Hacker News new | past | comments | ask | show | jobs | submit login
HNSearch API Contest (hnsearch.com)
115 points by andres on June 4, 2011 | hide | past | favorite | 31 comments



Every hnsearch contestant gets a dotCloud account to run their app easily, and for free.

Email hn@dotcloud.com to get started.


I sadly don't have time to implement it, but I'm still suggesting the idea.

You enter a keyword - say Python.

It shows a graph where nodes are submitted articles and comments intermixed. On each "Comment" node, we see the picture of the author (obviously if found), the karma and other possible information about him/her (Blog, website, startup company, etc.).

Also, on each nodes start new related nodes.. So, if I search for python, I might have a link to a new node called Django which would then link to useful comments and other articles/comments.

So, the overall idea is to really be able to "visually browse and learn through hacker news". "Vim" would give all plugins, author's plugin, etc.


Just for curiosity's sake, how is this dataset populated? Is ThriftDB just scraping HN? Is there some kind of back-end DB feed?


Well this is as far as I got in about an hour for adding a /hn tag to Unscatter.com - http://www.unscatter.com/search/?q=%2Fhn+facebook

Javascript paging still isn't and it needs to be prettied up, but I figure not bad for about an hours worth of work. And a good demonstration of the framework Unscatter.com is built on being suitable for rolling out new apis pretty quickly.

I'd wrap it up today but I got family stuff today, should be complete by 6/20. Until I finish the javascript and layout /hn won't be advertised on the sidebar, but anyone is free to start using it, I won't be turning it off.


Awesome! It would be great to start a discussion about ranking algorithms in the HNSearch developer forum (http://groups.google.com/group/hnsearch). We'd like to implement some ThriftDB features to tweak ranking algorithms automatically but that won't be ready for a while.


I haven't even gotten into ranking yet. I took the example url and rolled with that. I think the first thing I'll focus on is seeing if I can also offer /hncomment and /hnlinks to help people search different types, then work on the algorithms for all 3. Once I have /hn display looking good I can roll it out to the other 2 pretty quickly.


We just changed the contest date from June 15th to June 20th to give people more time to work on apps.


OK - I have no illusions about winning the contest, but just for fun I gave it a go by adding support to my existing Android app, SocialMediaTicker. My app is designed for when your device is charging or otherwise idle and is a bit of an eye-candy display. It's a 3D, interactive ticker-tape type display that scrolls facebook feeds, twitter streams, news headlines, photos, rss feeds, etc. Click once to stop an item, click on it again to call up the full photo, story, rss feed, etc.

For HN, I added a headline search capability and also made it part of my experimental Twitter Trends mashup where the search terms for content automatically come from the currently trending topics under Twitter. This can be hit-and-miss for HN as oftentimes Twitter trending terms don't match well to the topics found under HN, but other times I've been surprised about what pops up.

If you want to give the mashup a try, I'd turn most other content off, turn the mashup on, sit back and watch what happens! Configuration of all content can be found in the preferences menu. By default HN is off.

To give it a try on your Android, download the app here:

http://www.socialmediaticker.com/SocialMediaTicker.apk

Feedback appreciated!


Anyone wanna team up on this? I'm thinking of a WordPress plugin that shows links to related articles or discussions based on a post's title or tags ...


Here's my simple HNSearch API JavaScript example. It lists submissions from the last 3 hours and sorts by points.

http://www.hnjs.dotcloud.com/

http://news.ycombinator.com/item?id=2663282


This probably won't win, but it would be great if someone could write a tiny app that converts search results sorted by date into RSS feeds. I might hack something up myself after I'm done with exams, if no one else does.



That is for all(?) of Y Combinator, whereas you might be interested in new posts with some criterion (keyword, points, etc.)


Interesting. I wonder how you go about setting up an rss feed.


What makes the monitor "hacker-optimized"?


It's really big.


disclaimer: this is shameless self promotion for something i coded some time ago

well, is somebody hacks up something nice via node.js why not transform the whole hnsearch api into a nice npm package using

https://github.com/franzenzenhofer/apimaker

basically you put some nice URLs in -> magic -> a wrapped API comes out.


What are reasonable rates for querying the API?


For now, don't worry about rate limits.


Hello I've spent the last two days hacking together a mobile site for hacker new search API... the mobile site is fully function but still needs some UI updating I currently working on integrating into twilio to allow you to pull search results via TEXT kinda over kill but why not!

http://hnsearch.mobi


Here is one: http://HNTop100.com -- the top Hacker News posts of all time! Sort by most up voted or most commented on. This took me way longer than it need to because I kept finding awesome articles to read! Enjoy xD


If this is successful, will you consider running a similar contest for Arc Forum? http://news.ycombinator.com/item?id=2620297

Edit: Fixed link to have correct URL. Sorry for the mistake.


Has someone written a crawler for the forum? If so, I can help with the ThriftDB setup.


That would be amazing! I think there have been at least a couple crawlers written for it. (One being SearchYC's.) I'll let you know when I find something more concrete.

Is there anything else that would help make the ThriftDB setup go more smoothly?


I don't think so. Once we have the data it'll be easy to upload it. I'm emailing with the SearchYC creator so I'll ask him about arc forum data.


Kartik is happy to contribute his crawler, but wonders if it wouldn't be easiest to reuse the HNSearch crawler (since both forums have the same structure). http://arclanguage.org/item?id=14774


Thank you, Andres. I put up a message about it on the Arc Forum as well: http://arclanguage.org/item?id=14772. Please email me at evanrmurphy@gmail.com if you need assistance.


Put in a quick first entry for the contest (HNInstant.com):

http://news.ycombinator.com/item?id=2621144

More coming...


Looks like fun, gears are churning.


anyone looking for a coding buddy to do something or ideas? i have some free time atm


How are group entries supposed to share the prize? Refund for dinero?




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: