A task-Based multithread crawler framework for simple-minds.

=====

Version 1.0 beta.

#Features of version 1.0

• freaking task-based， 基于任务
• less network-panic 不需要拘泥于网络的细节，包括错误处理和重试
• more handlers 需要关心handlers，也就是你处理爬到的内容的逻辑实现
• simple http-controling(not implemented yet)-monitoring 基于http的监控和(控制)接口
• mess logs 不止一坨日志
• abysmal bugs maybe 可能潜在的一堆bugs

#Installation

• python 2.7
• And,

while running not OK:
	check what you are lacking of
	add them..

1. Copy the dir to the place you want.
2. Done.

#Try Demo

demo是挖voa的内容，但是voa的内容是被GFW认证的，所以你假如要试用的话，请先确认你的服务器能够连接到 http://learningenglish.voanews.com/

[cheungzee@pn-206-246 iPapa_voa]$cd bin
[cheungzee@pn-206-246 bin]$ls
AudioHandler.py          ClassPageHandler.pyc     ContentPageHandler.pyc   PicHandler.py            setup.py
AudioHandler.pyc         ContentMp3PageHandler.py MainPageHandler.py       PicHandler.pyc           setup.pyc
ClassPageHandler.py      ContentPageHandler.py    MainPageHandler.pyc      run.py
[cheungzee@pn-206-246 bin]$python run.py

wget localhost:28282
or type in localhost:28282 in your favourite broswer.

#Play it to the beat(not finished yet)

There are status of the task noted by task.status,

1. new
2. down
3. downed
4. parsing
5. done
6. failed

1. "page", we will get it once in worker, and put it in task['__data']. 对于page类型的task，worker实例会调用Fetcher 实例来一次性下载task.url指向的内容，并将内容保存到task['__data']中
2. "media", we may download it to the task.dest palce with serveral times, 5MB a chunk for each time. 对于media类型的task我们会进行分块下载，每次5MB，下载到task.dest指定的路径中

1. only task['_data'] is used by the framework. 我们仅仅使用_data 作为task的保留关键字。

1. Check our demo run.py
2. Read the code

def __init__(self, taskId, #taskId
				status='new', #status
				url='', #url
				handler='common', #call which handler to handle it in parsing stage
				handleBy='', #like a stamp saying, I am handling by which one...
               repeatTime=0,  #not implemented yet, when this task failed, shall we put it into the WorkManager.inQueue for a new birth.
               tryTimes=3, # retry times if got erroes in fetching the url 
               waitTime=30, # set timeout as waitTime in fetching the url
               taskType='page', # page or media?
               msg='new', #sth. like testament of the task...
               postdata={}, #if the request is a post request.
               data={}, #data in 'data' will be in task like values in a dict
               ref='', #ref url
               dest='') #put file in the dest, if the taskType is media

take cares of the threads, counters, queues, and blablabla, plz see the code and demo for details.

• WorkManager.inQueue type: Queue.Queue(), we put new tasks in it.
• WorkManager.outPQueue type: Queue.Queue(), Parser put the processed tasks in it.
• WorkManager.wThreads type: list, list of worker threads.
• WorkManager.pThreads type: list, list of parser threads.
• WorkManager.cThread controller thread, providing http interface for monitoring.
• WorkManager.activeTasks type list, storing the active tasks for referencing.

• WorkManager.start(self) start the crawler frame
• WorkManager.packTask(self, task) pack a task by setting the task.id
• WorkManager.addTask(self, task) add a task in self.inQueue and self.activeTasks
• WorkManager.rmTask(self, task) rm a task from self.activeTasks.

1. Get a new task from the inQueue of WorkManager.
2. Set the task.status as "down".
3. Begin to fetch/download the data pointed by task.url into task.dest or task['__data'].
4. Set the task.status to be "downed" or "failed".
5. Put the task into WorkManager.outQueue, i.e., WorkManager.inPQueue.

1. Get a new task from the WorkManager.inPQueue.
2. Parse it from task['__data'] for "page" task / do checking moving jobs for the "media" type task.
3. Set the task.status into "done" or "failed".
4. Put it into the WorkManager.outPQueue.