Programming capstone

[Furaime]

Finished the programming capstone yesterday and thought the challenge overall was a fun experience, even more knowing that this project will be used in the next capstone.

The first thing I want to point out for those that will do it later is that even though there are a dozen warning to not send pull request to the learntocloud repository, I was sure I had done everything right, forking the repository and then cloning it, but I still ended up sending a pull request to learntocloud, which I only understood why after the second time, when you make a pull request, its default repository is the main repository(learntocloud), followed by your own and then everyone that forked it by alphabetic order, felt a bit of shame screwing up like that, but after seeing the list of closed pull requests I realized I wasn't alone.

Python + FastAPI

I already had some exposure to python from a few years ago, but still went over all the chapter and did all the exercises, tried to go a bit further sometimes, making custom modules for testing, and even went back to early exercises to transform then in functions.
I already had some projects using flask and wanted to learn FastAPI so this was kinda nice too, learning how to work the docs helped a lot when doing the capstone project.
PS: I think the uvicorn link is broken or moved, it redirects to unvicorn.org, while uvicord.dev is working.

Postgres

I struggled a bit installing and running postgres, mostly because I'm running Linux Mint which kinda isn't directly supported in postgres, ended up learning how repository works for installation. For SQLAlchemy I think going over the SQL capstone at the same time as you go over the topics should be best, because the content is kinda long so you can just practice along the way, fro additional SQL practice there this game which makes you look into tables to solve a murder mystery highly recommend: Murder Mystery

Generative AI API

For this one I just went ahead and made an account in Openai, got the minimum prepaid credit of $5 by the end of the capstone I had use only $0.01. For the capstone I also went into the Github marketplace model playgrounds and tried the prompts until I found what I wanted. I couldn't get the Github model to work, looked like it got removed from the python-openai-demos, didn't spend too much that on it and went with the OpenAi account instead, if anyone got it working recently let me know.

The Capstone

This one I thought was a lot easier than the networking one, probably because I'm a bit used to python and SQL, only thing that give me trouble was Phase 4, and half of it was because I forgot some instructions in the Readme because its a bit long, I forgot to clear the placeholder for the API Key, other than that I really enjoyed the Github practice.

That's it from me, see you guys later on Cloud capstone which I took a peak and looks really cool.

[madebygps]

Thank you for the feedback.

You mentioned the readme was a bit long, anything you think should be removed?
The GitHub Model piece is becoming tricky with limits I believe, I might just have everyone use OpenAI/Azure OpenAI/ect, it is getting harder and harder to get free LLM usage.

[Furaime]

The readme is long due to necessity, its was just a comment due to some notes about the placeholder right at the start that I ended up forgetting when I got to Task 4, but reading everything again I'm not sure if I screwed up or it really works like that, but basically I had to unset the OPENAI_API_KEY variable, even tho I had updated the .env, it was still set as "placeholder".

From my understanding the OPENAI_API_KEY=placeholder gets set as a container/system variable, which gets priority over the values inside .env.

I'm not sure if that is intended as a learning experience, I did something wrong somewhere or that is some missing information. Maybe someone else that finishes can share how that part went?