FastAPI and Docker

Recording the progress of deploying the NLP project by fastAPI, docker and streamlit.

FastAPI

Create an api for others to use to generate scripts by sending a prompt.

Create a new project of app.
Write codes needed to handle the inputs by users.
Start the api by uvicorn <filename>:app --reload. (Run this under the dir of app.py)

1
from fastapi import FastAPI
2

3
app = FastAPI()
4

5
# define the input type class
6
class InputClass {
7
    prompt: str
8
}
9

10
@app.post('/generate')
11
fun generate_script(input: InputClass):
12
    # deal with the input

Docker

Use docker to pack the fastAPI and project.

Derive the requirement.txt file.
Create the dockerfile.
Set up the docker mirror.
Run the docker container (local).

1
# 1.derive requirement.txt
2
pip freeze > requirements.txt
3

4
# 2.create dockerfile
5
# 使用官方 Python 镜像
6
FROM python:3.9-slim
7

8
# 设置工作目录
9
WORKDIR /app
10

11
# 拷贝依赖和代码
12
COPY requirements.txt .
13
RUN pip install --no-cache-dir -r requirements.txt
14

15
COPY . .
16

17
# 启动 FastAPI 服务
18
CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "8000"]
19

20
# 3.setup docker mirror
21
docker build -t fastapi-llama-app .
22

23
# 4.run container locally
24
docker run -p 8000:8000 fastapi-llama-app