=========================================================================================================
INTRODUCTION
This is Distributed Microblog Spider which can used to getting the social network of users
and the microblog tweeted by users. Besides, it can update the latest information everyday.
The spider can mainly divided into two parts, the server and client. Server is responsable
for arranging tasks, collecting data, maintaining proxy pool and managing databases. While cli-
-ent is responsable for collecting data from website and sending them to server.
=========================================================================================================
DEPENDENCY
Language : Python 3.4
External Python Lib : tornado
pymongo
redis
pymysql
Database : MySQL
Redis
MongoDB
Others : nginx
URL : python : https://www.python.org/
MySQL : http://www.mysql.com/downloads/
MongoDB : https://www.mongodb.org/
Redis : http://redis.io/
tornado : http://www.tornadoweb.org/en/stable/
nginx : http://nginx.org/
=========================================================================================================
USAGE
Befor running this spider, you have to build several tables in MySQL first.
-------------------------------------------------------------
1. Build new database. Name : microblog_spider
Character Sets : gb2312
Collations : gb2312_chinese_ci
-------------------------------------------------------------
2. Create some tables. Name : accuracy_table
Column : float | accuracy
datetime| time
int | wanted_blog_num
-------------------------------------
Name : cache_atten_web
Column : varchar | from_uid
int | from_fans_num
int | from_blog_num
varchar | to_uid
int | to_fans_num
int | to_blog_num
-------------------------------------
Name : cache_attends
Column : varchar | uid
varchar | basic_page
varchar | name
varchar | gender
int | blog_num
varchar | description
int | fans_num
-------------------------------------
Name : cache_history
Column : datetime| latest_time
int | latest_timestamp
varchar | container_id
int | checkin_timestamp
datetime| is_dealing
The component and details of the project(__edition_0.9_)
---------------------------------------------------------------------------------------------------------
This is a Sina Microblog Spider archieved by server terminal and client terminal.
The Server terminal mainly deal with 4 tasks.
1. Get proxy from http proxy website. Unfortunately, not all proxy scratched from website
is valid. In fact, almost one quarter of them is valid. So, it is important to check if
those proxy is valid. One thread of server will manage a proxy pool.
Raw and unchecked proxy will scratched from website, and this thread will arrange these
raw proxy to several sub thread, each sub thread will check if the arranged proxy is
valid. If so, these proxy will be sent to valid proxy pool and ready for use.
Besides, the valid proxy may extend after some time. So it is necessary to check if the
proxy in proxy pool is valid.
2. The second part is communicating with clients. Clients will require proxy and task from
server. When client finish the task of getting info of users, they will return the user info
to server. When received this user info , server will store then in cache database.
3. The 3rd part is about receiving the data translated from client. Sometimes client will
translate data which is sized of hundred of MBs. When server is receiving there data, the query
from other client will be denied. So it's necessary to add some data server which is used to
receive data specially.
4. The 3rd part of function is dealing with the data in cache databases. The attends of user
will be checked if already exist in user_info_table. If not, the user will be inserted into
user_info_table. And no matter whether it exists, the user will be deleted from ready_to_get
table. And ,as for attends, if they are not contained in user_info_table, they will be inserted
to ready_to_get table.
==========================================================================================================
UPDATE HISTORY:
_0.9_ : add update function. Can update microblog
_0.8_ : start to use multi data servers and nginx
_0.7_ : the function of putting data to MongoDB is moved from server.py to server_database.py,
which improve the speed of reaction of main server
_0.6_ : add a data server to avoid the broken of server. And the data of weibo is translated in
small parts
_0.5_ : add the function of scratching the history of certain user
_0.4_ : if the average proxy size less than 30, refuse to assignment task to client
_0.3_ : drop the attends whoes fans num less than 1000
_0.2_ : add the partition to monitor the state of proxy pool
_0.1_ : the initial edition
==========================================================================================================
开机流程：
1.开启MySQL
启动 mysqld --user=mysql
2.开启MongoDB
启动 mongod --dbpath I:\MongoDB\data
3.将redis持久化关闭
config set save ""
4.将user info table 写入redis
5.启动nginx
6.开启花生壳
7.启动服务器
8.SSL连接云主机
9.启动客户端
==========================================================================================================
爬虫重新冷启动流程：
1.drop mongodb中的 assemble factory 和 formal两表
2.清空mysql 中的cache_history表
truncate table cache_history
3.将mysql-》user_info_table中的update_time,latest_blog和isGettingBlog设为null
update user_info_table set update_time=null,latest_blog=null,isGettingBlog=null