Explore Nearby: Separate API for Bot List loading

[zavreb]

Issue: Current Explore Nearby screen displays bots from the 'Bot List' API. Explore Nearby needs to have its own API.

Per @bengtan:

Explore Nearby should be using a separate list of bots to display (with it's own loading). It shouldn't be using the 'Bot List'. 👍 1

AC pending w/discussions w/ @thescurry

AC:

1. Prioritization of bots to display, should solely be based on distance. All nearby bots must be displayed (with proper visibility settings) within screen view.
2. Bot loading should currently not be an issue. (Should we cap # of bots displayed per screen view?) cc: @bengtan, @thescurry
3. Display all friend/follower accounts bots with appropriate visibility setting.
4. If one address has multiple bots please display first bot created or any criteria of your choosing. (This bit is still being worked out w/design).

[zavreb]

@aksonov since AC has just been determined feel free to consider this a part of the next sprint.

[zavreb]

@bengtan thoughts?

[thescurry]

@bengtan (cc: @zavreb) isn't this really a backend ticket?

[bengtan]

The server side API already exists.

What's missing ... is that there needs to be a discussion on how to limit the bots which appear in 'Explore nearby', and how to sort them by (some definition of) 'importance'.

I said something similar here:

#341 (comment)

but that seems to have been overlooked by everyone.

[thescurry]

I think I agree with everything except point #4 from @zavreb. The reason being is that we don't have that problem yet, we barely have content... so let's not prematurely optimize for what happens when we have too much content. Also we have some early mocks for too much content on a map... need to review those again. Off to look at #341 now.

[zavreb]

Per estimation meeting. For now we are loading all bots. In the future we need to tweak this ticket.

[aksonov]

@toland, @bernardd, @bengtan Could you help me with API? I'm using syntax taken from https://github.com/hippware/tr-wiki/wiki/Bot-Geosearch and getting exception:

rawOutput: <iq type='get' xmlns='jabber:client' id='49a1b776-e8b3-4154-88d0-ffe64dd813ad:iq' to='testing.dev.tinyrobot.com' from='46886766-fa85-11e6-bb1f-0eedd159a4c5@testing.dev.tinyrobot.com'><bots xmlns='hippware.com/hxep/bot' lat='11' lon='12.5'/></iq>
_dataRecv called<iq from='testing.dev.tinyrobot.com' to='46886766-fa85-11e6-bb1f-0eedd159a4c5@testing.dev.tinyrobot.com/20DAC82C1310E7391489590296314218' id='49a1b776-e8b3-4154-88d0-ffe64dd813ad:iq' type='error' xmlns='jabber:client'><error code='500' type='wait'><internal-server-error xmlns='urn:ietf:params:xml:ns:xmpp-stanzas'/><text xmlns='urn:ietf:params:xml:ns:xmpp-stanzas'>{{#{'__exception__' =&gt; true,
    '__struct__' =&gt; 'Elixir.Protocol.UndefinedError',
    description =&gt; &lt;&lt;&gt;&gt;,
    protocol =&gt; 'Elixir.Enumerable',
    value =&gt; nil},
  [{'Elixir.Enumerable','impl_for!',1,[{file,"lib/enum.ex"},{line,1}]},
   {'Elixir.Enumerable',reduce,3,[{file,"lib/enum.ex"},{line,116}]},
   {'Elixir.Enum',map,2,[{file,"lib/enum.ex"},{line,1776}]},
   {'Elixir.Wocky.Index',handle_call,3,
                         [{file,"lib/wocky/index.ex"},{line,46}]},
   {gen_server,try_handle_call,4,[{file,"gen_server.erl"},{line,615}]},
   {gen_server,handle_msg,5,[{file,"gen_server.erl"},{line,647}]},
   {proc_lib,init_p_do_apply,3,[{file,"proc_lib.erl"},{line,247}]}]},
 {'Elixir.GenServer',call,[wocky_index,{geosearch,&lt;&lt;"11"&gt;&gt;,&lt;&lt;"12.5"&gt;&gt;},5000]}}</text></error></iq>

[bernardd]

The exception is occurring because we don't have Algolia indices enabled on testing. Try the same thing on staging or us1 and it should work.
I'll add some code that will return a more informative error for this case.

[aksonov]

@bernardd It is pity because I can't run unit tests for staging (because of Digits auth). Any chance to make it work on development (so my unit tests work)? Also development of features is much faster with unit tests than with ios emulator.

[aksonov]

@bernardd Okey, I've run the test on staging with my credentials. However maybe we could make it work on development at the future.

[bernardd]

Any chance to make it work on development

There's no technical impediment. The main reason we didn't was that we're charged per query, so we didn't want large scale runs of unit tests running up our bills :) I might leave you, @thescurry and @toland to work through that one once you're all on line - or we could discuss it face to face next week :)

[aksonov]

@toland @bernardd We need to add 'owner' to avoid extra request for each bot to retrieve owner.

cc @bengtan

[thescurry]

@bernardd @aksonov (cc: @toland ) I'm open to all options on the table here. Happy to continue the discussion here or in SF.

[bengtan]

I'm open to all options on the table here. Happy to continue the discussion here or in SF.

To be continued here:

Discuss: Algolia search for unit tests
hippware/wocky#574

[zavreb]

Can we talk about this? (cc: @thescurry, @bengtan)

[zavreb]

Pending server deploy to staging (cc: @bengtan, @thescurry, @mstidham )

[zavreb]

Verified on Staging. Pushing ticket to "Verify on Prod"

[mstidham]

Verified on Prod