run npm install to install dependent modules, then run nodemon <filename> to start listening port 3000.
douban(豆瓣), cnode
run nodejs script on rented server whenever pull request reaches, then scrape curated tech news link and brief description from media and return back in JSON.
- How "curated"? use tagging and classfify the article by keywords, natural language processing!
- How "briefly summerize an article"? use given description as much as possible, while try to abstract "key sentences" from articles? Should I go this way?
- How to handle "pull request"? Run the scripts on server restlessly? Nope! But only running when received request will give a long reaction time. Use pointer! Avoid direct object comparison but use pointer comparison!