feat: refactor AI engine logic, remove cross-engine fallback, add retry with backoff, and bump version to 2.80.18

2026-05-08 01:24:12 +09:00
parent 6894152892
commit d451a082dd
10 changed files with 104 additions and 75 deletions
@@ -1,5 +1,5 @@
 {
  "result": "Final report with inconsistencies. This should be long enough to pass validation.",
-  "createdAt": 1778137049532,
+  "createdAt": 1778170647495,
  "modelVersion": "unknown"
 }
@@ -1,5 +1,5 @@
 {
  "result": "[CONFLICT WARNING] 성능이 200% 증가했습니다. vs 그러나 동시에 50% 감소했습니다. 최적화와 성능 저하가 동시에 발견됨.",
-  "createdAt": 1778137049529,
+  "createdAt": 1778170647483,
  "modelVersion": "unknown"
 }
@@ -1,5 +1,5 @@
 {
  "result": "Detailed Execution Plan: 1. Research 2. Analyze 3. Write report with high quality.",
-  "createdAt": 1778137049527,
+  "createdAt": 1778170647479,
  "modelVersion": "unknown"
 }
@@ -1,5 +1,5 @@
 {
-  "result": "---\nid: stress_conflict_1778137049510\ndate: 2026-05-07T06:57:29.533Z\ntype: knowledge_artifact\nstandard: P-Reinforce v3.0\ntags: [automated, connect_ai, brain_sync]\n---\n\n## 📌 Brief Summary\nFinal report with inconsistencies. This should be long enough to pass validation.\n\nFinal report with inconsistencies. This should be long enough to pass validation.\n\n---\n## 💡 Astra의 선제적 제안 (Proactive Next Actions)\nFinal report with inconsistencies. This should be long enough to pass validation.\n---\n## 🛡️ Reliability & Audit Summary\n> [!NOTE]\n> 이 문서는 ConnectAI의 **Intelligent Resilience** 엔진에 의해 검증 및 정제되었습니다.\n\n| Metric | Value | Status |\n| :--- | :--- | :--- |\n| **Conflict Risk** | `60/100` | ⚠️ Medium |\n| **Fallbacks Used** | `0` | ✅ None |\n| **Auto Retries** | `0` | ✅ Stable |\n| **Deduplication** | `0` | Standard |\n| **Processing Time** | `0.0s` | ✅ Fast |\n\n### 🔍 Decision Audit Trail\n- **[PLANNER]** 전략 수립 중... (16ms)\n- **[RESEARCHER]** 핵심 정보 수집 및 분석 중... (2ms)\n- **[WRITER]** 최종 리포트 작성 및 편집 중... (3ms)\n",
-  "createdAt": 1778137049533,
+  "result": "---\nid: stress_conflict_1778170647465\ndate: 2026-05-07T16:17:27.498Z\ntype: knowledge_artifact\nstandard: P-Reinforce v3.0\ntags: [automated, connect_ai, brain_sync]\n---\n\n## 📌 Brief Summary\nFinal report with inconsistencies. This should be long enough to pass validation.\n\nFinal report with inconsistencies. This should be long enough to pass validation.\n\n---\n## 💡 Astra의 선제적 제안 (Proactive Next Actions)\nFinal report with inconsistencies. This should be long enough to pass validation.\n---\n## 🛡️ Reliability & Audit Summary\n> [!NOTE]\n> 이 문서는 ConnectAI의 **Intelligent Resilience** 엔진에 의해 검증 및 정제되었습니다.\n\n| Metric | Value | Status |\n| :--- | :--- | :--- |\n| **Conflict Risk** | `60/100` | ⚠️ Medium |\n| **Fallbacks Used** | `0` | ✅ None |\n| **Auto Retries** | `0` | ✅ Stable |\n| **Deduplication** | `0` | Standard |\n| **Processing Time** | `0.0s` | ✅ Fast |\n\n### 🔍 Decision Audit Trail\n- **[PLANNER]** 전략 수립 중... (9ms)\n- **[RESEARCHER]** 핵심 정보 수집 및 분석 중... (5ms)\n- **[WRITER]** 최종 리포트 작성 및 편집 중... (12ms)\n",
+  "createdAt": 1778170647499,
  "modelVersion": "unknown"
 }
@@ -1,8 +1,8 @@
 {
-  "missionId": "stress_conflict_1778137049510",
+  "missionId": "stress_conflict_1778170647465",
  "status": "completed",
-  "startTime": "2026-05-07T06:57:29.510Z",
-  "totalElapsedMs": 24,
+  "startTime": "2026-05-07T16:17:27.465Z",
+  "totalElapsedMs": 36,
  "results": {
    "planner": "Detailed Execution Plan: 1. Research 2. Analyze 3. Write report with high quality.",
    "researcher": "[CONFLICT WARNING] 성능이 200% 증가했습니다. vs 그러나 동시에 50% 감소했습니다. 최적화와 성능 저하가 동시에 발견됨.",
