Skip to content

TiagoPrata/fastapi-frame-stream

Repository files navigation

fastapi-frame-stream

Package to easily stream individual frames (MJPEG) using FastAPI.

FastAPI server for publishing and viewing MJPEG streams.

Raw image files and images as base64 strings can be sent to a 'video stream' and then consumed by any client.

Quick start

Installing

pip install fastapi-frame-stream

Requirements

NOTE: This package will also automatically install:
  • imutils
  • opencv-python
  • python-miltipart

How to use

Server

You can create a simple FastAPI server where it is possible to publish and get multiple streams.

usage code

full code
from fastapi import FastAPI, File, UploadFile
import uvicorn
from pydantic import BaseModel
from fastapi_frame_stream import FrameStreamer

app = FastAPI()
fs = FrameStreamer()

class InputImg(BaseModel):
    img_base64str : str


@app.post("/send_frame_from_string/{stream_id}")
async def send_frame_from_string(stream_id: str, d:InputImg):
    await fs.send_frame(stream_id, d.img_base64str)


@app.post("/send_frame_from_file/{stream_id}")
async def send_frame_from_file(stream_id: str, file: UploadFile = File(...)):
    await fs.send_frame(stream_id, file)


@app.get("/video_feed/{stream_id}")
async def video_feed(stream_id: str):
    return fs.get_stream(stream_id)


if __name__ == '__main__':
    uvicorn.run(app, host="0.0.0.0", port=5000)

Client

Any client can view a published image (MJPEG) stream using a simple <img> tag:

usage code

full code
<!DOCTYPE html>
<html lang="en">
<head>
    <title>Testing fastapi-frame-stream</title>
</head>
<body>
    <img src="http://localhost:5000/video_feed/my_new_stream001">
</body>
</html>

All together

It is possible to upload an image file directly...

Server and client

... or to use any kind of application to convert the frames to base64 and send it to the web server:

Server and client

How it works

The frames sent throught the web server are stored in a temporary (in memory) SQLite DB...

User sending frame

... and the last frame of each stream is retrieved everytime a client wants to visualize the stream.

Retrieving stream