QRAPI - API for QR code extraction from PDFs

Dockerized Flask api with sole purpose of reading QR codes from PDFs. Decoding is done using pdf2image and pyzbar python libs.

Structure

Heroku automatically deploys all commits to the master branch where the CI pipeline had passed. The production server is served using gunicorn. When running on Heroku, the app uses a port supplied by Heroku, on localhost the port for testing is 5001. In case you need to debug the app, you can change the CMD at the end of Dockerfile to use Flask instead of Gunicorn.

In case you need to add debugging prints into the app, use app.logger.debug("foo") where app is the flask object.

Usage

When inside the directory, type

• make build to build the image
• make run to build and run the image with preset ENV vars and debug
• make stop to stop the running image
• make bash to access the running container's bash
• make logs show the docker logs
• make rbash run container only with bash

If you need to see the debug log use

• docker-compose up

When in production, do not forget to set the required secret ENV variable:

• API_AUTHORIZATION_TOKEN

Sample website is hosted at http://127.0.0.1/5001

Example

Upload request

curl --location --request POST 'localhost:5001/upload' \
--header 'Authorization: secret' \
--form 'file=@/Users/user123/file.pdf'

Response

Example response when 3 files are sent, 2 of them with qr codes at the first page.

[
  {
    "content": [
      {
        "data": "5499944158390",
        "page": 0,
        "type": "QRCODE"
      },
      {
        "data": "5499944158390",
        "page": 0,
        "type": "EAN13"
      }
    ],
    "filename": "example.pdf",
    "key": "somefile"
  },
  {
    "content": [],
    "filename": "LSTM_fixed.pdf",
    "key": "file[]"
  },
  {
    "content": [
      {
        "data": "5499944158390",
        "page": 0,
        "type": "QRCODE"
      },
      {
        "data": "5499944158390",
        "page": 0,
        "type": "EAN13"
      }
    ],
    "filename": "example.pdf",
    "key": "file[]"
  }
]

Dockerhub integration

To speed up the CI, we build our own image with both python and system requirements. This speeds up the CI pipeline as it just downloads the image instead of having to also build it locally. In order to upload the image (when we change/update dependencies), navigate to Actions > Workflows > Build docker and click the Run workflow button. This will build a new image with new dependencies and automatically push it to dockerhub.