Yesterday, I saw someone asking “which programming language is better for writing a web crawler? PHP, Python or Node.js?”and mentioning some requirements below.
-
The analytic ability to web page
-
Operational capability to database(MySQL)
-
Efficiency of crawling
-
The amount of code
Someone answered to the question.
“When you are going to crawl large-scale websites, then efficiency, scalability and maintainability are the factors that you must consider.”
Crawling large-scale websites involves many problems: multi-threading, I/O mechanism, distributed crawling, communication, duplication checking, task schedule, etc. And then the language used and the frame selected play a significant role at this moment.
PHP
The support for multithreading and async is quite weak and therefore is not recommended.
Node.js
It can crawling some vertical websites. But due to the support for distributed crawling and communications is relatively weaker than the other two. So you need to make a judgment.
Python
It’s strongly recommended and has better support for the requirements mentioned above, especially the scrapy framework. Scrapy framework has many advantages:
Support XPath
-
Good performance based on twisted
-
Has debugging tools
If you want to perform dynamic analysis of JavaScript, it’s not suitable to use casperjs under the scrapy framework and it’s better to create your own javescript engine based on the Chrome V8 engine.
C & C ++
I’m not recommended. Although they have good performance, we still have to consider many factors such as cost. For most companies it is recommended to write crawler program based on some open source framework. Make the best use of the excellent programs available. It’s easy to make a simple crawler, but it’s hard to make an excellent one.
Magic?
Truly, it’s hard to make a perfect crawler. But if there is such a software program that could meet your various needs, do you want to have a try?
The features of web crawlers:
-
Free yet powerful
-
Support data extraction of arbitrary HTML elements
-
Support distributed crawling
-
High concurrency
-
Deal with static pages and AJAX pages
-
Provide Data API
-
Connect to Database to export data
