joshlasdin / otterbot Goto Github PK

Right now when using the .pic command there is a decent chance that the image selected from google doesn't render in chat. Usually that is because

• The image link is broken (a 404)
• The host does not allow embedding and redirects to a full page
• The url isn't correctly rendered via plug because it contains characters that the plug image url doesn't recognize (e.g. "(")

In either case it renders a broken image link. We should check that the image can actually render for users before we post it into chat. Presumably just fetch the image url, check that the response is 200 and the content type looks like an image.

As per @jkupferman's suggestion: https://devcenter.heroku.com/articles/config-vars

After updating the PlugAPI, otterbot no longer greets people when they enter the room.

Should implement mocha testing with 95% coverage

Test case: .ud cthulhu

@jkupferman @jayteemo

We should probably start thinking on how to move otter over to Slack. Do either of you have any experience with Slack bots?

http://31.media.tumblr.com/a43479b0997b14076b8aa08adde3c8ee/tumblr_mw9tfkGfCl1t1udaco1_250.gif

bot.chatSingle() should take a string of any length and split it into chunks small enough for the destination to handle. The max value should be moved to the config. Current setup can be found in the .ud module.

Now that the plug chat server seems to be more stable we can turn scrobbling back on

Likely happened when moving config values.

Implement TravisCI to run tests on changes.

https://travis-ci.org/

Otter should greet Fred with the following gif: "https://api.plugcubed.net/proxy/http://i.imgur.com/eb55q5D.png"

README is a bit bleak. Make it more informative.

https://devcenter.heroku.com/articles/cedar-14-migration

https://issue.plugcubed.net/wiki/Plug3%3ARSS

Auto-deploy to heroku from TravisCI

http://docs.travis-ci.com/user/deployment/heroku/

Only after both Jersh and Jobu have secret handshake'd eachother, post a random secret handshake gif

When otterbot starts up and there is no song playing it dies after 30 seconds with the following stack trace.

/Users/jonathan/dev/otterbot/services/LastFm.js:26
            'artist': media.author,
                           ^
TypeError: Cannot read property 'author' of null
at _.extend.scrobble (/Users/jonathan/dev/otterbot/services/LastFm.js:26:28)
at /Users/jonathan/dev/otterbot/commands/newsong.js:17:24

newsong seems to get called even if nothing is playing, so this needs a check

Should include information on:

• Getting Grunt setup
• How to run tests

In a recent pricing change Heroku now requires that all free tier apps have at least 6 hours of "sleep" time per 24 hour period. While right now it is just a nag, according to their announcement by January of 2016 they will start enforcing it.

The two options seem to be:

1. Paying $7/mo for a hobby tier dyno and not worrying about it
2. Changing the way otterbot works so that she can get her doctor recommended 6 hours of sleep per night

1. is easy but collecting money from the group is kind of a pain in the butt and desirable if we can avoid it.

2. is do-able but will require some work. Right now otterbot is "kept awake" by an external service pinging otterbot every few minutes. Otterbot can safely sleep from 12AM-6AM EST since no one is around during those hours. What we'd need to do is change the pinging app to not ping during those hours so otter can sleep and at 6AM start pinging again to wake her up again.

This command relied on the imgtype=animated flag in google search which (apparently) was labeled as experimental and that experiment has ended.

This issue and potential solutions are discussed in more detail here:
hubot-scripts/hubot-google-images#10

Most likely from an API change, need to make sure when she parses a chat it makes sure it wasn't from her first. Causes infinite loop problems on commands like "florida"

Gives weird messages like <onlyinclude> and {{multiple image for certain searches. To replicate:

.wiki USWNT
.wiki penis

It'd be pretty sweet to be able to get a random themed gif from within the chatroom. Something like...
.gif drake
.gif cool story bro

WiffleGif can power this by using the search api which is accessible as follows:
http://wifflegif.com/gifs/search.json?q=drake
http://wifflegif.com/gifs/search.json?q=cool%20story%20bro

I'd recommend grabbing one of the results randomly and then extracting the "url" field which is a direct link to the gif.

There are often times when it would be useful to do a quick image search for a person or thing when people don't know what it is. Often times we end up doing a gif search for it but there are no results.

.pic Jenna Elfman
.pic tamagotchi

You can get google image search results using this JSON endpoint: http://ajax.googleapis.com/ajax/services/search/images?v=1.0&rsz=8&safe=active&q=cat

It should be really similar to the gifsearch command

Should return promise, resolving with the response object.

The config.js file contains real passwords and API keys. We probably want to make an example config file and .gitignore config.js.