先看看文档对于Scheduler的作用介绍
https://code4craft.gitbooks.io/webmagic-in-action/content/zh/posts/ch1-overview/architecture.html
之前我们也介绍过了，Scheduler主要负责爬虫的下一步爬取的规划，包括一些去重等功能。在主流程中也看到了Scheduler，现在来具体结合源码分析

源码

Scheduler是一个接口

public interface Scheduler {

/**
* add a url to fetch
*
* @param request
* @param task
*/
public void push(Request request, Task task);

/**
* get an url to crawl
*
* @param task the task of spider
* @return the url to crawl
*/
public Request poll(Task task);

}
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
其主要的实现是DuplicateRemovedScheduler，使用模板模式定义了push的步骤。

public abstract class DuplicateRemovedScheduler implements Scheduler {

protected Logger logger = LoggerFactory.getLogger(getClass());

private DuplicateRemover duplicatedRemover = new HashSetDuplicateRemover();

public DuplicateRemover getDuplicateRemover() {
return duplicatedRemover;
}

public DuplicateRemovedScheduler setDuplicateRemover(DuplicateRemover duplicatedRemover) {
this.duplicatedRemover = duplicatedRemover;
return this;
}

@Override
public void push(Request request, Task task) {
logger.trace("get a candidate url {}", request.getUrl());
if (!duplicatedRemover.isDuplicate(request, task) || shouldReserved(request)) {
logger.debug("push to queue {}", request.getUrl());
pushWhenNoDuplicate(request, task);
}
}

protected boolean shouldReserved(Request request) {
return request.getExtra(Request.CYCLE_TRIED_TIMES) != null;
}

protected void pushWhenNoDuplicate(Request request, Task task) {

}
}
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
我们来看看负责去重的接口DuplicateRemover，其实现类有HashSetDuplicateRemover使用HashSet来去重，RedisScheduler接触Redis来去重和BloomFilterDuplicateRemover使用BloomFilter去重。默认使用HashSetDuplicateRemover

public class HashSetDuplicateRemover implements DuplicateRemover {

private Set<String> urls = Sets.newSetFromMap(new ConcurrentHashMap<String, Boolean>());

@Override
public boolean isDuplicate(Request request, Task task) {
return !urls.add(getUrl(request));
}

protected String getUrl(Request request) {
return request.getUrl();
}

@Override
public void resetDuplicateCheck(Task task) {
urls.clear();
}

@Override
public int getTotalRequestsCount(Task task) {
return urls.size();
}
}
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
DuplicateRemovedScheduler抽象类有四个具体实现类QueueScheduler，PriorityScheduler，FileCacheQueueScheduler和RedisScheduler。默认使用QueueScheduler

@ThreadSafe
public class QueueScheduler extends DuplicateRemovedScheduler implements MonitorableScheduler {

private BlockingQueue<Request> queue = new LinkedBlockingQueue<Request>();

@Override
public void pushWhenNoDuplicate(Request request, Task task) {
queue.add(request);
}

@Override
public synchronized Request poll(Task task) {
return queue.poll();
}

@Override
public int getLeftRequestsCount(Task task) {
return queue.size();
}

@Override
public int getTotalRequestsCount(Task task) {
return getDuplicateRemover().getTotalRequestsCount(task);
}
}
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
其内部是使用了一个LinkedBlockingQueue这个无界队列来存储Request，我们应该看到了@ThreadSafe注解，那我抛一个问题吧。Scheduler是否存在线程同步问题呢，如果存在那是如何解决的呢？
再来看下一个

@ThreadSafe
public class PriorityScheduler extends DuplicateRemovedScheduler implements MonitorableScheduler {

public static final int INITIAL_CAPACITY = 5;

private BlockingQueue<Request> noPriorityQueue = new LinkedBlockingQueue<Request>();

private PriorityBlockingQueue<Request> priorityQueuePlus = new PriorityBlockingQueue<Request>(INITIAL_CAPACITY, new Comparator<Request>() {
@Override
public int compare(Request o1, Request o2) {
return -NumberUtils.compareLong(o1.getPriority(), o2.getPriority());
}
});

private PriorityBlockingQueue<Request> priorityQueueMinus = new PriorityBlockingQueue<Request>(INITIAL_CAPACITY, new Comparator<Request>() {
@Override
public int compare(Request o1, Request o2) {
return -NumberUtils.compareLong(o1.getPriority(), o2.getPriority());
}
});

@Override
public void pushWhenNoDuplicate(Request request, Task task) {
if (request.getPriority() == 0) {
noPriorityQueue.add(request);
} else if (request.getPriority() > 0) {
priorityQueuePlus.put(request);
} else {
priorityQueueMinus.put(request);
}
}

