QR Code Scanner

This repository contains a Web App and an API which extract QR codes from images and scans the content.

PLEASE: read everything

--

• Dev enviroment setup
• Web App
• API call
  • Test in FastAPI UI
  • Test with a python script in WSL CLI
  • Test with a cURL script in WSL CLI
• Reasoning behind

--

Dev enviroment setup

This repo only works locally with Windows WSL2 and Linux.

You need to install some dependencies (if you are in Windows):

• Install Windows terminal. This makes easier to switch between Windows and WSL (Windows Substystem in Linux) terminal
• Install Ubuntu with WSL

Inside WSL terminal

With Windows terminal you must open a new terminal inside WSL (Ubuntu):

Setup script

• Clone this repository

git clone git@github.com:alejlatorre/QRCodeScanner.git

• Inside the scripts folder run the config file

cd QRCodeScanner/
cd scripts/
chmod +x setup.sh
bash setup.sh

• After you run the setup.sh config file, you must locate on the code folder

cd ~/truora/QRCodeScanner

• Activate enviroment

source .venv/bin/activate

Web app

Deploy

1. Kill ports if they are in use by another program.

sudo lsof -t -i tcp:5000 | xargs kill -9

2. Run the program with poetry

poetry run python web_app.py

3. Open http://127.0.0.1:5000

• Usage:
  • Select a file stored in data/ folder
  • Click on submit
  • The input/output images and the extracted text will be shown
  • Test with more files. It also manages rotated QRs and images with multiple QR codes

API call

1. Kill ports if they are in use by another program.

sudo lsof -t -i tcp:8000 | xargs kill -9

2. Run the program with uvicorn

uvicorn api:app --reload

Test in FastAPI UI

• Enter into FastAPI UI

• Usage:

  • Click on Try it out
  • Select a file
  • Click on Execute
  • The response body will be shown

Test with a python script in WSL CLI

• With the program running, you need to open a new tab in Windows terminal

• Run api_request_example.py script with an image stored in data/ folder

  cd ~/truora/QRCodeScanner
  source .venv/bin/activate
  poetry run python api_request_example.py 'qrcode_sproutqr.jpg'
  

  

Test with a cURL script in WSL CLI

• With the program running, you need to open a new tab in Windows terminal

• Run the script with an image stored in data/ folder

  cd ~/truora/QRCodeScanner
  source .venv/bin/activate
  curl -X 'POST' 'http://127.0.0.1:8000/get_text' -H 'accept: application/json' -H 'Content-Type: multipart/form-data' -F 'img=@data/qrcode_sproutqr.jpg;type=image/jpeg'
  

  

Reasoning behind

First, we use a customized module called engine where is a class called QRCodeScanner.

This class have two methods:

1. read_image: It is used to read files with cv2 library when data is stored locally.
2. extract_info: It is used to extract three elements from decoded image (rectangle, polygons and text decoded)
   • Rectangle: Detects the perimeter where the QR code is in 2D
   • Polygons: Detects the shape of the QR code regardless of the rotation angle
   • Text: Extracts the decoded text from the QR code

Note: The extract_info returns a list of strings, so if we have an image with multiple QR codes, it will returns a list of decoded texts.

In the Web App version, we use Flask and POST method to send an input file through a form element. Then it validates allowed extensions (jpg, jpeg and png) and save the image in application/static/ folder in order to use it with QRCodeScanner class. Finally, the extracted text is iteratively shown through flask messages (with flash method) because we can have a list of decoded texts.

In the API version, we use FastAPI to create an app that contains a POST request url called http://127.0.0.1:8000/get_text and basically uploads a file that is open with Pillow and then the QRCodeScanner class is used to extract the list of decoded texts.