requests_spider 0.0.4

Web crawling framework like flask.

<div align=center>
<img src="./spider.png">
</div>

## 描述
requests_spider 是一个轻量级的异步爬虫框架，基于requests_html进行二次开发，类似flask

## 安装
pip install requests_spider

## 依赖
python: > 3.6
uvloop
requests_html

## 用法
##### 基础例子
```python3
import json
from spider import XField, Spider, Model, Response, Request


class Proxy(Model):
ip = XField(rule='//tr[contains(@class, "odd")]/td[2]', first=False)
port = XField(rule='//tr[contains(@class, "odd")]/td[3]', first=False)

async def process(self, response: Response):
with open('proxy1.txt', 'a+') as file:
for result in self.merge():
file.write(json.dumps(result) + '\n')


spider = Spider('proxy', workers=15)
spider.domains = ['www.xicidaili.com']
spider.init_requests = [
Request(url='http://www.xicidaili.com/nn/{}'.format(x), model=Proxy) for x in range(1, 10)
]

spider.async_limit = 5

if __name__ == '__main__':
spider.run()

```
爬取代理网站ip

##### 中间组件
```python3
import random
import re

from spider import Spider, Request, XRequest, Model, XField, RField, Response, Field, asyncio

# 获取某个用户的所有的视频信息 ===> 获取aid / page
videos_url = "https://space.bilibili.com/ajax/member/getSubmitVideos?mid={mid}" \
"&pagesize=30&tid=0&page={page}&keyword=&order=pubdate"

# 某个视频推荐的视频 ===> 获取aid
recommend_url = "https://comment.bilibili.com/playtag,{cid}-{aid}?html5=1"

# 用户信息 post csrf/mid
user_url = "https://space.bilibili.com/ajax/member/GetInfo"

# av页面，====> 获取下载视频的url、cid-aid, mid
av_url = "https://www.bilibili.com/video/av{aid}"


class AV(Model):
urls = RField(rule='"url":"(.*?)","backup_url"', first=False)
cid = RField(rule='cid=(.*?)&aid=')
aid = RField(rule='&aid=(.*?)&pre_ad=')
mid = RField(rule='"owner":{"mid":(.*?),')

async def process(self, response: Response):
print(self['urls'])
print(self['cid'])
print(self['aid'])
print(self['mid'])
print(self.json())
if self['mid'] and self['aid'] and self['urls'] and self['cid']:
# 推荐视频
yield Request(url=recommend_url.format(cid=self['cid'], aid=self['aid']), model=Recommend)

# 用户信息
yield Request(url=user_url, method='POST', data={'csrf': '', 'mid': self['mid']},
model=UserInfo, not_filter=True)

# 下载视频
for order, url in enumerate(self['urls']):
yield Request(url=url.replace('http', 'https'),
meta={'name': self['aid'] + '_' + str(order)}, model=Video)


class UserInfo(Model):
mid = Field()
name = Field()
sex = Field()
rank = Field()
face = Field()
regtime = Field()
birthday = Field()
sign = Field()
level_info = Field()

async def process(self, response: Response):
status = response.json().get('status')
if status:
data = response.json().get('data')
for k in self.keys():
if k in data:
self[k] = data[k]
with open('user_' + str(self['mid']) + '.txt', 'w') as f:
f.write(self.dumps() + '\n')


class Recommend(Model):

async def process(self, response: Response):
for data in response.json():
yield Request(av_url.format(aid=data[1]), model=AV)


class VideoInfo(Model):

async def process(self, response: Response):
status = response.json().get('status')
if status:
data = response.json().get('data')
pattern = 'mid=(\d+?)&pagesize=30&tid=0&page=(\d+?)&keyword=&order=pubdate'
patn = re.findall(pattern, response.url)[0]
print(patn)
yield Request(url=videos_url.format(mid=patn[0], page=int(patn[1]) + 1), model=VideoInfo),
for v in data['vlist']:
yield Request(url=av_url.format(aid=v.get('aid')), model=AV)


class Video(Model):

async def process(self, response: Response):
file_name = response.current_request.meta.get('name')
if file_name and response.status_code == 200:
with open(file_name + '.mp4', 'wb') as f:
for content in response.iter_content(chunk_size=512):
f.write(content)
f.flush()


spider = Spider('bilibili', workers=5)

spider.init_requests = [
Request(url=videos_url.format(mid='35789774', page=1), model=VideoInfo),
]
spider.async_limit = 5


@spider.Middleware('request')
async def test(request):
print(request.url)
if request.url.startswith('https://space.bilibili.com/'):
request.info.update({'headers': {'Referer': 'https://space.bilibili.com/'}})
else:
request.info.update({'headers': {'Referer': 'https://bilibili.com/'}})

asyncio.sleep(round(random.random() * 5))
return request


if __name__ == '__main__':
spider.run()
```
爬取bilibili用户视频，用户资料，视频资料，利用中间组件进行切换headers

## API
#### Spider
继承requests_html的HTMLSession

- **Spider.async_limit**

利用asyncio.Semaphore限制并发数量


- **Spider.queue_timeout**

从队列获取数据时候超时设置

- **Spider.request_depth**

请求的深度

- **Spider.init_requests**

初始化请求

- **Spider.domains**

爬取域名设置

- **Spider.rules**

从响应的数据中获取下次请求的信息，并加入队列

- **Spider.Middleware**

中间组件
Middleware('request'), request入队之前执行，返回request, response, None
Middleware('response'), response入队之前，返回request, response, None


##### Model
Model类似一个字典的数据模型

- **Model.keys**

类似字典的keys

- **Model.values**

类似字典的values

- **Model.items**

类似字典的items

- **Model.json**

获取所有Field的字典形式

- **Model.dumps**

获取所有的Field的字符串

- **Model.merge**

当所有的Field从响应数据获取的数据是列表的时候，将获取的列表合并成为json数据

- **Model.process**

处理响应数据


##### Field

- **Field**

不处理或待处理数据项

- **XField**

利用xpath从响应数据中获取数据

- **CField**

利用css获取数据

- **RField**

利用正则获取数据

##### Request

- **Request**
正常的请求

- **XRequest**

利用xpath, 用于Spider.rules

- **RRequest**

利用正则，用于Spider.rules



## 例子
examples目录下
bilibili.py 爬取哔哩哔哩用户信息、视频信息和视频
qidian.py 爬取起点小说月票排行包括评分
proxy.py 爬取代理ip网站代理
pearvideo.py 爬取梨视频网站的视频

# License
MIT