这篇文章专访segmentfault说到:
我们从一开始就会把搜索作为一个优先解决的技术问题。我们在做站内搜索时花了很大的力气。最开始我们用的是Lucene的引擎做了一个Java的一个后端搜索服务,后来我们是把整个搜索全部转移到Redis平台上,用Redis来做站内的搜索,这样效率会比较高,而且搜索起来准确度比较高。
很想了解这个redis如何实现站内搜索的。 感谢知道的人指点方向, 先谢谢啦
PHPz2017-04-21 10:59:29
The most basic framework of search is word segmentation + inverted index. Weights and so on are all details of the model (of course, for Baidu and Google, this is the most important thing, they need to go up pagerank and so on).
The set operation of redis is simply customized for inverted index.
黄舟2017-04-21 10:59:29
In fact, there are many ways to search. You can refer to this answer http://segmentfault.com/q/1010000000663373/a-1020000000702395
PHP中文网2017-04-21 10:59:29
I use Lucene more, which is more customizable, but the amount of development and maintenance is slightly larger. I recently tried to use Souyi's in-site search engine, which is good for websites with small and medium-sized data volumes.