YashIndane / web-sudoku

:white_square_button: A flask webapp to solve sudoku from image input

web-sudoku

A flask webapp to solve sudoku from image input

Demo link

AWS CLI configuration

Have user ready in AWS with admin and power user access. This user will be used to call the AWS Textract service. Create a S3 bucket in this users account.

Usage

$ sudo docker run -dit -p <PORT-NO>:1453  --name <NAME> yashindane/websudoku:v1 --aak="<AWS_ACCESS_KEY>" --ask="<AWS_SECRET_KEY>" --region="<REGION>" --bucketname="<BUCKET_NAME>"

On arm64 v8 machines

$ sudo docker run --platform linux/arm64/v8 -dit -p <PORT-NO>:1453  --name <NAME> yashindane/websudoku:arm64v8 --aak="<AWS_ACCESS_KEY>" --ask="<AWS_SECRET_KEY>" --region="<REGION>" --bucketname="<BUCKET_NAME>"

On arm64 v8 machines using podman

$ sudo podman run --platform linux/arm64/v8 -dit -p <PORT-NO>:1453  --name <NAME> docker.io/yashindane/websudoku:arm64v8 --aak="<AWS_ACCESS_KEY>" --ask="<AWS_SECRET_KEY>" --region="<REGION>" --bucketname="<BUCKET_NAME>"

How it works

The user first clicks a pic of there sudoku board and submits it. The data_uri of the image goes to backend app2.py script through the form on user click. Using base64, ìo and PIL libraries, the image is saved in the server and then uploaded to S3 bucket. Now using AWS Textract, the digits from the image in bucket are extracted and send back to the app2.py. This digits in a double array are fed to the function sudoku3.solve(), which returns the solution to the problem. This diits are then rendered on the output.html page.

Pull requests and new Ideas are welcome!