API Smash Battle

 API Smash Battle

That's right folks. Everyone loves a good mashup. So, today we're going to learn how to combine data from more than one API.

NewsAPI

👉 Let's get signed up for our APIs. First up is News API.

Hit 'Get API Key'

Create an account

Copy the API key

Add it as a secret to your repl called newsapi

Go to the get started guide

Choose the Get current top articles option and grab the full URL for the GET request  

 

Put it in your code as an fString. Replace the last bit with a {newsKey} variable (that will store the API key). We'll break it down further in a bit.

url = f"https://newsapi.org/v2/top-headlines?country=us&apiKey={newsKey}"

Import os

Now import the os library and add the API key in using the code in the secrets menu.  

import os

newsKey = os.environ['newsapi']

url = f"https://newsapi.org/v2/top-headlines?country=us&apiKey={newsKey}" 

Request time

👉 Now it's request time. I'm going to replace the hard coded 'country code' in the url with a variable to make it customizable later on.  

import os

newsKey = os.environ['newsapi']

country = "us"

url = f"https://newsapi.org/v2/top-headlines?country={country}&apiKey={newsKey}"  

👉 Next, create the request, send it the url, format the data returned as json and print it out so we can see what we get back.  

import requests, json, os

newsKey = os.environ['newsapi']

country = "us"

url = f"https://newsapi.org/v2/top-headlines?country={country}&apiKey={newsKey}"

##### The new bit ###################

result = requests.get(url)

data = result.json()

print(json.dumps(data, indent=2))  

👉 Inspecting the data returned tells us what information we can extract. I'm going to print out just the titles, URL links, and content by using a loop.

For this, I've commented out the print(json.dumps(data, indent=2)) line so that it doesn't output everything. This is useful for testing, so I've just made it a comment instead of deleting it completely. 

import requests, json, os

newsKey = os.environ['newsapi']

country = "us"

url = f"https://newsapi.org/v2/top-headlines?country={country}&apiKey={newsKey}"

result = requests.get(url)

data = result.json()

# print(json.dumps(data, indent=2))

##### The new bit #####################

for article in data['articles']:

  print(article['title'])

  print(article['url'])

  print(article['content'])  

Sometimes, however, the content isn't great. See below: 

OpenAI

The second half of our mashup uses openAI

Again, you'll need to create an account, verify your email, and validate with your mobile number.

You do only get a certain amount of use out of the free trial. So if you use this a lot then it will run out.

OpenAI allows us to tell a computer to do something in plain text.

👉 There are quite a few steps involved in setup, so read these next instructions carefully.

Go to your profile and click view API Keys

Create a new secret key

Copy the API key

Add it as a secret to your repl called openai

Now get your organization ID (from the 'settings' menu on the left of the 'View API Keys' screen).

Make a new secret called organizationID for this.

Next, move all of your previous code into a new file called 'news.py'

Bring your 'openai' secrets into main.py  

import requests, json, os

openai = os.environ['openai']

orgid = os.environ['organizationID']  

Go to the documentation menu on the website and scroll down until you find authentication

Copy the Example with the openai Python package code:  

 

Paste the code from openai into main.py and combine it with the secrets to remove the extra os import and to take advantage of Replit not needing to install libraries.

import requests, json, os

import openai

openai.organization = os.environ['organizationID']

openai.api_key = os.environ['openai']

openai.Model.list() 

Completion

👉 The openai API Reference page shows me lots of things that I can do. I'm going to start with a completion. This example should talk to openai and say 'this is a test'. I've printed the response to show that happening. 

response = openai.Completion.create(model="text-davinci-002", prompt="Say this is a test", temperature=0, max_tokens=6)

print(response) 

When you run this, you'll see that it comes back in a top level of 'choices' and second level 'text'.

👉 Let's manipulate it to see if we can get it to do something else.

I've replaced the hard coded prompt with a variable.  

prompt = "Who is the most handsome bald man?"

response = openai.Completion.create(model="text-davinci-002", prompt=prompt, temperature=0, max_tokens=6)

print(response) 

Unfortunately, openai has no definitive answer for this question!  

👉 To just output the text (and strip out the odd white space) I've changed the print to this:

print(response["choices"][0]["text"].strip())  

Whole code  

import requests, json, os

import openai

openai.organization = os.environ['organizationID']

openai.api_key = os.environ['openai']

openai.Model.list()

prompt = "Who is the most handsome bald man?"

response = openai.Completion.create(model="text-davinci-002", prompt=prompt, temperature=0, max_tokens=6)

print(response["choices"][0]["text"].strip())