feat: add OpenAI-compatible prompt token caching details and stream_options support

qorona_test.go

@@ -117,3 +117,78 @@ func TestModelsHandler(t *testing.T) {
     t.Fatalf("model not found in models list")
   }
 }
+
+func TestBuildUsage(t *testing.T) {
+  u := BuildUsage(100, 50, 80, 20)
+  if u.PromptTokens != 100 {
+    t.Fatalf("expected prompt_tokens 100, got %d", u.PromptTokens)
+  }
+  if u.CompletionTokens != 50 {
+    t.Fatalf("expected completion_tokens 50, got %d", u.CompletionTokens)
+  }
+  if u.TotalTokens != 150 {
+    t.Fatalf("expected total_tokens 150, got %d", u.TotalTokens)
+  }
+  if u.PromptTokensDetails == nil || u.PromptTokensDetails.CachedTokens != 80 {
+    t.Fatalf("expected cached_tokens 80, got %+v", u.PromptTokensDetails)
+  }
+  if u.CompletionTokensDetails == nil || u.CompletionTokensDetails.ReasoningTokens != 20 {
+    t.Fatalf("expected reasoning_tokens 20, got %+v", u.CompletionTokensDetails)
+  }
+
+  b, err := json.Marshal(u)
+  if err != nil {
+    t.Fatalf("failed to marshal usage: %v", err)
+  }
+  if !strings.Contains(string(b), `"cached_tokens":80`) {
+    t.Fatalf("expected json to contain cached_tokens: %s", string(b))
+  }
+}
+
+func TestWriteCompletionResponseUsage(t *testing.T) {
+  w := httptest.NewRecorder()
+  out := FinalOutput{
+    Content:          "Hello",
+    ReasoningContent: "Thinking...",
+    FinishReason:     "stop",
+    PromptTokens:     900,
+    CompletionTokens: 25,
+    CachedTokens:     850,
+    ReasoningTokens:  10,
+  }
+
+  WriteCompletionResponse(w, "chatcmpl-test", 1234567890, "qwen-3.8-max", out)
+  if w.Code != http.StatusOK {
+    t.Fatalf("expected 200 OK, got %d", w.Code)
+  }
+
+  var resp ChatCompletionResponse
+  if err := json.NewDecoder(w.Body).Decode(&resp); err != nil {
+    t.Fatalf("failed to decode response: %v", err)
+  }
+
+  if resp.Usage.PromptTokens != 900 {
+    t.Fatalf("expected prompt_tokens 900, got %d", resp.Usage.PromptTokens)
+  }
+  if resp.Usage.PromptTokensDetails == nil || resp.Usage.PromptTokensDetails.CachedTokens != 850 {
+    t.Fatalf("expected cached_tokens 850, got %+v", resp.Usage.PromptTokensDetails)
+  }
+}
+
+func TestStreamerUsage(t *testing.T) {
+  w := httptest.NewRecorder()
+  streamer := NewStreamer(w, nil, "chatcmpl-stream-test", 1234567890, "qwen-3.8-max")
+  usage := BuildUsage(900, 30, 850, 0)
+
+  streamer.EmitUsage(usage)
+  streamer.Done()
+
+  body := w.Body.String()
+  if !strings.Contains(body, `"cached_tokens":850`) {
+    t.Fatalf("expected streamed chunk to contain cached_tokens: %s", body)
+  }
+  if !strings.Contains(body, `data: [DONE]`) {
+    t.Fatalf("expected streamed body to contain [DONE]: %s", body)
+  }
+}
+