Top 100 podcasts API

Setup

• Open a terminal
• Preferably, activate a virtual environment
• Install the dependencies

cd podcasts-api/
pip3 install -r requirements.txt

In case you have trouble using requirements.txt, you can always install the modules individually:

pip3 install requests
pip3 install pandas
pip3 install SQLAlchemy
pip3 install flask
pip3 install flask_sqlalchemy
pip3 install flask_marshmallow
pip3 install marshmallow-sqlalchemy
pip3 install jwt

Running the code

Type in the terminal

python3 app.py

The first task that the code performs is the database creation. If it is the first time running the project, a message that says 'populating db...' will appear. When it its finished, it will display the message: 'db was populated succesfully!'

Create the username and password

You have to authenticate, in order to generate a token to use for calling the APIs.

• Open up postman or another REST client and use the endpoint: http://localhost:5000/user
• Define a name and a password like in the example

• If successful, a message with new user created! will get displayed.

Login

• Use the endpoint http://localhost:5000/login
• In Postman's Authorization tab, provide your username and password
• A token will get generated!

IMPORTANT: COPY THE PROVIDED TOKEN!

Using the token

• Create a x-access-token header, and paste the token you copied from last step.

Now you are ready to call the APIs!

Solutions used per exercise

Database and schema creation

Schema used for the exercise

I decided to split the data in three different tables: genre, podcast and genre_podcast, which is the middle table between the many to many relationship of genre and podcast. I have used pandas for this stage, given the simplicity for dropping duplicates and storing the dataframe directly to a table.

1. Retrieves the json from a remote URL
2. Separates the multivaluated column genres
3. Creates pandas dataframes of genre, podcast and genre_podcast
4. Create empty tables and their relationships.
5. Inserts dataframes into the tables.

Note: Given the fact SQL tables don't store the rows in insertion order, the pd index was used as a helper column to preserver the order. This index field is not visible when querying the data.

Exercise 1

Service to provide a search lookup within the podcasts

The function search_lookup implements the search and filters the podcasts table using a LIKE operator to search for a given name.

• use the endpoint http://localhost:5000/api
• use a POST request
• Send a json containing a name key and its value in string format

Exercise 2

Service that would allow to save the top 20 podcasts to a separate JSON File

The store_top_20 function uses the order_by function provided by flask-sqlalchemy to sort the podcasts using the index column.

• use the endpoint http://localhost:5000/api/top20
• use a GET request

Output got saved to json_oututs folder in project!

Exercise 3

• use the endpoint http://localhost:5000/api/swap
• use a GET request

A service to replace the top 20 podcasts for the bottom 20 to said JSON File

Using array slicing, its easy to swap the order of the rows in swap_top_bottom() function.

Output got saved to json_oututs folder in project!

Exercise 4

A service to remove a podcast, using a given identifier

• use the endpoint http://localhost:5000/api/360084272 (You can use any id of the podcast json)
• use a DELETE request.

The function delete_podcast(id) identifier used for deleting is 'id' (podcast id), given the fact it's been used as primary key of the podcasts table.

Excercise 5

Create a service to return the podcasts grouped by genres, even if they appear as duplicates within some of the categories.

• use the endpoint http://localhost:5000/api/grouped
• use a GET request

The podcasts_by_genres() performs a query to fetch the names of the genres and the id of the podcasts. Next, a for lop iterates over each row to perform a search of the podcast object and append it to a genres dictionary that wil return the desired output.