feat: add streaming support to LlamaCppClient

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

src/models/local/llamacpp.ts

@@ -1,4 +1,4 @@
-import type { ChatRequest, ChatResponse, ModelClient } from '../types.js';
+import type { ChatRequest, ChatResponse, ChatStreamEvent, ModelClient } from '../types.js';
 
 export interface LlamaCppClientConfig {
   endpoint: string;
@@ -15,6 +15,11 @@ interface LlamaCppResponse {
   usage: { prompt_tokens: number; completion_tokens: number };
 }
 
+interface LlamaCppStreamChunk {
+  choices: Array<{ delta?: { content?: string } }>;
+  usage?: { prompt_tokens: number; completion_tokens: number };
+}
+
 export class LlamaCppClient implements ModelClient {
   private endpoint: string;
   private authToken?: string;
@@ -68,4 +73,91 @@ export class LlamaCppClient implements ModelClient {
       },
     };
   }
+
+  async *chatStream(request: ChatRequest): AsyncIterable<ChatStreamEvent> {
+    const messages: LlamaCppMessage[] = [];
+
+    if (request.system) {
+      messages.push({ role: 'system', content: request.system });
+    }
+
+    for (const msg of request.messages) {
+      messages.push({ role: msg.role, content: msg.content });
+    }
+
+    const headers: Record<string, string> = {
+      'Content-Type': 'application/json',
+    };
+
+    if (this.authToken) {
+      headers['Authorization'] = `Bearer ${this.authToken}`;
+    }
+
+    try {
+      const response = await fetch(`${this.endpoint}/v1/chat/completions`, {
+        method: 'POST',
+        headers,
+        body: JSON.stringify({
+          messages,
+          max_tokens: request.maxTokens ?? 2048,
+          stream: true,
+        }),
+      });
+
+      if (!response.ok) {
+        const text = await response.text();
+        throw new Error(`llama-server error (${response.status}): ${text}`);
+      }
+
+      if (!response.body) {
+        throw new Error('No response body for streaming');
+      }
+
+      const reader = response.body.getReader();
+      const decoder = new TextDecoder();
+      let buffer = '';
+      let usage = { inputTokens: 0, outputTokens: 0 };
+
+      while (true) {
+        const { done, value } = await reader.read();
+        if (done) break;
+
+        buffer += decoder.decode(value, { stream: true });
+        const lines = buffer.split('\n');
+        buffer = lines.pop() ?? '';
+
+        for (const line of lines) {
+          const trimmed = line.trim();
+          if (!trimmed || !trimmed.startsWith('data: ')) continue;
+
+          const data = trimmed.slice(6);
+          if (data === '[DONE]') continue;
+
+          try {
+            const chunk = JSON.parse(data) as LlamaCppStreamChunk;
+
+            if (chunk.choices[0]?.delta?.content) {
+              yield { type: 'content', content: chunk.choices[0].delta.content };
+            }
+
+            if (chunk.usage) {
+              usage = {
+                inputTokens: chunk.usage.prompt_tokens,
+                outputTokens: chunk.usage.completion_tokens,
+              };
+            }
+          } catch {
+            // Skip malformed JSON
+          }
+        }
+      }
+
+      yield { type: 'done', usage };
+    } catch (error) {
+      yield {
+        type: 'error',
+        error: error instanceof Error ? error : new Error(String(error)),
+      };
+    }
+  }
 }