@@ -16,30 +16,30 @@
    {
      "from": "idle",
      "to": "planner",
-      "durationMs": 16,
+      "durationMs": 9,
      "message": "전략 수립 중...",
-      "ts": "2026-05-07T06:57:29.526Z"
+      "ts": "2026-05-07T16:17:27.474Z"
    },
    {
      "from": "planner",
      "to": "researcher",
-      "durationMs": 2,
+      "durationMs": 5,
      "message": "핵심 정보 수집 및 분석 중...",
-      "ts": "2026-05-07T06:57:29.528Z"
+      "ts": "2026-05-07T16:17:27.479Z"
    },
    {
      "from": "researcher",
      "to": "writer",
-      "durationMs": 3,
+      "durationMs": 12,
      "message": "최종 리포트 작성 및 편집 중...",
-      "ts": "2026-05-07T06:57:29.531Z"
+      "ts": "2026-05-07T16:17:27.491Z"
    },
    {
      "from": "writer",
      "to": "completed",
-      "durationMs": 3,
+      "durationMs": 10,
      "message": "미션 완료",
-      "ts": "2026-05-07T06:57:29.534Z"
+      "ts": "2026-05-07T16:17:27.501Z"
    }
  ],
  "resilienceMetrics": {
@@ -1,12 +1,12 @@
 {
  "name": "astra",
-  "version": "2.80.5",
+  "version": "2.80.18",
  "lockfileVersion": 3,
  "requires": true,
  "packages": {
    "": {
      "name": "astra",
-      "version": "2.80.5",
+      "version": "2.80.18",
      "license": "MIT",
      "dependencies": {
        "marked": "^18.0.2",
@@ -58,7 +58,6 @@
      "integrity": "sha512-CGOfOJqWjg2qW/Mb6zNsDm+u5vFQ8DxXfbM09z69p5Z6+mE1ikP2jUXw+j42Pf1XTYED2Rni5f95npYeuwMDQA==",
      "dev": true,
      "license": "MIT",
-      "peer": true,
      "dependencies": {
        "@babel/code-frame": "^7.29.0",
        "@babel/generator": "^7.29.0",
@@ -1943,7 +1942,6 @@
        }
      ],
      "license": "MIT",
-      "peer": true,
      "dependencies": {
        "baseline-browser-mapping": "^2.10.12",
        "caniuse-lite": "^1.0.30001782",
@@ -2841,7 +2839,6 @@
      "integrity": "sha512-NIy3oAFp9shda19hy4HK0HRTWKtPJmGdnvywu01nOqNC2vZg+Z+fvJDxpMQA88eb2I9EcafcdjYgsDthnYTvGw==",
      "dev": true,
      "license": "MIT",
-      "peer": true,
      "dependencies": {
        "@jest/core": "^29.7.0",
        "@jest/types": "^29.6.3",
@@ -4398,7 +4395,6 @@
      "integrity": "sha512-jl1vZzPDinLr9eUt3J/t7V6FgNEw9QjvBPdysz9KfQDD41fQrC2Y4vKQdiaUpFT4bXlb1RHhLpp8wtm6M5TgSw==",
      "dev": true,
      "license": "Apache-2.0",
-      "peer": true,
      "bin": {
        "tsc": "bin/tsc",
        "tsserver": "bin/tsserver"
@@ -2,7 +2,7 @@
  "name": "astra",
  "displayName": "Astra",
  "description": "The personal intelligence layer for Antigravity and VS Code. A private cognitive partner for deep project context, memory, and proactive strategic decision-making.",
-  "version": "2.80.17",
+  "version": "2.80.18",
  "publisher": "g1nation",
  "license": "MIT",
  "icon": "assets/icon.png",
@@ -1942,35 +1942,37 @@ export class AgentExecutor {
        temperature: number;
    }): Promise<{ response: Response; engine: 'lmstudio' | 'ollama'; apiUrl: string }> {
        const { baseUrl, modelName, reqMessages, temperature } = params;
-        const primaryEngine = resolveEngine(baseUrl);
-        const engines = primaryEngine === 'lmstudio' ? ['lmstudio', 'ollama'] as const : ['ollama', 'lmstudio'] as const;
+        const engine = resolveEngine(baseUrl);  // 사용자가 설정한 엔진만 사용
+        const apiUrl = buildApiUrl(baseUrl, engine, 'chat');
+        const messageVariants = this.buildEngineMessageVariants(reqMessages, engine);
+        const modelCandidates = this.buildModelCandidates(modelName, engine);
        let lastError: Error | null = null;

