xxl-job+工厂模式+token实现拉取（三）

目录
- 前言
- 整体设计
- - xxl-job说明
  - 表设计
  - 工厂模式设计
- 码上有戏
- - 核心代码
  - 测试
  - 简单说明
  - 源码地址

CREATE TABLE `sys_task_datetime` (`id` int(11) NOT NULL AUTO_INCREMENT COMMENT '主键id',`page_index` int(11) DEFAULT NULL COMMENT '页码',`api_name` varchar(50) DEFAULT NULL COMMENT '接口名称',`execute_time` datetime DEFAULT NULL COMMENT '创建时间',`after_min` int(11) DEFAULT NULL COMMENT '接口提前量',PRIMARY KEY (`id`) USING BTREE
) ENGINE=InnoDB AUTO_INCREMENT=12 DEFAULT CHARSET=utf8mb4 ROW_FORMAT=DYNAMIC;-- ----------------------------
-- Records of sys_task_datetime
-- ----------------------------
INSERT INTO `sys_task_datetime` VALUES ('1', '1', 'token', '2021-04-16 14:04:13', null);
INSERT INTO `sys_task_datetime` VALUES ('2', '1', 'leesin', '2021-04-14 11:17:03', null);
INSERT INTO `sys_task_datetime` VALUES ('3', '1', 'yurnero', '2020-08-18 16:07:01', null);

这里配置了三个接口，分别对应之前的token，盲僧接口，剑圣接口推送

工厂模式设计

以上是简化工厂模式uml图，事实上我们只需要用一个枚举值去控制每个具体xxApiRequest服务，在系统初始化向工厂里注入改组件即可，而实际调用时只需一行代码创建即可

 AbstractApiRequest apiRequest = factory.create(TaskEnumUtils.taskDatetimeType.LEESIN.key);apiRequest.excute();

事实上就可以获得具体子类，这也是oop充满魅力的地方

码上有戏

如图，为项目代码结构

核心代码

使用工厂模式就是为了防止模板发生变化，从而扩展多个产品线或者一个产品线下多个产品
首先是系统
注入服务

@Configuration
public class SystemConfig {@Resourceprivate LeesinApiRequest leesinApiRequest;@Resourceprivate YurneroApiRequest yurneroApiRequest;@Beanpublic ApiRequestFactory createFactory() {ApiRequestFactory factory = new ApiRequestFactory();Map<String, AbstractApiRequest> serviceMap = new HashMap<>(6);serviceMap.put(TaskEnumUtils.taskDatetimeType.LEESIN.key, leesinApiRequest);serviceMap.put(TaskEnumUtils.taskDatetimeType.YURNERO.key, yurneroApiRequest);factory.setServiceMap(serviceMap);return factory;}
}

如果有需要，可以扩展注入服务到工厂里的serverMap中

枚举核心配置

public class TaskEnumUtils {public enum taskDatetimeType {TOKEN("token"), LEESIN("leesin"), YURNERO("yurnero");public String key;private taskDatetimeType(String key) {this.key = key;}}public static EnumMap taskDatetimeTypeEnum = new EnumMap(taskDatetimeType.class) {{put(taskDatetimeType.TOKEN, "token");put(taskDatetimeType.LEESIN, "盲僧接口");put(taskDatetimeType.YURNERO, "剑圣接口");}};public enum taskApiUrlType {LEESIN("leesin"), YURNERO("yurnero");public String key;private taskApiUrlType(String key) {this.key = key;}public static String getValueByKey(String key) {for (taskApiUrlType item : values()) {if (item.key.equals(key)) {return (String) taskApiUrlTypeEnum.get(item);}}return null;}}public static EnumMap taskApiUrlTypeEnum = new EnumMap(taskApiUrlType.class) {{put(taskApiUrlType.LEESIN, "getinformation/leesinlist");put(taskApiUrlType.YURNERO, "getinformation/yurnerolist");}};
}

