FastAPI

lifespan

from contextlib import asynccontextmanager
from fastapi import FastAPI
import asyncio
import discord

bot = discord.Client(intents=discord.Intents.default())

@asynccontextmanager
async def lifespan(app: FastAPI):
    asyncio.create_task(bot.start("YOUR_TOKEN"))
    yield

app = FastAPI(lifespan=lifespan)

• Code before yield → runs when FastAPI starts
• Code after yield → runs when FastAPI shuts down

FastAPI Dependency Injection

Core idea

• Declare dependencies using Depends()
• FastAPI automatically calls them and injects results Basic usage

async def route(x: Type = Depends(dep_func)):

Key rules

• Do not call dependency yourself

    Depends(func)      # ✅ correct
    Depends(func())    # ❌ wrong

• Dependencies can have parameters → FastAPI fills them from:
  • query, path, body, headers
  • other dependencies

Passing custom values (patterns)

def dep_fixed():
    return func("value")
Depends(dep_fixed)
##
Depends(lambda: func("value"))
##
class Dep:
    def __init__(self, val):
        self.val = val
    def __call__(self):
        return func(self.val)

Depends(Dep("value"))
##
Depends(partial(func, val="value"))
##
def make_dep(val):
    def dep():
        return func(val)
    return dep

Depends(make_dep("value"))

Mental model

• You declare what you need
• FastAPI builds and injects it automatically
• Use wrappers/classes when you need custom inputs

cerery, boto for s3, docker, redis https://www.youtube.com/watch?v=RIVcqT2OGPA git TF-IDF creating corn jobs and auto run script in linux RAG eval future and Context managing, asycnio.semaphor https://www.youtube.com/watch?v=oAkLSJNr5zY KV cache optimization, sliding window attention, memory compression techniques context refinement

https://www.youtube.com/playlist?list=PL-osiE80TeTsak-c-QsVeg0YYG_0TeyXI this have a lot of thing that i want to learn this is fastapi tutorial it have almeic, boto aws, postgres, etc etc check it out