@nomad-e/bluma-cli 0.1.76 → 0.1.78

package/dist/main.js

@@ -14967,6 +14967,41 @@ function buildFactorHeaders(ctx) {
     "X-Company-Name": encodeHeader(ctx.companyName)
   };
 }
+var THINKING_TOKEN_BUDGET_BY_EFFORT = {
+  low: 256,
+  medium: 1024,
+  high: 2048
+};
+function getThinkingTokenBudgetForEffort(effort) {
+  if (!effort) return void 0;
+  return THINKING_TOKEN_BUDGET_BY_EFFORT[effort];
+}
+function buildVllmReasoningPayload(effort) {
+  if (!effort) return void 0;
+  return {
+    reasoning: { effort },
+    extra_body: {
+      thinking_token_budget: getThinkingTokenBudgetForEffort(effort)
+    }
+  };
+}
+function buildChatCompletionRequestBody(params, runtimeConfig = getRuntimeConfig(), stream = false) {
+  const tools = params.tools;
+  const hasTools = Array.isArray(tools) && tools.length > 0;
+  const effort = params.reasoning?.effort ?? runtimeConfig.reasoningEffort;
+  const reasoningPayload = buildVllmReasoningPayload(effort);
+  return {
+    model: params.model || runtimeConfig.model || "auto",
+    messages: params.messages,
+    tools: hasTools ? tools : void 0,
+    tool_choice: hasTools ? "auto" : void 0,
+    parallel_tool_calls: params.parallel_tool_calls ?? false,
+    temperature: params.temperature ?? 0,
+    ...reasoningPayload ?? {},
+    max_tokens: params.max_tokens,
+    ...stream ? { stream: true } : {}
+  };
+}
 function applyDeltaToolCallsToAccumulator(toolCallsAccumulator, deltaToolCalls) {
   if (!deltaToolCalls?.length) {
     return false;
@@ -15044,20 +15079,9 @@ var LLMService = class {
     if (!params.userContext) {
       throw new Error("LLMService.chatCompletion: userContext \xE9 obrigat\xF3rio");
     }
-    const tools = params.tools;
-    const hasTools = Array.isArray(tools) && tools.length > 0;
     const runtimeConfig = getRuntimeConfig();
     const resp = await this.client.chat.completions.create(
-      {
-        model: params.model || runtimeConfig.model || "auto",
-        messages: params.messages,
-        tools: hasTools ? tools : void 0,
-        tool_choice: hasTools ? "auto" : void 0,
-        parallel_tool_calls: params.parallel_tool_calls ?? false,
-        temperature: params.temperature ?? 0,
-        reasoning: params.reasoning ?? (runtimeConfig.reasoningEffort ? { effort: runtimeConfig.reasoningEffort } : void 0),
-        max_tokens: params.max_tokens
-      },
+      buildChatCompletionRequestBody(params, runtimeConfig, false),
       { headers: this.requestHeaders(params.userContext) }
     );
     return resp;
@@ -15069,21 +15093,9 @@ var LLMService = class {
     if (!params.userContext) {
       throw new Error("LLMService.chatCompletionStream: userContext \xE9 obrigat\xF3rio");
     }
-    const tools = params.tools;
-    const hasTools = Array.isArray(tools) && tools.length > 0;
     const runtimeConfig = getRuntimeConfig();
     const stream = await this.client.chat.completions.create(
-      {
-        model: params.model || runtimeConfig.model || "auto",
-        messages: params.messages,
-        tools: hasTools ? tools : void 0,
-        tool_choice: hasTools ? "auto" : void 0,
-        parallel_tool_calls: params.parallel_tool_calls ?? false,
-        temperature: params.temperature ?? 0,
-        reasoning: params.reasoning ?? (runtimeConfig.reasoningEffort ? { effort: runtimeConfig.reasoningEffort } : void 0),
-        max_tokens: params.max_tokens,
-        stream: true
-      },
+      buildChatCompletionRequestBody(params, runtimeConfig, true),
       { headers: this.requestHeaders(params.userContext) }
     );
     const toolCallsAccumulator = /* @__PURE__ */ new Map();

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@nomad-e/bluma-cli",
-	"version": "0.1.76",
+	"version": "0.1.78",
 	"description": "BluMa independent agent for automation and advanced software engineering.",
 	"author": "Alex Fonseca",
 	"license": "Apache-2.0",