start-fastapi
Version 2021, based on FastAPI, an easy-to-use web app developed upon Starlette Framework
Requirements
- python 3.6+ (for static typing check)
pip3 install -r ./requirements.txt- idea/pycharm (recommended) + venv
Structure
The web application is based on onion style~
The directory structure is:
- app: logics for your application, includes
__init__.pyas entry of user modules- handler: controllers
- middleware: router middleware, like cors
- model: basic data models and internal logics
- service: external logics (to users)
- cfg: config of different envs
- dev: configs of dev env
- app.cfg: python-dotenv cfg for fastapi
- uvicorn.json: cfg for uvicorn ASGI server, the launcher for fastapi
- logger.json: logging cfg for uvicorn
- prod: configs of production env
- dev: configs of dev env
- core: low level libraries and logics, includes
__init__.pyas entry of core modules, better make it able to be reused in other projects- handler
- lib: shortcut apis for user to build logics
- model
- service
- trans: transaction service, for managing tasks running in background
- misc: misc parts, like build, test scripts, etc
- build: build scripts
- dev: resources for dev usage
- gen_code.py: a script for generating codes
- doc: docs
- test: test scripts
- main.py: main entry, init uvicorn and start fastapi app
- requirements.txt: py package requirements
Quick Start
Launch App Server
Run ./main.py to start the example, which includes:
- core modules: health check handler, basic models, transaction service
- cors middleware configured with wildcard *
- test handlers, models and service: simulates a simple market trade system with products and customers
- examples of websocket and upload file are also included in test handlers
The internal steps are:
./main.pyloads the configs inside./cfg/{env}on cmd args, then callsuvicorn.runto start fastapi app at./app/__init__.py- in
./app/__init__.py, core modules and user handlers/models/services are loaded atstartupevent of fastapi app
You can put your launch scripts inside ./misc/build for your different launch options
Coding Guide
To build your logic, common steps are follows:
./main.pyruns server in dev environment in default, in which hot-reload is enabled- add handlers in
./app/handler, add correspondingimport&APP.include_routercodes in./app/__init__.py - add data models in
./app/model, add services in./app/service - add middlewares in
./app/middlewareif necessary
Some tips for coding:
- GET
/docsto test the routers on web page - Avoid using coroutines (
async deffunctions), as it may block the main evtloop, so that other requests are not handled in time.deffunctions will be invoked in different threads - codes of
./coreshould be shareable (for other projects), codes of./appshould fit with current project - You can use
./misc/dev/gen_code.pyto generate template codes for handlers, models & services. Exec it with working directory as project root directory - Code models based on
pydantic.BaseModel, it's powerful
HTTP Response
Most of the handled requests should contain a status code of 200
A simple solution is to use Resp model in ./core/model/handler.py to generate response body for your handlers
{
"success": bool,
"message": str,
"code": IntEnum,
"data": Any,
}
Use Resp.ok to generate success response and use Resp.err to generate error response
Misc
WebSocket
The test handler ./app/handler/test.py contains ws handler examples
To know more about it, see websocket documentation
Deployment
Run ./misc/build/pack.sh to pack the project into ./misc/build/start-fastapi.tar.gz
See ./misc/build/Dockerfile for an example of docker deployment
381 Dec 11, 2022
131 Jan 1, 2023
54 Dec 13, 2022
18 Oct 31, 2022
139 Dec 27, 2022
78 Jan 3, 2023
417 Jan 7, 2023
371 Jan 1, 2023
95 Jan 8, 2023
472 Jan 2, 2023
64 Nov 25, 2022
21 Jan 3, 2022
315 Dec 25, 2022
239 Dec 28, 2022
91 Dec 31, 2022
199 Dec 30, 2022
122 Dec 27, 2022
28 Oct 25, 2022