AI & Data

Introduction to Web APIs

• API: Application Programming Interface.
• Analogy: A menu in a restaurant, defining operations.
• Web API: An API accessed over the web (HTTP).
• REST (REpresentational State Transfer) is a popular style for designing networked applications.
• We use REST APIs to fetch data programmatically.

How REST APIs Work

• Resources: Data entities (e.g., users, posts).
• Endpoints: URLs that represent resources.
  • e.g., https://api.example.com/users
• HTTP Methods (Verbs): Actions on resources.
  • GET: Retrieve data.
  • POST: Create new data.
  • PUT: Update existing data.
  • DELETE: Remove data.
• Our focus will be on GET for data retrieval.

Making Your First API Request

We use the requests library in Python.

import requests

# Define the API endpoint URL
url = "https://jsonplaceholder.typicode.com/posts/1"

# Send a GET request
response = requests.get(url)

# The response object contains server's response
print(response.status_code)  # e.g., 200
print(response.json())       # The data in JSON format
    

Understanding the Response

• HTTP Status Codes:
  • 200 OK: The request was successful.
  • 404 Not Found: The requested resource could not be found.
  • 401 Unauthorized: Authentication is required.
  • 403 Forbidden: You don't have permission to access this.
  • 500 Internal Server Error: A problem on the server.
• Response Body: The actual data, often in JSON format (key-value pairs).

Exercise 1: Simple GET Request

• Goal: Fetch a list of users from a public API.
• API Endpoint: https://jsonplaceholder.typicode.com/users
• Task:
  1. Write a Python script using requests.
  2. Make a GET request to the endpoint.
  3. Check if the request was successful (status code 200).
  4. Print the name of the first user in the list.

Working with API Parameters

• Path Parameters: Part of the URL path to identify a specific resource.
  • e.g., /users/1 to get user with ID 1.
• Query Parameters: Key-value pairs at the end of the URL to filter or sort.
  • e.g., /posts?userId=1

Passing query parameters with requests:

params = {'userId': 1}
response = requests.get('.../posts', params=params)
print(response.url) # See the final URL
    

Exercise 2: Filtering API Results

• Goal: Fetch comments for a specific post.
• API Endpoint: https://jsonplaceholder.typicode.com/comments
• Task:
  1. Use query parameters to get comments for postId=1.
  2. Write a Python script to perform this request.
  3. Print the number of comments received.
  4. Print the email of the first commenter.

Handling Authentication

• Why? To identify the client, for security, and for rate limiting.
• Common methods:
  • API Keys: A unique string sent with each request.
  • Bearer Tokens (OAuth): For delegated access.
• Sending a key in the headers is a common practice:

headers = {'Authorization': 'Bearer YOUR_API_KEY'}
response = requests.get(url, headers=headers)
    

Exercise 3: Error Handling

• Goal: Write robust code that handles potential API errors.
• A request can fail (network issue, server error, etc.).
• Task:
  1. Request a non-existent post (e.g., /posts/99999).
  2. Check the status_code to see if it's 200.
  3. If not 200, print an informative error message.
  4. Wrap the request in a try...except block to catch network errors.

Integrating with Orange Data Mining

• Goal: Convert API data into an Orange Data Table.
• We can use a Python Script widget to run our code.
• The final result must be an Orange.data.Table assigned to out_data.

from Orange.data import Table, Domain, ContinuousVariable, StringVariable
import requests

# 1. Fetch data from API
users_data = requests.get('.../users').json()

# 2. Define the domain (the columns)
domain = Domain([
    StringVariable('name'),
    StringVariable('email'),
    StringVariable('city')
])

# 3. Create the data table
data = [[u['name'], u['email'], u['address']['city']] for u in users_data]
out_data = Table.from_list(domain, data)
    

Final Exercise: World Countries Data

• Goal: Fetch real-world country data and visualize it.
• API: We will use the public RestCountries API.
• Endpoint: https://restcountries.com/v3.1/all
• Task in an Orange Python Script widget:
  1. Fetch the data from the endpoint.
  2. Define a domain with `name`, `region`, `population`, `area`.
  3. Parse the JSON and build the data list for the Orange Table.
  4. Assign the result to out_data.
  5. Bonus: Connect the output to a GeoMap widget and visualize the countries!