autosnippet
Version:
Extract code patterns into a knowledge base for AI coding assistants
49 lines (48 loc) • 1.57 kB
JavaScript
/**
* AIRecallStrategy — 将 SignalCollector 的 AI 分析结果包装为 RecallStrategy 接口
*
* 读取 SignalCollector 的缓存快照中的 pendingSuggestions，
* 转换为标准 RecommendationCandidate。
*
* 依赖 AI Provider，不可用时返回空数组。
*/
export class AIRecallStrategy {
name = 'ai';
type = 'ai';
#signalCollector;
constructor(signalCollector) {
this.#signalCollector = signalCollector;
}
async recall(context) {
if (!this.#signalCollector) {
return [];
}
const snapshot = this.#signalCollector.getSnapshot();
const pending = snapshot.pendingSuggestions ?? [];
const existingSet = context.existingSkills ?? new Set();
return pending
.filter((s) => !existingSet.has(s.name))
.map((s) => ({
name: s.name || 'unknown',
description: s.description || '',
rationale: s.rationale || '',
source: 'ai:signal_collector',
priority: s.priority || 'medium',
signals: s,
body: s.body,
}));
}
isAvailable(context) {
if (!this.#signalCollector) {
return false;
}
const mode = this.#signalCollector.getMode();
// AI 策略只在 suggest/auto 模式下可用
return mode === 'suggest' || mode === 'auto';
}
/** 更新 SignalCollector 引用 (用于延迟注入) */
setSignalCollector(sc) {
this.#signalCollector = sc;
}
}
export default AIRecallStrategy;