Issue 175 rev1 (team 4)

[agiz]

• Events from culture.si
• reverse geo api

For issue: #175

I am using bits from #171

[mitar]

Please update from main repository.

Didn't we say that we will be accessing external data sources using background tasks and not from client?

[agiz]

Umm.. There was a discusion but we have not come to the conclusion. Plus that will depend on the nature of an API. If it can be accessed by client, there is no need in using our server's cycles. And you do not have to comment this at 3:00 am :) No need to rush, I can wait and I am always on stand-by haha :P

[mitar]

I think we decided that we will have a fixed set of locations (our nodes) and prefetch data regularly for them. Because also for clients not connected to our portal through our wireless nodes, we will pretend they are close to some existing node.

[agiz]

I do not understand how this changes anything. I am strongly against server prefetching where it is not necessary. As a top contributor I believe I have the right to my own opinion especially now that I have already made the whole API the way it is, including reverse geolocation that will be impossible to do on server side (will promptly overcome the limits). Do with this pull request whatever you like. Also no hard feelings :)

[mitar]

Please update from main repository.

[mitar]

Of course you have right to your opinion! I also have right to my opinion about your final mark! No hard feelings. ;-)

[mitar]

OK, seriously. I do not see any other way than doing it on the server. If venues are missing their geolocation, you have to map all missing geolocation venues' addresses to their geolocation. And if you do this on client and on server, there is same number of requests. But on the server you can do this with different databases, slower, with delays ... Also once you do this, search is much easier (get me all venues 100m around the node).

Also, on server side we can do much better reverse geolocation, because we can also use other services to do that.

Furthermore, we want to support search over all content on the webpage. If we prefetch this data, we can also much easier search over it.

I really do not see a reason for not prefetching it. See prefetching as a caching proxy to their API, where we also clean a bit data, and improve it. If we will have millions of users, do you want that poor their API falls apart because of us?

[mitar]

Anything new here? Please update from main repository, too.