A good web scraping framework helps to keep your code maintainable by, among other things, enabling and encouraging separation of concerns.
For example, Scrapy lets you implement different aspects of web scraping, like ban avoidance or data delivery, into separate components.
However, there are 2 core aspects of web scraping that can be hard to decouple: crawling, i.e. visiting URLs, and parsing, i.e. extracting data.
web-poet lets you write data extraction code that:
Makes your web scraping code easier to maintain, since your data extraction and crawling code are no longer intertwined and can be maintained separately.
Can be reused with different versions of your crawling code, i.e. with different crawling strategies.
Can be executed independently of your crawling code, enabling easier debugging and easier automated testing.
To learn more about why and how web-poet came to be, see From the ground up.