Refactoring A Tower Defense Game In Python // CODE ROAST
Vložit
- čas přidán 9. 07. 2024
- This video is a refactoring of a Tower Defense Game written in Python. In particular, I cover a few game architecture aspects and apply them to reduce the coupling in the example code. Get 20% off the Pro version of Tabnine, this video’s sponsor: bit.ly/ARJAN20. If you're a student, you can get Tabnine Pro for free here: www.tabnine.com/students.
The code I worked on in this example is available here: github.com/ArjanCodes/2021-co....
💡 Here's my FREE 7-step guide to help you consistently design great software: arjancodes.com/designguide.
🎓 Courses:
The Software Designer Mindset: www.arjancodes.com/mindset
The Software Designer Mindset Team Packages: www.arjancodes.com/sas
The Software Architect Mindset: Pre-register now! www.arjancodes.com/architect
Next Level Python: Become a Python Expert: www.arjancodes.com/next-level...
The 30-Day Design Challenge: www.arjancodes.com/30ddc
🛒 GEAR & RECOMMENDED BOOKS: kit.co/arjancodes.
🚀If you want to take a quantum leap in your software development career, check out my course The Software Design Mindset: www.arjancodes.com/mindset.
💬 Join my Discord server here: discord.arjan.codes
🐦Twitter: / arjancodes
🌍LinkedIn: / arjancodes
🕵Facebook: / arjancodes
👀 Channel code reviewer board:
- Yoriz
- Ryan Laursen
- Sybren A. Stüvel
🔖 Chapters:
0:00 Intro
0:47 Overview of the original code
3:39 Code analysis
7:38 A few quick minor fixes
14:34 About game engines
16:22 Creating a simple game engine
23:46 Game object structure
27:12 Creating the Tower Defense game subclass
31:28 Communication between game objects
35:57 Final thoughts
#arjancodes #softwaredesign #python
DISCLAIMER - The links in this description might be affiliate links. If you purchase a product or service through one of those links, I may receive a small commission. There is no additional charge to you. Thanks for supporting my channel so I can continue to provide you with free content each week!
This video was still a bit longer than I intended. I didn't really discuss this in the video, but if you know any good game engines for Python, feel free to share them here!
This is pure gold! This approach definitely brings out a different and interesting approach in talking about code. Thanks!
This video is the best python tutorial ever, imho.
Really, i wish it was even longer. This is to me a decoupling masterclass.
You are a gifted instructor and i felt like as it finished i was now getting into it more thoroughly and was waiting to see how you refactored the remaining globals. I felt that it would be as straightforward as adding them into the game function as a configuration with defaults since they will only have one instance.
What i love about this series is the application for things like event handlers, classifiers and abstractions in creating more general game classes. I am only a bit better than a beginner and I have a hard time imagining the efficient usecase for abstract classes like Enum even though i can understand how much more efficient it is to use it. I learn much more from this type of application even though i know it represents more work on your end. So thank you very much.
It is honestly some of the best instruction i have seen, with very calm delivery and overall i really love this series.
What i took away was these four things around decoupling:
1) Use gamestates or event handlers instead of global variables (i have a Tk project right now that could use this desperately)
2) Config is for when behavior is the same but attributes are different whereas inheritance is for when behavior is different
3) Reduce indentation level when possible by establishing basecases
4) Localize __name__ = '__main__'
The video could have been 2-4x longer and it still would have been engaging. You have a great pace and way of explaining decisions and tradeoffs that come up in program architecture and refactoring while not getting overly bogged down. Only engine I've encountered is Pygame/Pygin (Pygame Engine), but I haven't played with it.
This is gold. You're amazing. Seriously make a serious that puts you in position to be the "Decoupling Master."
Inspired to donate. Thanks for making these videos!
You're much, much more than that. I just use it as an example as I don't know if you realize just how special your content really actually is
Nothing more satisfying than deleting dozens of lines of code during a nice refactor. Warms the soul.
