Update

aiapi/src/main/kotlin/eu/m724/aiapi/AiApi.kt

@@ -7,6 +7,7 @@ import io.ktor.client.engine.cio.CIO
 import io.ktor.client.plugins.contentnegotiation.ContentNegotiation
 import io.ktor.client.plugins.sse.SSE
 import io.ktor.client.plugins.sse.SSEBufferPolicy
+import io.ktor.client.plugins.sse.deserialize
 import io.ktor.client.plugins.sse.sse
 import io.ktor.client.request.header
 import io.ktor.client.request.setBody
@@ -15,14 +16,11 @@ import io.ktor.http.HttpMethod
 import io.ktor.http.contentType
 import io.ktor.http.headers
 import io.ktor.serialization.kotlinx.json.json
-import kotlinx.coroutines.channels.awaitClose
+import io.ktor.sse.TypedServerSentEvent
 import kotlinx.coroutines.flow.Flow
-import kotlinx.coroutines.flow.callbackFlow
-import kotlinx.coroutines.flow.flow
-import kotlinx.coroutines.flow.map
 import kotlinx.coroutines.flow.mapNotNull
-import kotlinx.coroutines.launch
 import kotlinx.serialization.json.Json
+import kotlinx.serialization.serializer
 
 class AiApi(
     private val session: String
@@ -50,17 +48,16 @@ class AiApi(
                 method = HttpMethod.Post
                 contentType(ContentType.Application.Json)
                 setBody(requestBody)
+            },
+            deserialize = {
+                    typeInfo, jsonString ->
+                val serializer = Json.serializersModule.serializer(typeInfo.kotlinType!!)
+                Json.decodeFromString(serializer, jsonString)!!
             }
         ) {
             block(
-                incoming.mapNotNull {
-                    try {
-                        val streamGptEvent = Json.decodeFromString<StreamGptEvent.Completion>(it.data!!)
-                        streamGptEvent
-                    } catch (e: Exception) {
-                        // TODO
-                        null
-                    }
+                incoming.mapNotNull { event: TypedServerSentEvent<String> ->
+                    deserialize<StreamGptEvent.Completion>(event.data)
                 }
             )
         }

src/main/kotlin/App.kt

@@ -14,6 +14,7 @@ import com.jakewharton.mosaic.layout.size
 import com.jakewharton.mosaic.layout.wrapContentSize
 import com.jakewharton.mosaic.layout.wrapContentWidth
 import com.jakewharton.mosaic.modifier.Modifier
+import com.jakewharton.mosaic.ui.Color
 import com.jakewharton.mosaic.ui.Text
 import eu.m724.modifier.BorderedBoxWithTabs
 import eu.m724.modifier.RunTitle
@@ -41,6 +42,11 @@ fun App(
         },
         subTitle = runs[selectedTabIndex].model ?: ""
     ) {
+        Text(
+            value = runs[selectedTabIndex].reasoningContent.wrap(90),
+            color = Color(200, 200, 200)
+        )
+
         Text(
             value = runs[selectedTabIndex].content.wrap(90)
         )

src/main/kotlin/ViewModel.kt

@@ -28,10 +28,11 @@ class ViewModel {
         var index = 0
         _runs.update {
             index = it.size
-            it + Run(RunState.Queued, "")
+            it + Run(RunState.Queued, "", "")
         }
 
-        var response = ""
+        var reasoningContent = ""
+        var responseContent = ""
 
         val timeSource = TimeSource.Monotonic
         val startMark by lazy {
@@ -43,6 +44,7 @@ class ViewModel {
         GlobalScope.launch(Dispatchers.IO) {
             aiApi.requestCompletion(
                 requestBody = StreamGptRequestBody(
+                    model = "free-model",
                     messages = listOf(
                         ChatMessage(
                             role = ChatMessage.Companion.Role.User,
@@ -56,15 +58,25 @@ class ViewModel {
                         is StreamGptEvent.Completion -> {
                             val now = timeSource.markNow()
 
+                            event.choices[0].delta.reasoning?.let { chunk ->
+                                reasoningContent += chunk
+                                tokens++
+                            }
+
                             event.choices[0].delta.content?.let { chunk ->
-                                response += chunk
+                                responseContent += chunk
                                 tokens++
                             }
 
                             _runs.update {
                                 it.toMutableList().apply {
                                     val tps = (tokens.toDouble() / (now - startMark).inWholeSeconds).toInt()
-                                    this[index] = Run(RunState.InProgress(tps), response, model = event.model)
+                                    this[index] = Run(
+                                        state = RunState.InProgress(tps),
+                                        reasoningContent = reasoningContent,
+                                        content = responseContent,
+                                        model = event.model
+                                    )
                                 }
                             }
                         }
@@ -84,6 +96,7 @@ class ViewModel {
 
     data class Run(
         val state: RunState,
+        val reasoningContent: String,
         val content: String,
         val model: String? = null
     )

src/main/kotlin/modifier/RunTitle.kt

@@ -46,7 +46,7 @@ fun RunTitle(
             }
         )
 
-        if (state is RunState.InProgress) {
+        if (state is RunState.InProgress && state.tokensPerSecond != Int.MAX_VALUE) {
             Text(
                 value = " ${state.tokensPerSecond}t/s",
                 color = Color(0, 120, 0)