Fast OCR is a free and opensource Optical Character Recognition Service.

What The the service does?

This application receive a http multipart request with a image file, extract text using OCR from Tesseract and returns the result. I know it's not an UNIQUE project but my main goal is to make a fast and less resource intensive OCR web service.

Live Demo.

https://ocr.0x30c4.dev

Built With.

• Docker - Platform and Software Deployment
• FastApi - Backend Frame-work.
• Nginx - For Load Balancing.
• ReactJs - Front-end Frame-work

Prerequisites.

Make sure you have git, nodejs, make and Docker installed.

Version.

• Production
• Development

Deploy Docker Container (Production Version.).

Clone the repo.

$ git clone https://github.com/0x30c4/FastOCR

Create the data/media/uploads directory where all the images will be.

$ mkdir -p data/media/uploads 

You can change the uploads directory location by changing it from the .env.prod

Build the front-end with node. (Optional step)

$ cd front-end
$ npm run build
$ cd -

Create a docker volume for the postgres database to be persistent.

$ docker volume create pgdata

If you want to take backup of the databse then you can backup the /var/lib/docker/volumes/pgdata directory.

Build the container.

$ make build

Now deploy the production version of the image using make.

$ make up # to run in the foreground.
or 
$ make up op=-d # to run the container in background.

Check logs.

$ docker logs nginx_proxy # this is the nginx container.
$ docker logs ocr_api_srv # this is the FastApi app.
$ docker logs postgres # this is the postgres database.

You can add docker logs -f/--follow <container_name> to follow log output

Stop the container.

$ make down

Deploy Docker Container (Development Version.).