Build Your Own Wordle in Bash with the Data API
Rate this code example
I wanted to brush up on my Bash scripting skills, so I thought, “Why not create the Wordle game in Bash?” I figured this would be a good exercise that would include some
ifstatements and loops. However, the word list I have available for the possible Wordles is in a MongoDB collection. Well, thanks to the new Atlas Data API, I can now connect to my MongoDB database directly from a Bash script.
Let’s get to work!
The game will run inside a while loop that will accept user inputs. The loop will go on until either the user finds the right word or has reached five tries without finding the right word.
First, we’ll start by creating a variable that will hold the word that needs to be guessed by the user. In Bash, you don’t need to initialize variables; you can simply assign a value to it. To access the variable, you use the dollar sign followed by the variable's name.
Next up, we will need a game loop. In Bash, a
whileloop uses the following syntax.
Finally, we will also need an if statement to compare the word. The syntax for
ifin Bash is as follows.
To get started with the game, we will create a variable for the while condition, ask the user for input with the
readcommand, and exit if the user input matches the word we have hard-coded.
Save this code in a file called
wordle.sh, set the execute permission on the file, and then run it.
So far, so good; we now have a loop that users can only exit if they find the right word. Let’s now make sure that they can only have five guesses. To do so, we will use a variable called TRIES, which will be incremented using
exprat every guess. If it reaches five, then we change the value of the GO_ON variable to stop the main loop.
Let’s now compare the value that we got from the user and compare it with the word. Because we want the coloured squares, we will need to compare the two words letter by letter. We will use a for loop and use the index
iof the character we want to compare. For loops in Bash have the following syntax.
We will start with an empty
STATEvariable for our round result. We will add a green square for each letter if it’s a match, a yellow square if the letter exists elsewhere, or a black square if it’s not part of the solution. Add the following block after the
readline and before the
Note how we then output the five squares using the
echocommand. This output will tell the user how close they are to finding the solution.
We have a largely working game already, and you can run it to see it in action. The only major problem left now is that the comparison is case-sensitive. To fix this issue, we can transform the user input into uppercase before starting the comparison. We can achieve this with a tool called
awkthat is frequently used to manipulate text in Bash. Right after the
readline, and before we initialize the empty STATE variable, add the following line to uppercase the user input.
That’s it; we now have a fully working Wordle clone.
We now have a fully working game, but it always uses the same start word. In order for our application to use a random word, we will start by populating our database with a list of words, and then pick one randomly from that collection.
When working with MongoDB Atlas, I usually use the native driver available for the programming language I’m using. Unfortunately, no native drivers exist for Bash. That does not mean we can’t access the data, though. We can use curl (or another command-line tool to transfer data) to access a MongoDB collection using the new Data API.
Let’s start with adding a single word to our
wordscollection, in the
wordledatabase. Each document will have a single field named
word, which will contain one of the possible Wordles. To add this document, we will use the
insertOneendpoint of the Data API.
Create a file called
insert_words.sh. In that file, create three variables that will hold the URL endpoint, the API key to access the data API, and the cluster name.
Next, use a curl command to insert a single document. As part of the payload for this request, you will add your document, which, in this case, is a JSON object with the word “MONGO.” Add the following to the
Running this file, you should see a result similar to
This tells you that the insert was successful, and that the new document has this
You can add more words to the list, or you can import the official list of words to your MongoDB cluster. You can find that list in the
words.jsonfile in this project’s . You can change the
insert_words.shscript to use the raw content from Github to import all the possible Wordles at once with the following curl command. This command will use the
insertManyendpoint to insert the array of documents from Github.
Now back to the
wordle.shfile, add two variables that will hold the URL endpoint, the API key to access the data API, and cluster name at the top of the file.
Next, we’ll use a curl command to run an aggregation pipeline on our Wordle database. This aggregation pipeline will use the
$samplestage to return one random word. The curl result will then be piped into
jq, a tool to extract JSON data from the command line. Jq will return the actual value for the
wordfield in the document we get from the aggregation pipeline. All of this is then assigned to the WORD variable.
Right after the two new variables, you can add this code.
And that’s it! Now, each time you run the
wordle.shfile, you will get to try out a new word.
That’s it! You now have your very own version of Wordle so that you can practice over and over directly in your favorite terminal. This version only misses one feature if you’re up to a challenge. At the moment, any five letters are accepted as input. Why don’t you add a validation step so that any word input by the user must have a match in the collection of valid words? You could do this with the help of the data API again. Don’t forget to submit a pull request to the repository if you manage to do it!