Scrapy Tutorial: Installation, Project Structure, Basic Usage, and Real‑World Example

Scrapy is a fast, high‑level Python framework for web crawling and data extraction, allowing developers to build spiders with minimal code.

The framework consists of several components: Engine, Scheduler, Downloader, Spider, Item, Pipeline, Downloader Middlewares, Spider Middlewares, and Scheduler Middlewares.

Installation can be performed via pip: $ pip install scrapy or by downloading the package first:

$ pip download scrapy -d ./
# Using a domestic mirror
$ pip download -i https://pypi.tuna.tsinghua.edu.cn/simple scrapy -d ./

After downloading, install the wheel: $ pip install Scrapy-1.5.0-py2.py3-none-any.whl Creating a project is done with: scrapy startproject mySpider Generate a spider: scrapy genspider demo "demo.cn" Typical workflow includes four steps: creating a project, generating a spider, extracting data (e.g., using XPath or CSS selectors), and saving the data via pipelines.

Running a spider can be done from the command line: scrapy crawl qb # qb is the spider name or programmatically in PyCharm:

from scrapy import cmdline
cmdline.execute("scrapy crawl qb".split())

The project directory contains configuration files such as scrapy.cfg, the Python module folder mySpider/, items.py, pipelines.py, settings.py, and the spiders/ directory where spider code resides.

Example items.py definition:

import scrapy
class MyspiderItem(scrapy.Item):
    pass  # define fields here

Example pipelines.py for CSV output:

from itemadapter import ItemAdapter
import csv
class MyspiderPipeline:
    def __init__(self):
        self.f = open('Zcool.csv', 'w', encoding='utf-8', newline='')
        self.writer = csv.DictWriter(self.f, fieldnames=['imgLink','title','types','vistor','comment','likes'])
        self.writer.writeheader()
    def process_item(self, item, spider):
        self.writer.writerow(dict(item))
        return item
    def close_spider(self, spider):
        self.f.close()

A sample spider that extracts items from ZCOOL:

import scrapy
class DbSpider(scrapy.Spider):
    name = 'db'
    allowed_domains = ['douban.com']
    start_urls = ['http://douban.com/']
    def parse(self, response):
        pass  # extraction logic here

Data extraction uses selectors, e.g., response.xpath() or response.css(), with methods like extract(), extract_first(), get(), and getall().

Pagination can be handled by following the "next" link:

next_href = response.xpath("//a[@class='laypage_next']/@href").extract_first()
if next_href:
    next_url = response.urljoin(next_href)
    yield scrapy.Request(next_url)

Alternatively, construct URLs manually using a page counter.

Running the crawler via a helper script ( start.py) simplifies execution:

from scrapy import cmdline
cmdline.execute('scrapy crawl zc'.split())

After execution, the scraped data is saved to Zcool.csv, confirming successful collection.