feat: 适配器解析引擎，用 CSS 选择器从提交的 HTML 中自动提取数据进行校验

2026-07-01 01:25:16 +08:00
parent e27a5b59ed
commit e56f1d4b92
9 changed files with 439 additions and 0 deletions
--- a/par-api/src/main/java/com/par/api/controller/AdapterController.java
+++ b/par-api/src/main/java/com/par/api/controller/AdapterController.java
@@ -0,0 +1,87 @@
+package com.par.api.controller;
+
+import com.fasterxml.jackson.databind.ObjectMapper;
+import com.par.core.adapter.AdapterConfig;
+import com.par.core.adapter.AdapterEngine;
+import com.par.core.dto.ApiResponse;
+import com.par.core.entity.AccessResult;
+import com.par.core.entity.SiteConfig;
+import com.par.core.mapper.AccessResultMapper;
+import com.par.core.mapper.SiteConfigMapper;
+import com.par.core.storage.CachedStorageService;
+import lombok.RequiredArgsConstructor;
+import lombok.extern.slf4j.Slf4j;
+import org.springframework.web.bind.annotation.*;
+
+import java.util.Map;
+
+/**
+ * 适配器解析控制器
+ */
+@Slf4j
+@RestController
+@RequiredArgsConstructor
+public class AdapterController {
+
+    private final AdapterEngine adapterEngine;
+    private final SiteConfigMapper siteConfigMapper;
+    private final AccessResultMapper accessResultMapper;
+    private final CachedStorageService storageService;
+    private static final ObjectMapper objectMapper = new ObjectMapper();
+
+    /**
+     * 用指定适配配置解析指定访问结果
+     */
+    @PostMapping("/api/v1/admin/adapter/parse")
+    public ApiResponse<Map<String, Object>> parse(
+            @RequestBody Map<String, Object> payload) {
+
+        Long configId = toLong(payload.get("configId"));
+        Long accessResultId = toLong(payload.get("accessResultId"));
+
+        if (configId == null || accessResultId == null) {
+            return ApiResponse.error(400, "configId and accessResultId are required");
+        }
+
+        // 加载适配配置
+        SiteConfig siteConfig = siteConfigMapper.selectById(configId);
+        if (siteConfig == null) return ApiResponse.error(404, "配置不存在");
+
+        String configJson;
+        try { configJson = storageService.read(siteConfig.getStoragePath()); }
+        catch (Exception e) { return ApiResponse.error(500, "读取配置失败: " + e.getMessage()); }
+        if (configJson == null) return ApiResponse.error(404, "配置内容为空");
+
+        AdapterConfig adapterConfig;
+        try { adapterConfig = objectMapper.readValue(configJson, AdapterConfig.class); }
+        catch (Exception e) { return ApiResponse.error(400, "配置格式无效: " + e.getMessage()); }
+
+        // 加载访问结果 HTML
+        AccessResult accessResult = accessResultMapper.selectById(accessResultId);
+        if (accessResult == null) return ApiResponse.error(404, "访问结果不存在");
+
+        String html;
+        try { html = storageService.read(accessResult.getStoragePath()); }
+        catch (Exception e) { return ApiResponse.error(500, "读取访问结果失败: " + e.getMessage()); }
+        if (html == null) return ApiResponse.error(404, "访问结果内容为空");
+
+        // 执行解析
+        Map<String, Object> result;
+        try {
+            result = adapterEngine.extract(adapterConfig, html, accessResult.getUrl());
+        } catch (Exception e) {
+            log.error("Adapter parse error", e);
+            return ApiResponse.error(500, "解析异常: " + e.getMessage());
+        }
+
+        return ApiResponse.success(result);
+    }
+
+    private Long toLong(Object val) {
+        if (val instanceof Number n) return n.longValue();
+        if (val instanceof String s) {
+            try { return Long.parseLong(s); } catch (NumberFormatException ignored) {}
+        }
+        return null;
+    }
+}