fix(plan): 修正Task状态值、Product状态值、删除死代码RULES块、修复PointsLog查询语法

docs/superpowers/plans/2026-08-21-user-profile-recommendation.md

@@ -559,9 +559,9 @@ public class ProfileComputeServiceImpl implements ProfileComputeService {
         LambdaQueryWrapper<PointsLog> pqw = new LambdaQueryWrapper<>();
         pqw.eq(PointsLog::getFamilyMemberId, memberId)
            .ge(PointsLog::getCreatedAt, java.sql.Timestamp.valueOf(sevenDaysAgo.atStartOfDay()))
-           .eq(PointsLog::getAmount, e -> true); // positive amounts only
-        long pointsEarned = pointsLogMapper.selectList(pqw).stream()
-            .filter(p -> p.getAmount() != null && p.getAmount() > 0)
+           .gt(PointsLog::getAmount, 0);
+        List<PointsLog> pointsList = pointsLogMapper.selectList(pqw);
+        long pointsEarned = pointsList.stream()
             .mapToInt(PointsLog::getAmount).sum();
         m.put("points_velocity_7d", pointsEarned);
         return m;
@@ -897,30 +897,6 @@ import java.util.*;
 @Service
 public class RecommendServiceImpl implements RecommendService {
 
-    // 规则映射：标签code -> 匹配的Content查询条件
-    private static final Map<String, RecommendRule> RULES = new LinkedHashMap<>();
-    static {
-        RULES.put("sleep_deficit", new RecommendRule("sleep", 0.8, 
-            q -> q.like(Article::getTags, "睡眠").or().like(Article::getTags, "助眠"),
-            q -> q.eq(Activity::getDimensionCode, "body").or().like(Activity::getDescription, "睡眠")));
-        RULES.put("low_activity", new RecommendRule("action", 0.7,
-            q -> q.like(Task::getCategory, "运动").or().eq(Task::getDimensionCode, "body"),
-            null));
-        RULES.put("high_stress", new RecommendRule("mind", 0.85,
-            q -> q.like(Article::getTags, "压力").or().like(Article::getTags, "情绪管理"),
-            q -> q.eq(Activity::getDimensionCode, "mind").or().like(Activity::getTitle, "冥想")));
-        RULES.put("low_mood", new RecommendRule("mind", 0.9,
-            q -> q.like(Article::getTags, "情绪").or().like(Article::getTags, "心理"),
-            q -> q.eq(Activity::getDimensionCode, "mind")));
-        RULES.put("attention_weak", new RecommendRule("wisdom", 0.85,
-            q -> q.like(Article::getTags, "注意力").or().like(Article::getTags, "专注力"),
-            q -> q.eq(Activity::getDimensionCode, "wisdom").or().like(Activity::getTitle, "注意力")),
-            null); // task已包含
-        RULES.put("task_avoidance", new RecommendRule("action", 0.75,
-            q -> q.like(Task::getCategory, "入门").or().eq(Task::getDifficultyLevel, 1),
-            null));
-    }
-
     @Autowired private ProfileSnapshotMapper snapshotMapper;
     @Autowired private ArticleMapper articleMapper;
     @Autowired private ActivityMapper activityMapper;
@@ -985,8 +961,8 @@ public class RecommendServiceImpl implements RecommendService {
         for (String tag : tags) {
             LambdaQueryWrapper<Task> qw = new LambdaQueryWrapper<>();
             qw.like(Task::getCategory, tag).or().like(Task::getDescription, tag)
-               .eq(Task::getStatus, "active")
-               .orderByAsc(Task::getSortOrder)
+               .eq(Task::getStatus, "pending")
+               .
                .last("LIMIT 5");
             List<Task> tasks = taskMapper.selectList(qw);
             for (Task t : tasks) {
@@ -1052,7 +1028,7 @@ public class RecommendServiceImpl implements RecommendService {
     private List<Map<String, Object>> queryProducts(Set<String> tags) {
         List<Map<String, Object>> list = new ArrayList<>();
         LambdaQueryWrapper<Product> qw = new LambdaQueryWrapper<>();
-        qw.eq(Product::getStatus, "on_sale")
+        qw.eq(Product::getStatus, "on_shelf")
            .isNotNull(Product::getGrowthCategory)
            .last("LIMIT 8");
         List<Product> products = productMapper.selectList(qw);
@@ -1482,6 +1458,6 @@ git push origin cfclub
 1. **异步触发画像计算**：使用 `new Thread(...)` 避免阻塞接口响应，生产环境建议替换为 Spring 的 `@Async` 或消息队列
 2. **画像计算幂等性**：`computeAndSave` 内部用 upsert，重复调用不会产生脏数据
 3. **历史数据清理**：`cleanOldHistory` 只在写入新记录时顺带清理，不单独定时任务
-4. **规则可扩展**：`RULES` Map 集中在 `RecommendServiceImpl` 顶部，后续添加新规则只需在此增改
+4. **规则可扩展**：推荐匹配逻辑集中在 `RecommendServiceImpl.getRecommendations()` 中，后续添加新规则只需在此增改 if 分支
 5. **前端 no `?.` 语法**：小程序禁止可选链，用 `&&` 替代
 6. **日期格式**：统一使用 `yyyy-MM-dd` 或 ISO 8601，禁止 `toLocaleString()`