Skip to main content

Add your description here

Project description

DataHarvest

DataHarvest 是一个用于数据搜索🔍、爬取🕷、清洗🧽的工具。

AI时代,数据是一切的基石,DataHarvest 能够帮助获取干净有效的数据。

DataHarvest

搜索支持

搜索引擎 官网 支持
tavily https://docs.tavily.com/
天工 https://www.tiangong.cn/ coming soon

数据爬取&清洗支持

网站 内容 url pattern 爬取 清洗
百度百科 词条 baike.baidu.com/item/
百度百家号 文章 baijiahao.baidu.com/s/
B站 文章 www.bilibili.com/read/
腾讯网 文章 new.qq.com/rain/a/
360个人图书馆 文章 www.360doc.com/content/
360百科 词条 baike.so.com/doc/
搜狗百科 词条 baike.sogou.com/v/
搜狐 文章 www.sohu.com/a/
头条 文章 www.toutiao.com/article/
网易 文章 www.163.com/\w+/article/.+
微信公众号 文章 weixin.qq.com/s/
马蜂窝 文章 www.mafengwo.cn/i/
小红书 文章 www.xiaohongshu.com/explore/、/xhslink.com/

其他情况使用基础playwright数据爬取和html2text数据清洗,但并未做特殊适配。

安装

pip install dataharvest
playwright install

使用

分为搜索、爬虫、数据清洗三个主要模块,互相独立,您可以按需使用对应模块。

爬取和清洗做了根据URL的自动策略匹配,您只需要使用AutoSpider和AutoPurifier即可。

最佳实践

搜索

from dataharvest.searcher import TavilySearcher

api_key = "xxx"  # 或者设置环境变量 TAVILY_API_KEY

searcher = TavilySearcher(api_key)
searcher.search("战国水晶杯")
SearchResult(keyword='战国水晶杯', answer=None, images=None, items=[
    SearchResultItem(title='战国水晶杯_百度百科', url='https://baike.baidu.com/item/战国水晶杯/7041521', score=0.98661,
                     description='战国水晶杯为战国晚期水晶器皿,于1990年出土于浙江省杭州市半山镇石塘村,现藏于杭州博物馆。战国水晶杯高15.4厘米、口径7.8厘米、底径5.4厘米,整器略带淡琥珀色,局部可见絮状包裹体;器身为敞口,平唇,斜直壁,圆底,圈足外撇;光素无纹,造型简洁。',
                     content='')])

爬虫

from dataharvest.spider import AutoSpider

url = "https://baike.so.com/doc/5579340-5792710.html?src=index#entry_concern"
auto_spider = AutoSpider()
doc = auto_spider.crawl(url)
print(doc)

代理

我们需要实现 一个代理生成类,并实现他的__call__方法。 使用时可以在爬虫初始化时,将配置添加进去,也可以在调用时传入。

from dataharvest.proxy.base import BaseProxy, Proxy
from dataharvest.spider import AutoSpider
from dataharvest.spider.base import SpiderConfig


class MyProxy(BaseProxy):

    def __call__(self) -> Proxy:
        return Proxy(protocol="http", host="127.0.0.1", port="53380", username="username", password="password")


def test_proxy_constructor():
    proxy_gene_func = MyProxy()
    auto_spider = AutoSpider(config=SpiderConfig(proxy_gene_func=proxy_gene_func))
    url = "https://baike.baidu.com/item/%E6%98%8E%E5%94%90%E5%AF%85%E3%80%8A%E7%81%8C%E6%9C%A8%E4%B8%9B%E7%AF%A0%E5%9B%BE%E8%BD%B4%E3%80%8B?fromModule=lemma_search-box"

    doc = auto_spider.crawl(url)
    print(doc)


def test_proxy_call():
    proxy_gene_func = MyProxy()
    auto_spider = AutoSpider()
    config = SpiderConfig(proxy_gene_func=proxy_gene_func)
    url = "https://baike.baidu.com/item/%E6%98%8E%E5%94%90%E5%AF%85%E3%80%8A%E7%81%8C%E6%9C%A8%E4%B8%9B%E7%AF%A0%E5%9B%BE%E8%BD%B4%E3%80%8B?fromModule=lemma_search-box"
    doc = auto_spider.crawl(url, config=config)
    print(doc)

清洗

from dataharvest.purifier import AutoPurifier
from dataharvest.spider import AutoSpider

url = "https://baike.so.com/doc/5579340-5792710.html?src=index#entry_concern"
auto_spider = AutoSpider()
doc = auto_spider.crawl(url)
print(doc)
auto_purifier = AutoPurifier()
doc = auto_purifier.purify(doc)
print(doc)

效果:

整合

搜索+爬取+清洗

import asyncio

from dataharvest.base import DataHarvest
from dataharvest.searcher import TavilySearcher

searcher = TavilySearcher()
dh = DataHarvest()
r = searcher.search("战国水晶杯")
tasks = [dh.a_crawl_and_purify(item.url) for item in r.items]
loop = asyncio.get_event_loop()
docs = loop.run_until_complete(asyncio.gather(*tasks))

鸣谢

伙伴们如果觉着这个项目对你有帮助,那么请帮助点一个star✨。如果觉着存在问题或者有其他需求,那么欢迎在issue提出。当然,我们非常欢迎您加入帮忙完善。

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

dataharvest-0.2.6.tar.gz (13.4 kB view details)

Uploaded Source

Built Distribution

dataharvest-0.2.6-py3-none-any.whl (25.4 kB view details)

Uploaded Python 3

File details

Details for the file dataharvest-0.2.6.tar.gz.

File metadata

  • Download URL: dataharvest-0.2.6.tar.gz
  • Upload date:
  • Size: 13.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.10.10

File hashes

Hashes for dataharvest-0.2.6.tar.gz
Algorithm Hash digest
SHA256 9e2f95490fbcf95fa5ac60b481dd7428482e097caa0cba0b20fd4320e97a5ec8
MD5 4ec1d69d9e1fbff093a21df93c611bf3
BLAKE2b-256 ca25611cc5ba797e73eae22a445370a0728d8ed62256f92745fe2587bfe67af2

See more details on using hashes here.

File details

Details for the file dataharvest-0.2.6-py3-none-any.whl.

File metadata

  • Download URL: dataharvest-0.2.6-py3-none-any.whl
  • Upload date:
  • Size: 25.4 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.10.10

File hashes

Hashes for dataharvest-0.2.6-py3-none-any.whl
Algorithm Hash digest
SHA256 9f5ce8d580b9580c8589e0fa6d8445192f099b11e9f07094d856eed75c0dffca
MD5 02f0157c85df81368776e391ffea8ba5
BLAKE2b-256 b3325ee68b77e70c3f7dec12fbd83327d4fdc7aad5e09f4f6d6ab19a31810fc4

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page