README.md

OCR service API server.

this service support just English OCR.

requirement

I implemented it with below. I don't confirm with other enviromnemet due to short time to implement

python = 3.9 docker = 20.10.12

quick start

Below explain on the basis that you can use docker and python in your environment.

1. Please install some python modules

pip install -r requirement.txt

2. Please build image and run it

# Change to docker directory
cd docker 

# Let's build images
docker compose build

# Try to run container
docker compose up -d
cd  ..

3. Try to run some api

I had prepared to run api with python code easily.

But you can also confirm with http://localhost:5000/<api endpoint>

Please confirm you are in /OCR_service

• If you want to run 'POST /image-sync'

  python test/client_test.py image-sync --image-path <image path>
  # e.g
  python test/client_test.py image-sync --image-path test/data/test.png
  
  # It may print like below.
  # <Response [200]>
  # {'task_id': '< any id >'}

• If you want to run 'POST /image'

  python test/client_test.py post-image --image-path <image path>
  
  # It may print like below.
  # <Response [200]>
  #　{'task_id': '< any id >'}
  

• If you want to run 'GET /image'

  python test/client_test.py get-image --id < any id >
  
  # It may print like below.
  #　{'task_id': '= Se a Ba\nunnicr\n\nQ......'}
  # or
  #　{'task_id': None}

• if you want to run 'POST /image' and 'GET /image' at the same time

  python test/client_test.py image --image-path test/data/test.png
  # It may print like below.
  # {'task_id': 'ed68fd36-bee7-4ff7-9bc0-224274514e9f'}
  # {'task_id': None}

  Please send request 'GET /image'

  python test/client_test.py get-image --id < any id >
  
  # It may print like below.
  #　{'task_id': 'Challenge to make OCR!!!'}

1. clean up

cd docker 
docker compose down -t 0

structure

For future versatility, we have divided the API server and the OCR server into two separate container. By doing this, we expect to be able to adapt autoscaling using kubernetes, etc. and not add to the API server.

I using ubuntu as OCR-server container and API-server container.

And I using MySQL by mysql server and access by sqlalchemy which is python module.

data flow

when you request POST /image-sync, data-flow is like below.

when you request POST /image and GET /image, data-flow is like below.

when the API-server request Image and id to the OCR-server, the API-server don't wait to come back and return id for user.

I inplement with python async for this logic.

Data base

column name	type	overview
id	uuid.UUID	primary key
text	LONGTEXT	OCR result
Image path	string	image apth in OCR server
datetime	datetime	insert date time of OCR result

test

change to ocr_service root

pip install pytest
python -m pytest test