A personal project to generate lists of names that exist with identical spelling in different spoken languages.
The idea to do this came to me while searching for my first junior dev role when two friends who were expecting a baby, a Spaniard and Frenchman, said they had chosen the name for their child, but that they wouldn't reveal it before the birth. Their only hint was that it was spelt the same in both French and Spanish...
- Simply fork or clone this repo, and run
npm install - Run
npm start, open a browser and navigate to http://localhost:3000/ - Enter the names of two countries (not their languages - i.e. 'France', not 'French') in the boxes, and hit the search button
- After a few seconds, an alphabetized list of matches should appear below, along with a tally for how many have been found
- Tests can be run by running
npm testin the main directory
You can also skip all the above and use a verison of the app deployed to Heroku at https://cryptic-hamlet-10312.herokuapp.com
The final version allows a user to compare names from any two countries provided for by the thomas-bayer.com API (around two dozen or so). So not just France and Spain. The full list of supported countries can be found here.
- JavaScript/Node.js
- Express
- Jasmine/Mocha/Zombie/Chai
- HTTP requests
- Parsing XML
- Promises
So apart from entertaining myself, I decided to undertake this little project a) because I'd just finished writing something else in Node.js and wanted to do more, and b) I saw it would present opportunities for me to improve my understanding of how to deal with asynchronicity in my code.
Writing tests as I went, I started by putting together the simplest object in the entire app - the 'Comparer', which takes two lists of strings and returns a third list made up of matching strings that are present in both of them.
Next came the 'Scraper', originally named this because I initially decided to make HTTP requests to regular baby name websites, then parse their HTML and extract the names. It was difficult finding websites with significant numbers of names on a single page, and those that I did find had deeply complex and confusing structures - one, for instance, storing hundreds of names in giant strings in each individual cell in a table. After some research, I found a nice REST API that served the same data up in a cleaner XML format.
To navigate through the data, my 'ListBuilder' object uses the xml2js library, returning an array of the names from the XML it is passed.
At this stage, I was passing my Scraper the ListBuilder's 'listNames' function as a callback, in order to get the latter to run after the HTTP response data had arrived. This didn't feel too clever, and the idea of sequencing a long chain of callback functions began to look increasingly horrific. I had a dull idea that promises were the solution - I had used those emitted by Node's http.get function, but had never written my own. After a lot of reading, video-watching and experimenting, I hit on a solution that would run both API queries in parallel and (using Promise.all), and, after both had succeeded, run a series of functions with the resulting data. Perfect, et voila, my list appeared!
UPDATE:
A couple of months after building the back-end to this project, I came back to it and built a (very) basic front-end in Express, with a single view allowing users to enter the names of two countries in a form, and then iterating through the response data to display a bullet-pointed list of matching names underneath.
PS. In the end, they went for 'Felix'.
Not much, except to say that while my API (http://www.thomas-bayer.com/restnames/) has great data, it is slow, slow, slow - sometimes a full response can take longer than a minute - and sometimes (like right now, as I write this) it's just completely offline.
Additionally - I believe I can blame this on the API data - some accented characters are rendered as question marks.