added chat streaming

packages/orchestration-service/src/chat/index.js

@@ -93,4 +93,60 @@ async function chat(externalId, userMessage, options = {}) {
     };
 }
 
-module.exports = { chat };
+async function chatStream(externalId, userMessage, onChunk, options = {} ) {
+    // 1. Resolve or create session
+    let session = await memory.getSessionByExternalId(externalId);
+    if (!session) session = await memory.createSession(externalId);
+
+    // 2. Context assembly
+    const recentEpisodes = await memory.getRecentEpisodes(session.id, RECENT_EPISODE_LIMIT);
+    const recentIds = new Set(recentEpisodes.map(e => e.id));
+    const semanticEpisodes = await getSemanticEpisodes(userMessage, session.id, recentIds)
+
+    // 3. Assemble Prompt
+    const prompt = buildPrompt(recentEpisodes, semanticEpisodes, userMessage);
+
+    // 4. Open stream to inference service
+    const res = await inference.completeStream(prompt, options);
+
+    let fullText = '';
+    let model = '';
+    let tokenCount = 0;
+
+    // 5. Parse SSE chunks
+    for await (const chunk of res.body){
+        const lines = chunk.toString().split('\n');
+
+        for (const line of lines) {
+            if (!line.startsWith('data: ')) continue;
+            const raw = line.slice(6).trim();
+            if (raw === '[DONE]') continue //stream closed sentinel
+
+            try {
+                const data = JSON.parse(raw);
+                if (data.model) model = data.model
+
+                if (data.response) {
+                    fullText += data.response;
+                    onChunk(data.response);
+                }
+
+                if (data.done && data.eval_count !== undefined) {
+                    tokenCount = (data.eval_count || 0) + (data.prompt_eval_count || 0)
+                }
+            } catch {
+                //partial chunk
+                //skip and wait for next
+            }
+
+        }
+    }
+
+    // 6. Write Complete episode to memory
+    memory.createEpisode(session.id, userMessage, fullText, tokenCount)
+        .catch(err => console.error('[orchestration] Failed to save streamed episode:', err.message))
+    
+    return {model, tokenCount};        
+}
+
+module.exports = { chat, chatStream };