-        for (const engine of engines) {
-            const apiUrl = buildApiUrl(baseUrl, engine, 'chat');
-            const messageVariants = this.buildEngineMessageVariants(reqMessages, engine);
-            const modelCandidates = this.buildModelCandidates(modelName, engine);
-
-            for (const candidateModel of modelCandidates) {
-                for (const variant of messageVariants) {
-                    const streamBody = {
-                        model: candidateModel,
-                        messages: variant.messages,
-                        stream: true,
-                        ...(engine === 'lmstudio'
-                            ? { max_tokens: 4096, temperature }
-                            : { options: { num_ctx: 32768, num_predict: 4096, temperature } }),
-                    };
+        // 같은 엔진 내에서만 model candidate / message variant retry
+        for (const candidateModel of modelCandidates) {
+            for (const variant of messageVariants) {
+                const streamBody = {
+                    model: candidateModel,
+                    messages: variant.messages,
+                    stream: true,
+                    ...(engine === 'lmstudio'
+                        ? { max_tokens: 4096, temperature }
+                        : { options: { num_ctx: 32768, num_predict: 4096, temperature } }),
+                };

+                // 일시적 네트워크 오류용 retry (최대 2회, 지수 backoff)
+                const MAX_RETRIES = 2;
+                for (let attempt = 0; attempt <= MAX_RETRIES; attempt++) {
                    try {
+                        if (attempt > 0) {
+                            const delay = 500 * Math.pow(2, attempt - 1);  // 500ms, 1000ms
+                            await new Promise(r => setTimeout(r, delay));
+                            logInfo('AI streaming request retry.', { engine, attempt, model: candidateModel });
+                        }
                        logInfo('AI streaming request started.', {
-                            engine,
-                            apiUrl,
-                            model: candidateModel,
-                            variant: variant.name,
-                            messageCount: variant.messages.length,
-                            roles: variant.messages.map(message => message.role),
-                            firstUserPreview: summarizeText(String(variant.messages.find(message => message.role === 'user')?.content || ''), 300)
+                            engine, apiUrl, model: candidateModel,
+                            variant: variant.name, messageCount: variant.messages.length,
+                            attempt
                        });
                        const response = await fetch(apiUrl, {
                            method: 'POST',
@@ -1988,7 +1990,12 @@ export class AgentExecutor {
                        if (!response.ok) {
                            const errText = await response.text();
                            lastError = new Error(`AI Engine error (${engine}/${variant.name}): ${response.status} - ${summarizeText(errText, 300)}`);
-                            logError('AI streaming request returned non-OK status.', { engine, variant: variant.name, apiUrl, status: response.status, body: summarizeText(errText, 500) });
+                            logError('AI streaming request returned non-OK status.', {
+                                engine, variant: variant.name, apiUrl,
+                                status: response.status, body: summarizeText(errText, 500)
+                            });
+                            // 4xx는 재시도해도 의미없음. 5xx만 재시도.
+                            if (response.status >= 400 && response.status < 500) break;
                            continue;
                        }

@@ -1996,13 +2003,26 @@ export class AgentExecutor {
                        return { response, engine, apiUrl };
                    } catch (error: any) {
                        lastError = error instanceof Error ? error : new Error(String(error));
-                        logError('AI streaming request failed.', { engine, variant: variant.name, apiUrl, model: candidateModel, error: lastError.message });
+                        // AbortError는 사용자가 취소한 것이므로 retry 금지
+                        if (lastError.name === 'AbortError') {
+                            throw lastError;
+                        }
+                        logError('AI streaming request failed.', {
+                            engine, variant: variant.name, apiUrl, model: candidateModel,
+                            attempt, error: lastError.message
+                        });
                    }
                }
            }
        }

-        throw lastError || new Error('Unable to connect to AI engine.');
+        // 명확한 에러 메시지: 어느 엔진이 실패했는지 사용자에게 알림
+        const engineLabel = engine === 'lmstudio' ? 'LM Studio' : 'Ollama';
+        throw new Error(
+            `${engineLabel} 엔진에 연결할 수 없습니다. ` +
+            `${engineLabel}가 실행 중이고 모델 '${modelName}'이 로드되어 있는지 확인하세요. ` +
+            `(원인: ${lastError?.message || 'unknown'})`
+        );
    }

    private normalizeMessages(messages: ChatMessage[]) {
@@ -90,6 +90,14 @@ export function deactivate() {
 }