Haha, can’t agree more!
so true! last week at work I deleted I shaved like 20 or 30 unecessary lines from my API code and it's so much cleaner now
i refactored an old tower defense game once. it went from 2800 lines to 1500, i had to delete over one hundred lines at once
honestly cant believe i coded like that back then
Haha, so true...! I remember a couple of years back when working on with C. When I started working with this specific software the code base was one file with about 6000 lines of code, when I finished the whole code base was about 3000 lines and was using several different source-files... I felt so got damn pleased... :D
Dude what, I always feel bad deleting a lot of code, wasted time, good bye little code friends
I think a Game Engine Design series would be extremely valuable to learn about design patterns. I‘d even pay for that when done by you
Yandere dev code review is really great if you like this type of content
I second that .
This 100%
Game engines from scratch are absolutely gigantic in scope.
I recommend a channel called The Cherno then
These are so great. As a self-taught dev who is always striving to improve, it's incredibly helpful to see "working" code being refined like this.
Thanks, glad you liked it!
Haven't watched the whole thing but at 9:55 you actually change the meaning of the code in the update function and do the exact opposite. I hope you catch that later on!
Well spotted! Indeed the ‘not’ should not be there. In the final version that’s on Git, I think I actually used the game state for this.
Came looking for this comment :)
There are dozens of us. DOZENS!
@@BasicallySteve23 did you also watch to the end to see if he mentions it himself before going to the comments?
@@DavidvanDeijk I did not ;)
I love this Code Roast format.
Seasoned C++ dev here. Came here to see some real life Python, stayed for the refactoring and showing the value of good architecture. Great work! I second all the comments about a game engine series :)
This is so useful. More videos with detailed code reviews like this please. I think you might have found an interesting niche.
Thanks Jason, glad you like it and will do!
A game engine series would be great. Game engines cover so much ground that once you're comfortable with them 'Enterprise Software' seems like a walk in the park.
Fully agree!
I also agree, games are inherently complex and lots of things happen at once. Lots of programming tutorials have toy examples, but with a working game examples are closer to the "real thing" and also easier to teach.
A video series on game development with python would be fantastic!
I’ve had some successes with pygame, but I’ve seen nothing on how to handle player levels and what in some other games might constitutes evolutions, upgrades, or tiers). The same goes for keeping player movement fluid and camera tracking.
Happy to dive into it on your dedicated Discord channel!
Thanks and noted!
@@ArjanCodes Indeed, it would be valuable to me since I'm currently working on a simple space shooter on pygame as my first real project in it. Right now: I'm focusing on improving the structure, it's still probably not as good as it could be, but at least it's way better than having all sorts of classes in a single python file, making it a complete nightmare to navigate through said script and find what you want.
This is the my favourite series! I always has trouble "reading" others code as a beginner.
Being able to have an expert explain the reading process by cutting all parts apart is very, very helpful!
I appreciated the positive attitude in the "Overview" section
You know what would be cool? A short playlist for building games. A rewrite the 1980s might be interesting: Making a text based game (e.g. Zork), top down rpg (e.g. Dragon Quest or Legend of Zelda), and a side scrolling platformer.
Great suggestion, thank you!
I"m a JS/TS/C# developer and don't even know how to write Python but these principles seem to be language agnostic. Very pleasing to watch!
Also a "roast" implies Java, clearly!
The game engine’s Update() and Paint() are concepts nicely covered by The Coding Train, Coding Math, and One Lone Coder who makes a cpp game engine. I think Decoupling is the most valuable lesson because of the complexity, size and typical “need” of globals from beginners.
I would absolutely love to see your take on more gaming, complicated, types of projects. Thanks for the walk through of this one! And thanks/great job to the person who submitted their code for review!
Everything you said in 14:34, in the "About game engines" section, I am trying to juggle in my own game. Especially coupling and event handling. If you could make a video dealing with these, or a whole series about game engines, it would be so helpful.
Your videos have been a great source of help thus far. Thanks for the great content, Arjan!
Thank you and you’re welcome!
@@Borgilian Modernization happens for a reason. Even with low level programming you'll still want to decouple your code as much as possible. The irony of your statement is that these "experienced" programmers all retire in the next decade and all of the forward thinking people will have the tools to use their "high level" languages in virtually the same capacity as a low level language with some outlier exceptions.
Accept forward change, or be left in the dust.
The fact that a basic python refactor video has triggered you this hard speaks bundles about your state of mind. Go get help, you need to be medicated (I mean this with no sarcasm, it's abnormal).
@@Borgilian Woah... easy on the scare quotes son.
Are you contending that our host is a crappy programmer? And if so, on what grounds?
Loved the last video, was genuinely excited when I saw this one - keep them coming!
Thanks, glad you liked it!
Arjan makes these coding videos as exciting as they can be. This is amazing.
Really nice of you to not just flame the hell outta the programmer but positively criticise the code!
This is huge! I am amazed at how much value just this one video provides. First of this kind of series I've come across. It's game changing and a lot more helpful than the typical tutorial or language feature review channels. Really appreciate it!
Great to hear & glad this is helpful to you!
Just stumbled upon this video, and I gotta say that this is SO HELPFUL! Hope you get your 100k subs soon and many more!!
Thank you so much!
Glad you liked it! ☺️
This code roast series the best! I just subscribed. Thank you for these interesting and professional videos.
Awesome video. Seeing refactoring a real world example is priceless. I'm looking forward to the continuation of this.
Oh my gosh yes! A code review and improvement! More of these please if it is possible !!!
Arjan Thank you for doing this series. It's fun to watch and your explanations are easy to understand. I'm learning quite a bit from it.
Well done! Really enjoyed watching the refactor and hearing your explanations. Looking forward to more - keep up the effort!
“If not done” is the same as “if done == False”, so you didn’t change the condition and flipped the logic to run the block if done instead of doing nothing if done.
yeah, I've noticed that one as well.
@@Gunth0r me too...at 9:45 for those who wonder
That's why we have unit tests, and run them as we refactor.
I was yelling at my screen just now... 😅
I almost questioned my existence when I saw that
Nice! Thanks for doing these! I'm getting ready to do a massive overhaul myself of a Python project. It's all research code doing some mathy things written by a bunch of people of differing experience levels. This provides me with some motivation and good ideas.
This is awesome! As a Jr. engineer seeing you refactor a decent size project is massively useful. Your refactoring of his if statements was a real eye opener for me.
Thanks, glad you liked it!
Hey Arjan, didn't know you made videos like this! Absolutely awesome. I can see that being helpful to many.
The only thing I'd recommend doing (or rather would recommend to anyone else attempting a big refactor) is to frequently test that things still work. It wouldn't be the first time I've gotten ahead of myself, only to have a hard time figuring out where in the process I broke stuff.
Subscribed :) Keep it up!
It might be hard to find a well unit-tested project to refactor, but if you do: I would love to see a continual-test refactoring.
That was amazing!
Please make a series about game dev!
Thank you Dimas, glad you liked it!
I've been trying to get better at python for a while now and I am always amazed when I learn something cool that I wish I knew about before. like at 11:02 where you show how python supports more complex comparisons like xmin < x < xmax.
Thank you Arjan. Love your voice really calm and easy to listen to.
Learned a lot. Looking forward to other refactoring for something maybe not game related. (Web Development, Bigger Backend, Service)
Thank you for this really practical example. Thank you Arjan and Gooshrlific so much
This is great, I think the guy who sent this code in realized how poorly organized it was and knew that a lot of people can learn from his early mistakes! Great job!
Great video as always, I always feel like I learn a lot (thanks to you I learned about type hinting), would’ve liked to see you test the game was still working after all the refactoring though! 😁
Cannot care less about game development but you're such a good teacher that made this so interesting! Thank you for these videos, Arjan! Again, groetjes uit Den Haag :P
Thanks Alessandro!
This video is amazingly helpful as someone relatively new to coding looking for ways to improve. Even though I have no intention to make any games, I can see how these tips can be applied in my own projects. Thank you very much for this.
Thank you, glad you liked it!
Can you explain Protocols more? I'm not seeing much information about them when I search and I didn't understand the benefit in this video. Thanks!
Absolutely! There’s a video coming very soon where I spend more time on Protocol classes.
I found your channel because of this video. Keep up the good work; I really enjoy your content!
Thanks Andrew, glad you like the content!
I just found your channel today. Great videos! I agree with the past commenters, a game engine series would be very beneficial. You're an exceptional instructor.
This video was awesome. Please finish the refactor in a part 2. I need a tutor like you to teach me design patterns and strategy.
Thank you, Jacin!
Great to see how it's properly done!
It will be very interesting and informative to see a video covering game engines from your experience. Thanks for sharing you knowledge. Im particularly learning a lot to code in an efficient and ordered way fron your videos. Thanks again. 👍
Hi Eduardo, you’re most welcome - I’m happy the videos are helping you.
I haven't watched this video yet but my god, this format might have finally discovered an actual way to present code on CZcams in a way that isn't mid-numbingly boring.
I loved your explanation of including a game state which can trigger certain dependant events. I've been working a lot with React lately, so that really connected the web dev and game dev worlds for me. I'd love to see a game engine series from you, I bet there are many more design patterns that I'm overlooking
Game Engine and Framework design would be a great series. Even if it's talked above my head, seeing how more professional coders tackle problems can be in the back of my head when hobby shopping some of my own code
I would love to see you teaching design patterns in game dev
You are a very good teacher! Keep up the good work!
You might also consider streaming a full refactor. I would totally watch it!
Very nice video! I have not coded in python in years but the concepts are universal and well explained, plus I enjoyed the 'real time' feeling
This is a super nice format! Got this in my recommended feed after watching your exception videos.
What I really enjoyed the most was the explanations. I do developer support in my team and I feel that succinctly explaining why something is not a good idea is where I struggle the most. This really helped in pointing out what I do wrong. Also, it was very entertaining :)
Thanks for the great content!
Thanks, glad to hear you enjoyed this. I have more code roasts in the pipeline.
These videos are amazing! The more the merrier of code roasts
Thanks! Glad you like the videos.
I *really* appreciate you checking the cringe response at the beginning. I get so tired of this attitude out in the wild.
Completely agree - I'm realising I'm guilty of this with newer Devs, but remembering that they are learning and that I am also still learning makes a massive difference. Currently trying to therefore make the focus of my code reviews and pair programming focus on how they grow in knowledge, not just do we end up with code that I like.
Wow, just reading how self-aggrandising that is when I say/type it out loud...
@@nathanbrown2640 It’s not self aggrandizing to know you’re good at what you do. 😉
As a new dev who is entering the industry with a lot of prior programming experience, I am learning new things every day. But one thing that confuses me still is this. I was given the assignment to load and save to a file format called HDF5. The api to interact with the library distributed by the developers of the format is horribly unintuitive so I made a series of abstractions to make it much easier for my serializer code to work with so as to not repeat code that could potentially do horrible things like buffer overflow. The whole abstraction set was only about 1.5k lines and I set it up to be extensible if we ever need more functionality from hdf5 in the future, you could just derive a class, slightly modify the functionality, and use it. During my code review to check in the entire serializer, my reviewer had issues with it because the abstractions were relatively complex and apparently all I needed to do was parse the format. So I ended up spending hours explaining the abstractions I made and why only to eventually have it accepted. But the main complaint I got is that I an thinking about a million things at once. Id be talking about one thing and Id say I made a different choice because it was faster, or I made a different choice because I wanted the abstractions to ensure there were no buffer overflows rather than the user of the library. Then when I add failsafes to the abstractions like bounds checking, I had to fight hard to get that accepted because apparently "the way I use the library in the serializer that condition wont be hit." The thing is, I wrote it such that I wanted arbitrary code to be able to use it safely so I dont understand why bounds checking is something I have to fight for. Oh well, sorry about the rant, I just dont get that about the industry so far. Apparently its bad to think ahead and cover all my bases because it makes my code harder to explain.
Edit: I guess I just want each function to be correct for any given set of parameters unless the caller 100 percent knows that their parameters have to be checked before the function call. In the case of my code, they were checked before the call, but that doesnt stop an arbitrary caller from giving it bad parameters which could break the code. So I only exclude those checks if the code was called by an internal caller which verified the parameters for it
@@KayOScode Sounds like you might be over-engineering your code. Whilst it's great to make your code fail safe and extensible, you have to be mindful of how long this extra work takes and whether the time would be better spent elsewhere. Also, it will be harder for someone else to read and maintain your code.
@@DrewCocker you might be right. Before I started industry work, I just wrote the code and pushed without much thought. It led to simpler code than that pull request. Maybe theres value in simplicity. Something about doing it for a job makes me want to write perfect code
I would like to praise youtube algo for finally showing me an usefull recomendation
It may be long, but you made it worth watching every minute!
Thanks Andrew, glad you liked it!
These series are very instructive and well structured. Thank you so much. BTW you have a video about your stack in NodeJS and TypeScript. It would be nice if you compare them with Python from your experience in web development. your channel deserves and gets 100k soon
Thanks Omid, great suggestion!
This was great. Thank you!
Already at 6 mins into this video I've learned some great tips on how to think more like a professional developer - great vid!
Glad it was helpful!
Love the code roast format. Very enjoyable
I've learnt a lot from this video alone, thank you for this. Also, i'd love those videos about creating a game engine!
This video is a great example why you should not manually refactor without tests: The error around the 9:40 mark that others already pointed out and the changed behavior after deleting the try/catch statements later on. You made an implicit Finally of the code that was after this block, while it was originally only hit in the catch.
Write tests first, refactor later. You don't need to create the tests yourself: there are unit test generators and in combination with code coverage tooling you can see if every line is covered (or at least the lines you want to refactor.)
Speaking about tooling; refactoring can (and should) be done with tools too. All IDEs support refactoring patterns (like the switch conditional) that can be used while guaranteeing not to break anything. You can even go wilder with tools like Sourcery that have very clever refactorings.
While it is good to do refactorings by hand (to learn how they work and why they exist), don't do this with production code without tools or tests.
I enjoyed this, and that was some really effective and satisfying refactoring. You mentioned that people are at different stages on their learning journey. I think its true to say the same of projects and code. The game you started with reached the working prototype stage in its journey. It was cobbled together, iterated on a little - all focusing on the game play. It seemed to play well and if the goal was to accomplish that, then it was a success and "good" code by that definition. To allow it to evolve further, maybe build some more levels, or monsters etc, or to stand as an open source example to learn and build on - then it needs to move further on its journey with the kind of refactoring you did.
When inspiration strikes and you want to make something from nothing, taking the most direct route from start to something that works is absolutely the right decision. Its usually impossible to know what the right structure is until you've got something that maybe almost works and the problems and their solutions start to emerge. And then, when it does work, don't stop just yet. You push on a little further to clean things up once the shape and requirements have started to become clear. I've found good code is a result of a series of discovery, testing and refinement loops. With experience you can sometimes skip some steps, but when you are learning, this process can be opaque - as people usually only commit and share the end product, not the indirect path they took to get there.
Thanks again for such a well presented example of taking some code a little further on its journey, to become something we all can learn from.
You can learn more from this 30 minute video than in hours of online lessons/classes. This started to feel like Unity after you were done. Only comment is I like the subclass for Monsters, even though it was just different attributes, is probably a good forward looking design.
Very cool video! I wasn't expecting the video to be so constructive because of the name but you spent most of the time helping and giving advices! Not really a roast but good content!
Love these refactoring videos. I find them very informative. "Here's the way you probably wrote your code, Here's how not to do it stupidly"
Thank you, glad you like them!
That's some confident way of working with legacy code if I've ever seem one. I think I would try hard to back all my changes up with unittests before I'd even get going on the refactoring.
Learn a lot, keep going!
Finally, ASMR for getting better at coding
I would love a game engine series. Also, amazing videos. Keep the great work. Cheers from Portugal
w8ing for next part. and game engine videos. subbed
Would love to see a game design series!!!❤
You really know what you're talking about, great vid!
This is just incredible
Arjan,
Before coming to the point of this comment, I would like to compliment you on the style and the information in most of your videos. It is interesting and informative to see a developer's viewpoint on software (especially his/her viewpoint on the underlying code).
Now for the constructive criticism; You mention that you would not be refactoring the 'Tower Defense' completely due to time constraints. I would like to point out that not all of your viewers have an attention span of only a few minutes.
In fact, I regularly skip any video under 15 minutes that I would view to learn from. The previous refactoring video was very informative in its completeness.
Thank you for this inside look into the steps used by a professional software developer.
Hi John, thanks for your comment - much appreciated! I've been playing around with the format of these code roast videos. In more recent episodes, I've stuck to longer roasts, but split into a few different parts. I think that's a reasonable balance between keeping the videos to the point, but still providing a full refactor.
Hi , just discovered your channel . Just a thought that what do you generally look for in refactoring?
Sounds like there's a ton of value in generalizing commands instead of directly referencing individual events. I have been interested in learning about coding for so long now but I get intimidated by how specific everything seems to be, but seeing this really puts it into a different perspective.
Knowing how to code is a really valuable skill to have, I highly recommend you dive into it! Start with a simple project that's fun to work on. For example, you could create a classic text-based adventure game, or write a small application that does currency conversion. From there on, introduce new things to broaden your skillset.
Yeah, architecture for a game engine or some sort of simulation environment would be really awesome to see from you!
YES PLEASE! Game engines and GameDev architectures! Definitely the architectures
This is a really useful and fun series. As you mentioned in the video, please consider doing a series on game engines! Great work.
Thanks for the suggestions, Tamas, I've put it on the list.
Good Video : ) I would like to see more of this series. You can implement one of the design principles you created a video about and reference that in the same video. I started reading SOLID principle books since I started watching your videos about them. But maybe seeing someone thinking about what design principle they should apply when refactoring might improve the series.
This type of Video is great and super educational
Thank you - glad it's helpful!
youtube algorythm now recommends me Refactorigs xd
I love these. Holy moly.
Thank you Josh, glad you liked it!
this is great, totally going to send in my noob project when i am done
I'm learning so much, thank you for doing god's work!
Thank you, I’m happy the videos are helping you!
Very instructive video, per the usual. Thanks for sharing. I second the idea of a video on game engine development.
Thanks Christian!
I would very much so be interested in more videos about game engine architecture.
Totally on with this game design pattern serie !
Really interests me since i'm trying to create a game from scratch (without a game engine), in a python-like language.
Also would like to ear if you have some suggestions about a functionnal approach to game design. Most of the patterns I see on the web are about OO and not FP.
Yes! Make a game programming series please!
Hell yeah, I would watch a video on game engine architecture. I contribute to multiple open source games based on just SDL and architecture is the hardest thing to get right in projects of some size.
This is wonderful content, thank you so much. It would be very helpful if you say how much time it takes you to go through the code, understand and then refactor the code so we have somewhat of a realistic timeline
Amazing, even if i dont code games, ive learnd a lot from this!
Thanks Tobi, glad to hear it was useful!
That's actually very interesting.
Of course, the refacto is very far from finished, but you introduced the different concepts mandatory for a proper game engine.
Also I feel like you started with C or C++ looking how you code is definitely refreshing.
Cause I definitely feel like my eyes are burning whenever I look at some python, but not here.
Thanks, glad you liked it! Actually, I started out in Basic but I'm quite happy you don't see that back in the code, haha. After Basic, I did move on pretty quickly to C++ though.
Please make more videos like this man! Would also be cool to watch you build an application from scratch! Like a web app with mongo or sqlite.
Maybe I don't understand protocols but didn't you almost completely remove the implementations of the update and paint methods? Or did I miss something?
I would love to see more of this kind of video... I don't use Python3 for making games but I learned both game-design and general Python coding tricks. Do you plan on making more of these refactor videos?
Your coding is way too complicated for me to fully understand it now, but I can still learn much from your videos. I hope one day I will get everything you say!
Hi Arjan. I would love for you to go over best practices while creating games in Pygames. It would be awesome if you could plan a series regarding this in the future.
Thanks for the idea!
I think a video game design series would be an awesome thing to watch
Thanks Liam, noted!