ML API-SERVICES

This is the source code to run api-services server

Structure of project

1. api-services
2. celery-worker
3. flower

Detail each component

api-services

1. FastAPI: web framework for developing RESTfullAPIs
2. Uvicorn: ASGI web server
3. Background Tasks: it is a function (or services of fastapi) used for the task that has consumer long time
4. AsyncIO function (Task producer)

Some document:
 - https://fastapi.tiangolo.com/tutorial/background-tasks/
 - https://leimao.github.io/blog/Python-Concurrency-High-Level/

celery-worker

1. Celery: The framework that support asynchronous task queue or job queue
2. Pool-config
3. Logger in celery
4. Task consumer

flower

1. Flower: web tools which monitoring Celery Tasks / Events (Manage Tasks or Events)

Other knowledge

1. Redis broker
2. Redis database
3. Logging file

How to run

Run to follow the code

docker compose up -d --build

How to use

1. API docs

xxxx.xxxx.xxxx.xxxx:8081/api/docs

2. Tasks Management

xxxx.xxxx.xxxx.xxxx:5555

How to see the log

1. APIs logs: See at folder system_logs/api-services-logs/
2. Worker logs: See at folder system_logs/celery-logs/

How to see the storages

Please see at: /system_storages/

To do in future

• Replace Redis_broker to RabbitMQ
• Display CPU consumer with task or event
• Deloy OCR model
• Combine Flower with Prometheus metrics
• Combine Prometheus metrics to Grafana monitoring

---

License Copyright (c) [2022] [Nguyễn Thái Học nguyenthaihoc1996@outllok.com]