Train TJBot to Suggest a Lunch Spot in Node-RED
Today we’ll train TJBot how to suggest a lunch spot using the Watson Conversation service.
Now, let me start this skill with a couple of warnings. First, like probably a lot of teams out there, the team of developer advocates I work with have a tough choice to make everyday. It happens about noontime every weekday we’re not at an event hanging out with awesome developers like you eating the free food (honestly, we work for food). And no, the tough choice doesn’t have to do with which Watson service we want to use next. Well…okay…this too.
Someone peeks up from their laptop and says they’re hungry. We all look up and ask, where do we want to go grab some lunch? The struggle is real here in Silicon Valley.
So, today, we’ll train TJBot to pick a place using the Watson Conversation service. Now, this isn’t a fully baked…solution…a future blog post shall follow this idea as it rises and spreads like hotsauce! Ah, fresh bread right out of the oven.
This is just a fun little (sesame) seed to show how Watson Conversation can randomize the selection of the responses right out of the (lunch) box and get to the meat of the issue: we never know what to eat and end up picking a random place. So, let’s just have Watson and TJBot help with this dilemma. Let’s keep this blog on a roll…white, wheat, or sourdough? Toasted?
Second, please don’t judge the selection of example lunch options as I can’t fudge my foodie skills beyond this point…I obviously have none…this is why we need help picking where we have lunch! Give me a (lunch) break…the struggle is gnawing at me and this slice of inspiration is all that’s left over, and must be used by today’s date. Gosh, this post is turning sour. Wait for it…it’s heating up in here!
Let’s squash this down to the basic ingredients: 1 TJBot, 1 Node-RED editor, and 1 Watson Conversation service.
Create the Watson Conversation service. Train Watson with the #find_lunch intent. Be creative, enter all the phrases you hear in the office at lunch time.
Now we’re cooking with gas! For those foodies out there, add an entity, @ cuisine so we can satiate those hangry folks with the applicable type of cuisine. For simplicity, I’ve added american, chinese, and asian, and left indian and mexican cuisines for you to fill in. Again, I’m not a foodie and need to learn my Nutrition Facts.
Pause for the commercial break (where everything comes out perfect, yeah right) and we’ve found some places and categorized them.
Put the Watson Conversation credentials and workspace ID into the TJBot configuration back in the Node-RED editor.
And here we go. Places to eat. Serves, um, as many people as you need.
Here’s a video of how to train TJBot to suggest a lunch spot.
That’s it for today’s skill(et). What can you train TJBot to do now that it can suggest lunch spots? And remember, keep TJBot’s electronics safe and don’t feed the bots. Even if TJBot speaks nicely to you.
Come back tomorrow and we’ll train TJBot to take a photo and tweet it.
If you enjoyed the sense of humor in this post, please leave a comment or clap for me. Please, no throwing tomatos.
This post is part of a series of skills you can train TJBot to perform.
