[Feature] I need solutions for re-unicode Chinese character and adjust floating point precision

[SMFDrummer]

What happened?

My original data is like

{
    "sd": {
        "n": "\u81ea\u4fe1\u7684\u5927\u5634\u82b1",
        "lmz": 1.000000,
        "rsd": {
            "wr": 0.300000,
            "lr": 0.300000,
            "bc": 0
        }
    }
}

When I use parseToJsonElement and encodeToString, this json string changed like

{
    "sd": {
        "n": "自信的大嘴花",
        "lmz": 1.0,
        "rsd": {
            "wr": 0.3,
            "lr": 0.3,
            "bc": 0
        }
    }
}

I'd like to

The reason why I not make this json to a @serializable data class is because this json is so mass, thousands of lines. I literally cannot handle this. I have to use JsonElement as JsonObject instead. So how to keep the unicode non-convert, and how to make the float value maintain 6 digit precision. Please give me some tips.

[SMFDrummer]

I've tested for KSerializer

import kotlinx.serialization.KSerializer
import kotlinx.serialization.Serializable
import kotlinx.serialization.descriptors.PrimitiveKind
import kotlinx.serialization.descriptors.PrimitiveSerialDescriptor
import kotlinx.serialization.descriptors.SerialDescriptor
import kotlinx.serialization.encodeToString
import kotlinx.serialization.encoding.Decoder
import kotlinx.serialization.encoding.Encoder
import kotlinx.serialization.json.Json
import org.jetbrains.annotations.TestOnly

@Serializable
data class Data(
    val rs: Int,
    @Serializable(with = UnicodeStringSerializer::class)
    val n: String
)

object UnicodeStringSerializer : KSerializer<String> {
    override val descriptor: SerialDescriptor = PrimitiveSerialDescriptor("UnicodeString", PrimitiveKind.STRING)
    override fun deserialize(decoder: Decoder): String {
        val decoded = decoder.decodeString()
        return decoded.split("\\u")
            .filter { it.isNotEmpty() }
            .joinToString("") {
                it.toInt(16).toChar().toString()
            }
    }

    override fun serialize(encoder: Encoder, value: String) {
        val unicodeEscaped = value.toCharArray().joinToString("") {
            "\\u" + it.code.toString(16).padStart(4, '0')
        }
        encoder.encodeString(unicodeEscaped)
    }
}

@TestOnly
fun main() {
    val data = Data(rs = 1703239116, n = "自信的大嘴花")
    val json = Json { encodeDefaults = true }
    val jsonString = json.encodeToString(data)
    println(jsonString)
}

then I get result like

{"rs":1703239116,"n":"\\u81ea\\u4fe1\\u7684\\u5927\\u5634\\u82b1"}

but I want is

{"rs":1703239116,"n":"\u81ea\u4fe1\u7684\u5927\u5634\u82b1"}

is kotlin cannot handle this? Please give me some advice.

[sandwwraith]

We do not have any setting to control output formatting, unfortunately. You can try to use JsonUnquotedLiteral together with some JsonTransformer: https://github.com/Kotlin/kotlinx.serialization/blob/master/docs/json.md#encoding-literal-json-content-experimental

[SMFDrummer]

DAMN Thank YOU AAAAAAAAAAAAA
I used a third-party lib: @nomisRev: kotlinx-serialization-jsonpath, drived by Arrow. This library is so sick.(Apology for at)

fun String.addQuotes(): String = "\"$this\""
fun String.ensureAscii(): String = this.toCharArray().joinToString("") { "\\u" + it.code.toString(16).padStart(4, '0') }

val origin = """{"sd":{"n":"\u81ea\u4fe1\u7684\u5927\u5634\u82b1","lmz":1.000000,"rsd":{"wr":0.300000,"lr":0.300000,"bc":0}}}"""

val data = Json.parse(origin) as JsonObject // {"sd":{"n":"自信的大嘴花","lmz":1.0,"rsd":{"wr":0.3,"lr":0.3,"bc":0}}}
// I can modify others here, and do something else..., and then ->
val data1 = JsonPath.path("sd.n").modify(data) {
    JsonUnquotedLiteral(it.jsonPrimitive.content.ensureAscii().escaped().addQuotes())
} // {"sd":{"n":"\u81ea\u4fe1\u7684\u5927\u5634\u82b1","lmz":1.0,"rsd":{"wr":0.3,"lr":0.3,"bc":0}}}

val data2 = JsonPath.path("sd.lmz").modify(data1) {
    JsonUnquotedLiteral(it.jsonPrimitive.content.toBigDecimal().setScale(6).toString())
} // {"sd":{"n":"\u81ea\u4fe1\u7684\u5927\u5634\u82b1","lmz":1.000000,"rsd":{"wr":0.3,"lr":0.3,"bc":0}}}

val data3 = JsonPath.path("sd.rsd.wr").modify(data2) {
    JsonUnquotedLiteral(it.jsonPrimitive.content.toBigDecimal().setScale(6).toString())
} // {"sd":{"n":"\u81ea\u4fe1\u7684\u5927\u5634\u82b1","lmz":1.000000,"rsd":{"wr":0.300000,"lr":0.3,"bc":0}}}

val data4 = JsonPath.path("sd.rsd.lr").modify(data3) {
    JsonUnquotedLiteral(it.jsonPrimitive.content.toBigDecimal().setScale(6).toString())
} // {"sd":{"n":"\u81ea\u4fe1\u7684\u5927\u5634\u82b1","lmz":1.000000,"rsd":{"wr":0.300000,"lr":0.300000,"bc":0}}}

println(Json.encodeToString(data4) == origin) // true

Although this issue has been resolved, as I am not very familiar with the Arrow library and am new to using the JsonPath library, I do not know how to combine them together to avoid declaring so many intermediate variables. I apologize for taking up your time. As a beginner, I would like to ask you how to reduce intermediate variables?(My code is not very standardized, please forgive me)

[sandwwraith]

Unfortunately, I'm not familiar with JsonPath or Arrow so I can't help you here

[SMFDrummer]

Unfortunately, I'm not familiar with JsonPath or Arrow so I can't help you here

Hi I'm back, you and other kotliner could watch this issue to modify JsonElement: nomisRev/kotlinx-serialization-jsonpath#69