IAmEddieDean/autodesk_challenge

a fun task

             // | |                                                               
            //__| |            __  ___  ___      ___   /  ___      ___     / ___  
           / ___  |   //   / /  / /   //   ) ) //   ) / //___) ) ((   ) ) //\ \   
          //    | |  //   / /  / /   //   / / //   / / //         \ \    //  \ \  
         //     | | ((___( (  / /   ((___/ / ((___/ / ((____   //   ) ) //    \ \


              //   ) )                                                            
             //        / __      ___     // //  ___       __      ___      ___    
            //        //   ) ) //   ) ) // // //___) ) //   ) ) //   ) ) //___) )
           //        //   / / //   / / // // //       //   / / ((___/ / //        
          ((____/ / //   / / ((___( ( // // ((____   //   / /   //__   ((____     

Thanks for a fun and demanding code challenge! A couple quick things to note:

• I'm not certain this does what it's supposed to do, more below
• I'm not skilled with Python or Docker/Compose... yet
• I am an extremely curious person by nature, so the science-y bits were an additional bit of fun, though frequently perplexing

Setup

Assuming you have git and docker/compose on your machine -

Clone this repo

git clone https://github.com/IAmEddieDean/autodesk_challenge.git

Run

docker-compose up

Go to http://localhost:3000 in your browser - (optional: marvel at my lack of design skills)

What it does (and hopefully it's the right thing)

Upon entering a snippet of a DNA sequence into the text field, a node process sends a request to a python service. This python service creates a Seq object from the provided snippet and compares it to the list of available genomes. If a match is found, the service returns a simple JSON object containing the id of the organism, the organism's long name, and the index of the genome where the beginning of the match is found. Node takes this object and stores it in redis as a string, which can be parsed later. Once the node process responds to the frontend, the table is refreshed and the new data is seen as a new row.

Comments on the challenge

Without pointing out too much in a public readme, there were a couple flaws in the challenge prompt:

• The provided list are all complete genomes, not proteins
• The above bullet makes it pretty challenging to figure out what you should do, and even after a clarification over email, I was still perhaps confused... I guess that's up to you to decide.
• Hopefully redis will serve the purpose of data retention in this case, as the prompt does mention this is to be treated as a PoC. Besides, I love redis.
• I did not go through the trouble of managing concurrent requests in python... sorry. I wasn't ignoring the challenge prompt, it simply wasn't necessary (in my humble opinion,) for this first pass. Not only that, I don't know what framework - if any - Autodesk is using in python, and they all do it a bit differently.

In closing

Thank you very much for this fun and - for me - unique challenge. I sincerely hope it's up to snuff (enough) to warrant an in-office interview. - Chris

Updates!!! Improvements!!!

I'll be honest with you, I was annoyed that I didn't manage concurrency with the python application. To that end, I did a little reading on gunicorn, which seems to be a very popular way to manage additional workers in the synchronous world of python. You'll now find that the python app has two workers on startup. Not only that, I removed the public port for that container, making it only accessible within the network docker creates.