Mobile friendly web app to perform OCR.
Local (client side) and Remote (server side) options are available.
Benchmark option compares local and remote performances.
Uses Docker and Kubernetes to deploy a cluster on Google Container Engine. Uses MongoDB replica set of 3 nodes with persistent disk. Uses Google Cloud Storage to save source images.
Allowed image extensions:
- jpg
- jpeg
- png
Not compatible with Safari.
To run it locally follow these steps:
# Install needed dependencies
apt-get update --fix-missing
apt-get install build-essential python python-dev python-pip python-virtualenv libjpeg-dev libffi-dev libssl-dev tesseract-ocr --assume-yes
# Clone the code
git clone https://github.com/federico-fiorini/ocr-benchmark.git
cd ocr-benchmark
# Install virtualenv
pip install --upgrade pip virtualenv
# Install mongodb
# On Mac
brew install mongodb
# On Ubuntu
sudo apt-get install -y mongodb-org
# Run mongodb deamon in a new terminal
mongod
# Back to previous terminal: create virtual environment
virtualenv .env
.env/bin/pip install -r requirements.txt
# Set env variables
export UPLOAD_FOLDER="/tmp"
export GOOGLE_APPLICATION_CREDENTIALS="path/to/credentials/google-cloud-auth.json"
# Run
./run.py
# Open the browser at the given IP address
To deploy on google cloud follow these steps:
# 1) Install gcloud and authenticate
sudo ./install_gcloud.sh
# 2) Run deployment script
sudo ./deploy.sh
# 3) Open the browser at the given IP address
To roll out an update run:
# Deploy the last version of the code
./roll_update.sh v2
Login with one of the following test users:
- username:
user1, password:password1 - username:
user2, password:password2 - username:
user3, password:password3
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent