Dependency Injection Explained Like You’re 5 (with FastAPI Examples)

Learning dependency injection can be hard. In this simple FastAPI tutorial we will go over dependency injection and that Depends Keyword.

🎓 Courses can be found:

👨‍💻 Hi, I am Eric!👨‍💻
Hey! I’m a seasoned developer and Cloud enthusiast. I love sharing my knowledge and helping others to discover the exciting world of programming and cloud technologies. For more tutorials, tips, and tech magic, don’t forget to hit that subscribe button and ring the bell for notifications!

🙏 Support The Channel 🙏
If you found this tutorial helpful, please give it a thumbs up 👍, comment, and share it with your network 🌍

#fastapi #dependencyinjection #python

[tarangsuri]

but its just a function calling? what diffrence will it make if i call those function that return objects with in the main function, whats the big diffrence?

[calebmatthias]

This is so far the best video I have seen on this subject! Thank you for the post

[ssswdon]

Nice video, but I think the pattern here is IOC(inversion of control)

[inhbtnt]

Hi Eric! I have a question for clarification - so if our function (or class) depends on some object that's already created (like Logging) - we should use DI, but if we need to create a new object every time (e.g. some context manager) - we would use composition (pass a class as an argument and then create new object every time) - is this correct?

[mantle_bearer]

Well explained and thanks for the detailed explanation.

Please between using a decorator function and dependency injection which would you recommend for user authorization?

[vikrantwangal]

If we create an object in another module and import it and use it then will be efficient than DI?

[CodePythonPro]

Great video Eric, I hope to built videos like this soon (to help python community in Spanish), thanks for sharing your knowledge.

[ConfirmedPlayer]

Why can't you just declare the class outside the function (router)? It will be initialized only once, like with dependency injection. I still don't realize the difference.

[BritScientist]

Thank You. Why don't you organize you code in `__init__()` in your classes without creating unnecessary functions next to classes? Dependency injection in FastAPI requires callable, so it may be a class, not only function.

[farazahmed]

which theme are you using in this tutorial?

[hipolitoribeirocavalcantej]

Maravilha!! Ficou muito claro com exemplos práticos! Parabens!

[AGunzOw]

still don't see the point why moving the instantiation onto input param is a better thing

[Wolfci]

More of these please lol just subbed great explanation

[codingwithroby]

There it is folks, dependency injection. What’s your biggest challenge as a developer right now?

[ankitmoura]

ok so whole point is to save resources and not create new instance on every api call ryt? but like when we create a new function which returns the new instance for us won't that fn be called on every api call and return a new instance i might be wrong but it feels like we r doing same thing with extra steps help?. also great content

[wait]

Could you make a video on how to use dependency injection without fastapi?

[dcgrc]

Beginner here. Trying to understand how it's different from just doing a "def log_message(message, logger=get_logger())"

[shafiq_ramli]

For once I thought DI is a security vulnerability like sqli 😅

[ZumaTech]

Start from 4:49, before it's just bullshit to make video longer

[awakenwithoutcoffee]

Hey Eric, thanks for the awesome lesson! we encountered an issue with dependency injection naming (?) :
"EmailResponse = Annotated[EmailService, Depends(get_email_service)]" -> WORKS,
email_response_dependency = Annotated[EmailService, Depends(get_email_service)] -> ERROR:

AI Response which we could use feedback on as you don't seem to have this issue:
" It fails with "Variable not allowed in type expression" because Python's type system interprets names with lowercase and underscores as variables, not as type names. In type annotations, only proper type names (typically PascalCase) are allowed as aliases for complex types.
FastAPI's dependency injection system relies on Python's typing module, which follows these naming conventions strictly. Type aliases should follow class naming conventions (PascalCase), while variables use snake_case."