主要控制接口的相对url和服务的key

ResponseModel作为顶层返回类，通过向下转型可以获取任意子类类型，并且在模板中声明为抽象，交给子类

public abstract class AbstractApiRequest {@Resourceprivate RestTemplate restTemplate;@Resourceprivate ISysTaskTokenService iTskTokenService;@Resourceprivate ApiTokenRequest tokenRequest;@Resourceprivate ISysTaskDatetimeService taskDatetimeService;protected volatile String param;/*** 一次性插入数据的数量*/protected final int limit = 50;@Retryable(value = {RemoteAccessException.class}, maxAttempts = 3, backoff = @Backoff(delay = 5000L, multiplier = 1))public void excute() throws IOException {//1 获取tokenString token = iTskTokenService.getToken().getToken();if (token == null || "".equals(token)) {XxlJobLogger.log("token为空!");tokenRequest.excute();}//2 获取任务执行配置String apiName = initApiName();SysTaskDatetime tskDateTime = taskDatetimeService.getByApiName(apiName);if (tskDateTime == null) {XxlJobLogger.log("接口【" + apiName + "】未配置或不存在");return;}//执行时间String executetime = DateUtil.format(tskDateTime.getExecuteTime(), "yyyy-MM-dd'T'HH:mm:ss");//当前时间String currenttime = DateUtil.format(new Date(), "yyyy-MM-dd'T'HH:mm:ss");//读取的首页数int pageIndex = tskDateTime.getPageIndex();//读取的下一页int nextPage = pageIndex + 1;// 是否有异常boolean isE = false;while (pageIndex <= nextPage) {ResponseEntity<? extends ResponseModel> responseEntity = null;Map<String, Object> map = getStringObjectMap(executetime, currenttime, pageIndex);Map<String, Object> requestExtMap = getRequestExtMap();if (null != requestExtMap) {map.putAll(requestExtMap);}String relativeApiUrl = initRelativeApiUrl();Class<? extends ResponseModel> resultMappingClass = initMappingClass();try {responseEntity = restTemplate.postForEntity(ApiParam.API_BASE_URL +ApiRequestUrlUtil.getRequestUrl(relativeApiUrl, token), ApiRequestUrlUtil.getHttpEntity(map, apiName), resultMappingClass);if (null != responseEntity.getBody()) {int result = responseEntity.getBody().getResult();if (result == 1) {if (responseEntity.getBody().getPagesize() > 0) {insertOrUpdate(responseEntity.getBody());pageIndex++;if (responseEntity.getBody().getNextpage() == 0 || responseEntity.getBody().getNextpage() == null) {nextPage = 0;break;} else {nextPage = responseEntity.getBody().getNextpage();}} else {pageIndex = 1;break;}} else {XxlJobLogger.log("任务名称:{}获取数据异常,异常信息为{}", apiName, responseEntity.getBody().getMsg());break;}} else {pageIndex = 1;break;}} catch (RestClientException e) {isE = true;XxlJobLogger.log("接口远程异常,{}", e.getMessage());String exceptionMessage = "调用任务接口【" + apiName + "】异常: " + e.getMessage() + e.getStackTrace().toString();throw new RemoteAccessException(exceptionMessage);} catch (Exception e) {nextPage = 0;isE = true;XxlJobLogger.log("接口异常,{}", e.getMessage());tskDateTime.setPageIndex(pageIndex);taskDatetimeService.updateById(tskDateTime);}}//3 执行完成后，将页码初始化，并保存当前时间if (!isE) {tskDateTime.setPageIndex(1);tskDateTime.setExecuteTime(new Date());taskDatetimeService.updateById(tskDateTime);}}@Recoverpublic void recover(RetryException e) {XxlJobLogger.log("recovery,{}", e.getMessage());}public abstract String initApiName();public abstract String initRelativeApiUrl();public abstract Class<? extends ResponseModel> initMappingClass();public abstract void insertOrUpdate(ResponseModel entity);private Map<String, Object> getStringObjectMap(String executetime, String currenttime, int pageIndex) {Map<String, Object> map = new HashMap<>(3);map.put("pageindex", pageIndex);map.put("starttime", executetime);map.put("endtime", currenttime);return map;}public abstract Map<String, Object> getRequestExtMap();public void setParam(String param) {this.param = param;}
}

通过抽象拉取业务的模板方法，将可抽象的方法交由子类去完成，比如initApiName()，initRelativeApiUrl()，initMappingClass()。当然改模板方法可以按需再扩展

具体工厂创建

public class ApiRequestFactory extends Factory {/*** 保存注入的所有ApiXXXRequest*/private Map<String, AbstractApiRequest> serviceMap;@Overrideprotected AbstractApiRequest createApiRequest(String apiName) {if (StringUtils.isEmpty(apiName) || null == this.serviceMap) {XxlJobLogger.log("初始化AbstractApiRequest失败");return null;}return serviceMap.get(apiName);}public void setServiceMap(Map<String, AbstractApiRequest> serviceMap) {this.serviceMap = serviceMap;}
}

而具体的工厂最核心的就是将注入的service按照key去标识，我们只需要声明如下形式就可以调用相应的xxApiRequest,并执行模板方法

 AbstractApiRequest apiRequest = factory.create(TaskEnumUtils.taskDatetimeType.LEESIN.key);
apiRequest.excute();

测试

我们需要先启动jwt+shiro+模板模式实现推送（二）中的代码作为推送端，启动xxljob的admin和客户端。然后通过bean模式注册客户端的bean，如下图所示

可分别执行上述盲僧和剑圣接口，即可执行客户端通过 @XxlJob声明的方法，如 @XxlJob(“leesinJobHandler”)

简单说明

当然这里只是一个demo，实际中我们可以扩展许多功能，比如多条产品线的不同业务拉取，同一个产品线的不同产品拉取等

源码地址

友情提示，需先下载推送端，在下载本demo代码github地址