openai4s - OpenAI Client for Scala
openai4s is OpenAI client for functional programming.
Here is an example Scala script with scala-cli in Scala 2.
First of all you need to set the following environment variables
export OPENAI_API_KEY=YOUR_API_KEY_FROM_OPENAIIf you don't have the API key yet, you should generate it first. Visit OpenAI's account page to do it.
openai4s-app.scala
//> using scala "2.13.14"
//> using options -Xsource:3
//> using dep "org.typelevel::cats-core:2.12.0"
//> using dep "io.kevinlee::openai4s-core:0.1.0-alpha12"
//> using dep "io.kevinlee::openai4s-config:0.1.0-alpha12"
//> using dep "io.kevinlee::openai4s-api:0.1.0-alpha12"
//> using dep "io.kevinlee::openai4s-http4s:0.1.0-alpha12"
//> using dep "com.github.pureconfig::pureconfig-cats-effect:0.17.7"
import cats.data.NonEmptyList
import cats.effect.*
import cats.syntax.all.*
import fs2.io.net.Network
import org.http4s.Uri as H4sUri
import org.http4s.ember.client.EmberClientBuilder
import scala.concurrent.duration.*
import eu.timepit.refined.types.*
import openai4s.api.ApiCore
import openai4s.api.chat.ChatApi
import openai4s.config.OpenAiConfig
import openai4s.http.HttpClient
import openai4s.types.common.*
import openai4s.types.chat.*
import eu.timepit.refined.auto.*
object MyAiApp extends IOApp.Simple {
override def run: IO[Unit] =
runChat[IO]
def runChat[F[*]: Async: Network]: F[Unit] = {
for {
openAiConfig <- pureconfig.module.catseffect.loadConfigF[F, OpenAiConfig]
_ <- EmberClientBuilder
.default[F]
.withTimeout(120.seconds)
.withIdleConnectionTime(120.seconds)
.build
.use { client =>
val httpClient = HttpClient(client)
for {
openAiApiUri <- Sync[F]
.catchNonFatal(
H4sUri.fromString(
openAiConfig.apiUri.chatCompletions.value
)
)
.rethrow
apiCore = ApiCore(openAiConfig.apiKey, httpClient)
chatApi = ChatApi(openAiApiUri, apiCore)
chat = Chat(
model = Model.gpt_4o,
messages = NonEmptyList.of(
Message(
Message.Role("user"),
Message.Content(
"Jane is faster than Joe. Joe is faster than Sam. Is Sam faster than Jane? Explain your reasoning step by step."
),
),
),
temperature = Temperature(0.1f).some,
maxTokens = none,
)
_ <- Sync[F].delay(println(show"Sending $chat"))
response <- chatApi.completion(chat)
_ <- Sync[F].delay(println(show"Response: $response"))
} yield ()
}
} yield ()
}
}Run
scala-cli run openai4s-app.scalaCompiling project (Scala 2.13.12, JVM (17))
Compiled project (Scala 2.13.12, JVM (17))
Sending Chat(model = Gpt_4o(value=gpt-4o, description=GPT-4o
Our most advanced, multimodal flagship model that’s cheaper and faster than GPT-4 Turbo. Currently points to gpt-4o-2024-05-13., maxTokens=128000, maxOutputTokens=4096, trainingData=Some(2023-10)), messages = NonEmptyList(Message(role = user, content = Jane is faster than Joe. Joe is faster than Sam. Is Sam faster than Jane? Explain your reasoning step by step.)), temperature = Some(0.1), maxTokens = None)
Response: Response(id = chatcmpl-A9Yv3oSHqqpsd0X5HiMNwhOE4VWwP, object = chat.completion, created = 2024-09-20T14:29:05Z, model = Gpt_4o_2024_05_13(value=gpt-4o-2024-05-13, description=gpt-4o currently points to this version., maxTokens=128000, maxOutputTokens=4096, trainingData=Some(2023-10)), usage = Usage(promptTokens = 32, completionTokens = 307, totalTokens = 339), choices = List(Choice(message = Message(role = assistant, content = To determine whether Sam is faster than Jane, let's analyze the information given step by step:
1. We are told that "Jane is faster than Joe." This establishes a relationship where Jane > Joe in terms of speed.
2. Next, we are told that "Joe is faster than Sam." This establishes another relationship where Joe > Sam in terms of speed.
3. To compare Sam and Jane, we can use the transitive property of inequality, which states that if A > B and B > C, then A > C.
4. Applying the transitive property to the relationships we have:
Since Jane > Joe and Joe > Sam, we can infer that Jane > Sam.
Therefore, based on the given information, Sam is not faster than Jane. In fact, Jane is faster than Sam.), finishReason = stop, index = 0)))