa7b"
        }
    ],
    "output": {
        "__type": "org.some.awsome.code.common#queryShowSizeReply",
        "shoeSize": {
            "value": "$ion_1_0 'org.some.awsome.model.processing.ShoeMeasurementBI@1.0'::'org.some.awsome.model.processing.FeetScience@1.0'::{customer_id:\\\"983017317\\\",measureInstrumentIdentifierTilda:\\\"595909-48d2-6115-85e8-b3aa7b\\\",foot_owner:\\\"Oedipus\\\",toe_code:\\\"LR2X10\\\",account_number_token:\\\"1234-2838316-1298470\\\",token_status:VALID,country_code:GRC,measure_store_format:METRIC}"
        }
    }
}
"""

两种方法都能让提取后的JSON字符串保留正确的转义引号，从而被Python的json模块正常解析。

内容的提问来源于stack exchange，提问作者primfaktor