capgaga / image-video-analyzer-api

FastAPI & ONNXmodel video frames analyzer

Video frames analyzer API

How to use:

Setup and run

Clone this repository with git clone [email protected]:CAPGAGA/image-video-analyzer-API.git

Setup with docker compose via docker-compose up -d

After that, you can access API via localhost/docs or 127.0.0.1/docs depending on your local DNS setup

API endpoints and what they do

To submit your video for analysis use POST request to the /analysis endpoint with fields media file in request body

example:

curl -X 'POST' \
  'http://127.0.0.1:8000/analysis/' \
  -H 'accept: application/json' \
  -H 'Content-Type: multipart/form-data' \
  -F 'media=@855538-hd_1920_1080_25fps.mp4;type=video/mp4'

example of successful respond:

{"requestId": 18}

requestId is an id of submitted file, which can be used to analysis data about of each frame of the video as shown in next example

To get information of each frame use GET request with reuqestId in URL as shown:

example:

curl -X 'GET' \
  'http://127.0.0.1:8000/analysis/18/' \
  -H 'accept: application/json'

example of successful respond:

[
  {
    "file_id": 18,
    "result": "African elephant, Loxodonta africana",
    "id": 1,
    "created": null
  },
  {
    "file_id": 18,
    "result": "African elephant, Loxodonta africana",
    "id": 2,
    "created": null
  },
  {
    "file_id": 18,
    "result": "African elephant, Loxodonta africana",
    "id": 3,
    "created": null
  },
  ...
]

Keep in mind that analysis is done in background, so results might not be accessible right away!

To delete analysis results send DELETE request with requestId in URL as shown:

example:

curl -X 'DELETE' \
  'http://127.0.0.1:8000/analysis/18/' \
  -H 'accept: application/json'

there will be no response here if all is successful