 async function runInitialSetup(context: vscode.ExtensionContext) {
+    // 이미 사용자가 URL을 설정했다면 자동 감지를 스킵
+    const existingUrl = vscode.workspace.getConfiguration('g1nation').get<string>('ollamaUrl');
+    if (existingUrl && existingUrl.trim()) {
+        context.globalState.update('setupComplete', true);
+        logInfo('Initial setup skipped: ollamaUrl already configured.', { existingUrl });
+        return;
+    }
+
    try {
        let engineName = '';
        let modelName = '';
@@ -51,6 +51,7 @@ export class SidebarChatProvider implements vscode.WebviewViewProvider, BridgeIn
    private _currentSessionBrainId: string | null = null;
    private _currentNegativePrompt: string = '';
    private readonly _chronicle = new ProjectChronicleManager();
+    private _modelDiscoveryInFlight = false;

    constructor(
        private readonly _extensionUri: vscode.Uri,
@@ -75,13 +76,18 @@ export class SidebarChatProvider implements vscode.WebviewViewProvider, BridgeIn
        };

        // [State Persistence Fix] 사이드바가 다시 보여질 때 세팅값 자동 복원
+        let _lastVisibilityRefresh = 0;
        webviewView.onDidChangeVisibility(() => {
-            if (webviewView.visible) {
-                logInfo('Sidebar became visible, restoring state...');
-                void this._sendModels();
-                void this._sendBrainProfiles();
-                void this._sendAgentsList();
-            }
+            if (!webviewView.visible) return;
+            const now = Date.now();
+            // 5초 이내에 이미 갱신했으면 건너뜀
+            if (now - _lastVisibilityRefresh < 5000) return;
+            _lastVisibilityRefresh = now;
+            
+            logInfo('Sidebar became visible, restoring state...');
+            void this._sendModels();
+            void this._sendBrainProfiles();
+            void this._sendAgentsList();
        });

        webviewView.webview.html = this._getHtml(webviewView.webview);
@@ -1950,26 +1956,26 @@ export class SidebarChatProvider implements vscode.WebviewViewProvider, BridgeIn

    private async _sendModels() {
        if (!this._view) return;
+        if (this._modelDiscoveryInFlight) {
+            logInfo('Model discovery already in progress, skipping.');
+            return;
+        }
+        this._modelDiscoveryInFlight = true;
        try {
            const config = getConfig();
            const url = config.ollamaUrl;
            let defaultModel = config.defaultModel;
            let models: string[] = [];

-            const primaryEngine = resolveEngine(url);
-            const engines = primaryEngine === 'lmstudio' ? ['lmstudio', 'ollama'] as const : ['ollama', 'lmstudio'] as const;
-
-            for (const engine of engines) {
-                const modelsUrl = buildApiUrl(url, engine, 'models');
-                try {
-                    logInfo('Model discovery started.', { engine, modelsUrl });
-                    const res = await fetch(modelsUrl, { signal: AbortSignal.timeout(5000) });
-                    const rawText = await res.text();
-                    if (!res.ok) {
-                        logError('Model discovery returned non-OK status.', { engine, modelsUrl, status: res.status, body: summarizeText(rawText, 300) });
-                        continue;
-                    }
-
+            const engine = resolveEngine(url);  // 단일 엔진만
+            const modelsUrl = buildApiUrl(url, engine, 'models');
+            try {
+                logInfo('Model discovery started.', { engine, modelsUrl });
+                const res = await fetch(modelsUrl, { signal: AbortSignal.timeout(5000) });
+                const rawText = await res.text();
+                if (!res.ok) {
+                    logError('Model discovery returned non-OK status.', { engine, modelsUrl, status: res.status, body: summarizeText(rawText, 300) });
+                } else {
                    const data = rawText ? JSON.parse(rawText) as any : {};
                    models = engine === 'lmstudio'
                        ? (data.data || []).map((m: any) => m.id)
@@ -1977,11 +1983,10 @@ export class SidebarChatProvider implements vscode.WebviewViewProvider, BridgeIn

                    if (models.length > 0) {
                        logInfo('Model discovery succeeded.', { engine, count: models.length, modelsPreview: models.slice(0, 5) });
-                        break;
                    }
-                } catch (e: any) {
-                    logError('Model discovery failed.', { engine, modelsUrl, error: e?.message || String(e) });
                }
+            } catch (e: any) {
+                logError('Model discovery failed.', { engine, modelsUrl, error: e?.message || String(e) });
            }

            if (models.length === 0) {
@@ -2017,8 +2022,8 @@ export class SidebarChatProvider implements vscode.WebviewViewProvider, BridgeIn
            this._view.webview.postMessage({ type: 'modelsList', value: { models, selected: defaultModel } });
        } catch (err) {
            logError('Model list update failed.', err);
-            const fallbackModel = getConfig().defaultModel;
-            this._view.webview.postMessage({ type: 'modelsList', value: { models: fallbackModel ? [fallbackModel] : [], selected: fallbackModel } });
+        } finally {
+            this._modelDiscoveryInFlight = false;
        }
    }