@Override
public synchronized Request poll(Task task) {
Request poll = priorityQueuePlus.poll();
if (poll != null) {
return poll;
}
poll = noPriorityQueue.poll();
if (poll != null) {
return poll;
}
return priorityQueueMinus.poll();
}

@Override
public int getLeftRequestsCount(Task task) {
return noPriorityQueue.size();
}

@Override
public int getTotalRequestsCount(Task task) {
return getDuplicateRemover().getTotalRequestsCount(task);
}
}
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
我们看到了两个PriorityBlockingQueue和一个LinkedBlockingQueue。在poll的时候存在一个顺序。
继续

public class FileCacheQueueScheduler extends DuplicateRemovedScheduler implements MonitorableScheduler {

private String filePath = System.getProperty("java.io.tmpdir");

private String fileUrlAllName = ".urls.txt";

private Task task;

private String fileCursor = ".cursor.txt";

private PrintWriter fileUrlWriter;

private PrintWriter fileCursorWriter;

private AtomicInteger cursor = new AtomicInteger();

private AtomicBoolean inited = new AtomicBoolean(false);

private BlockingQueue<Request> queue;

private Set<String> urls;

public FileCacheQueueScheduler(String filePath) {
if (!filePath.endsWith("/") && !filePath.endsWith("\\")) {
filePath += "/";
}
this.filePath = filePath;
}

private void flush() {
fileUrlWriter.flush();
fileCursorWriter.flush();
}

private void init(Task task) {
this.task = task;
File file = new File(filePath);
if (!file.exists()) {
file.mkdirs();
}
readFile();
initWriter();
initFlushThread();
inited.set(true);
logger.info("init cache scheduler success");
}

private void initFlushThread() {
Executors.newScheduledThreadPool(1).scheduleAtFixedRate(new Runnable() {
@Override
public void run() {
flush();
}
}, 10, 10, TimeUnit.SECONDS);
}

private void initWriter() {
try {
fileUrlWriter = new PrintWriter(new FileWriter(getFileName(fileUrlAllName), true));
fileCursorWriter = new PrintWriter(new FileWriter(getFileName(fileCursor), false));
} catch (IOException e) {
throw new RuntimeException("init cache scheduler error", e);
}
}

private void readFile() {
try {
queue = new LinkedBlockingQueue<Request>();
urls = new LinkedHashSet<String>();
readCursorFile();
readUrlFile();
} catch (FileNotFoundException e) {
//init
logger.info("init cache file " + getFileName(fileUrlAllName));
} catch (IOException e) {
logger.error("init file error", e);
}
}

private void readUrlFile() throws IOException {
String line;
BufferedReader fileUrlReader = null;
try {
fileUrlReader = new BufferedReader(new FileReader(getFileName(fileUrlAllName)));
int lineReaded = 0;
while ((line = fileUrlReader.readLine()) != null) {
urls.add(line.trim());
lineReaded++;
if (lineReaded > cursor.get()) {
queue.add(new Request(line));
}
}
} finally {
if (fileUrlReader != null) {
IOUtils.closeQuietly(fileUrlReader);
}
}
}

private void readCursorFile() throws IOException {
BufferedReader fileCursorReader = null;
try {
fileCursorReader = new BufferedReader(new FileReader(getFileName(fileCursor)));
String line;
//read the last number
while ((line = fileCursorReader.readLine()) != null) {
cursor = new AtomicInteger(NumberUtils.toInt(line));
}
} finally {
if (fileCursorReader != null) {
IOUtils.closeQuietly(fileCursorReader);
}
}
}

private String getFileName(String filename) {
return filePath + task.getUUID() + filename;
}

@Override
protected void pushWhenNoDuplicate(Request request, Task task) {
if (!inited.get()) {
init(task);
}
queue.add(request);
fileUrlWriter.println(request.getUrl());
}

@Override
public synchronized Request poll(Task task) {
if (!inited.get()) {
init(task);
}
fileCursorWriter.println(cursor.incrementAndGet());
return queue.poll();
}

@Override
public int getLeftRequestsCount(Task task) {
return queue.size();
}

@Override
public int getTotalRequestsCount(Task task) {
return getDuplicateRemover().getTotalRequestsCount(task);
}
}
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
会将url和已经执行的url指针存在两个文件中，创建了scheduleExecutor定期的flush，所有内存中的url还是存在BlockingQueue中。
RedisScheduler不是很懂。。目前还没有接触过：）

使用

具体使用过程还是需要自己根据自己的爬虫特点然后选择特定的Scheduler及DuplicateRemover，只有懂得其原理才能选择最合适的组件。
WebMagic组件都可以自行设置这点真的太棒了~

Scheduler的更多相关文章

  1. AndroidStudio3.0无法打开Android Device Monitor的解决办法(An error has occurred on Android Device Monitor)

    ---恢复内容开始--- 打开monitor时出现 An error has occurred. See the log file... ------------------------------- ...

