SnakePickle.snakeToCamel crashes on keys with leading, trailing, or consecutive underscores

[haskiindahouse]

Hit this while decoding a JSON payload that contained a "_request_id" field. SnakePickle.snakeToCamel calls x(0) on each segment without checking for empty strings, which throws StringIndexOutOfBoundsException whenever split("_", -1) produces an empty segment — i.e. for any key with leading, trailing, or consecutive underscores.

Reproducer (scala-cli):

//> using scala 3.8.3

@main def repro(): Unit =
  // Same logic as core/src/main/scala/sttp/ai/core/json/SnakePickle.scala (lines 13-14)
  def snakeToCamel(s: String): String =
    val res = s.split("_", -1).map(x => s"${x(0).toUpper}${x.drop(1)}").mkString
    s"${s(0).toLower}${res.drop(1)}"

  snakeToCamel("_type")        // throws
  snakeToCamel("some__key")    // throws
  snakeToCamel("trailing_")    // throws

Output:

java.lang.StringIndexOutOfBoundsException: Index 0 out of bounds for length 0
  at SnakePickle$.snakeToCamel(SnakePickle.scala:14)

Source pointer:

sttp-ai/core/src/main/scala/sttp/ai/core/json/SnakePickle.scala

Lines 13 to 14 in fa739fa

	private def snakeToCamel(s: String): String = {
	val res = s.split("_", -1).map(x => s"${x(0).toUpper}${x.drop(1)}").mkString

Any OpenAI-style API field whose JSON name starts with _ (private/internal fields are common) will trigger this. Filtering empty segments — s.split("_", -1).filter(_.nonEmpty) — fixes it without changing behaviour for well-formed keys. There's a related but milder issue in camelToSnake where strings starting with an uppercase letter end up with a leading underscore, but that one isn't a crash. Happy to PR both.

[adamw]

Thanks for the report - PRs welcome!