  2. 从scheduler is shutted down看程序员的英文水平

    我有个windows服务程序,今天重点在测试系统逻辑.部署后,在看系统日志时,不经意看到一行:scheduler is shutted down. 2016-12-29 09:40:24.175 {& ...

  3. Spring 4 + Quartz 2.2.1 Scheduler Integration Example

    In this post we will see how to schedule Jobs using Quartz Scheduler with Spring. Spring provides co ...

  4. VMware中CPU分配不合理以及License限制引起的SQL Scheduler不能用于查询处理

    有一台SQL Server(SQL Server 2014 标准版)服务器中的scheduler_count与cpu_count不一致,如下截图所示: SELECT  cpu_count ,      ...

  5. Windows Task Scheduler Fails With Error Code 2147943785

    Problem: Windows Task Scheduler Fails With Error Code 2147943785 Solution: This is usually due to a ...

  6. Fair Scheduler 队列设置经验总结

    Fair Scheduler 队列设置经验总结 由于公司的hadoop集群的计算资源不是很充足,需要开启yarn资源队列的资源抢占.在使用过程中,才明白资源抢占的一些特点.在这里总结一下. 只有一个队 ...

  7. Fair Scheduler中的Delay Schedule分析

    延迟调度的主要目的是提高数据本地性(data locality),减少数据在网络中的传输.对于那些输入数据不在本地的MapTask,调度器将会延迟调度他们,而把slot分配给那些具备本地性的MapTa ...

  8. 【Cocos2d-x 3.x】 调度器Scheduler类源码分析

    非个人的全部理解,部分摘自cocos官网教程,感谢cocos官网. 在<CCScheduler.h>头文件中,定义了关于调度器的五个类:Timer,TimerTargetSelector, ...

  9. Linux IO Scheduler（Linux IO 调度器）

    每个块设备或者块设备的分区,都对应有自身的请求队列(request_queue),而每个请求队列都可以选择一个I/O调度器来协调所递交的request.I/O调度器的基本目的是将请求按照它们对应在块设 ...

  10. Pair Project: Elevator Scheduler [电梯调度算法的实现和测试]

    作业提交时间:10月9日上课前. Design and implement an Elevator Scheduler to aim for both correctness and performa ...

随机推荐

  1. IUAP--单点登录

    登录组件: 提供统一的登录组件 身份.证明验证身份 支持多种身份标识,用户名.邮箱.手机号 支持多个域,从与得到用户响应的角色,权限进行验证用户时候能进行操作. 支持会话管理和安全管理 支持多种验证策 ...

  2. 本地 vs. 云：大数据厮杀的最终幸存者会是谁？— InfoQ专访阿里云智能通用计算平台负责人关涛

    摘要: 本地大数据服务是否进入消失倒计时?云平台大数据服务最终到底会趋向多云.混合云还是单一公有云?集群规模增大,上云成本将难以承受是误区还是事实?InfoQ 将就上述问题对阿里云智能通用计算平台负责 ...

  3. Rol租车网项目总结

    现在自行车的租聘如此火爆,我们是否需要加入这个浩浩荡荡的行列? 相比起现在ofo共享单车,摩拜单车.而我们的竞争力在何处? 如何能让我们的项目脱颖而出? 而我们的Rol租车网为什么要叫Rol呢? Ri ...

  4. 关于spring中<util:/>的配置

    解决redis设置缓存时间找到的帖子,我这个初学者需要学习的还是很多的. 原文地址:http://www.doc100.net/bugs/t/216322/index.html 探索<util/ ...

  5. python 模块的导入

  6. No PostCSS Config found解决办法

    npm install报错 Module build failed: Error: No PostCSS Config found 解决办法是同级package.json文件新建postcss.con ...

  7. 2019-6-23-WPF-网络-request-的-read-方法不会返回

    title author date CreateTime categories WPF 网络 request 的 read 方法不会返回 lindexi 2019-06-23 11:26:26 +08 ...

  8. python 异常处理技巧

  9. 【NS2】在linux下安装低版本GGC

    1.下载安装包,cd到文件所在目录 sudo dpkg -i gcc41-compat-4.1.2-ubuntu1210_i386.deb g++41-compat-4.1.2_i386.deb 2. ...

  10. day13 SQLAlchemy

    ORM:也叫关系对象映射 本篇要点: 原生模块 pymsql ORM框架 SQLAchemy pymysql pymsql是Python中操作MySQL的模块,其使用方法和MySQLdb几乎相同. 需 ...