The current version HDPA 1.1 is compatible with Linux OS (reference OS - Ubuntu 18.04)

Running HDPA:
-------------
1. Download and extract HDPA_v1_1.zip
2. Install Anaconda 3 (https://www.anaconda.com/distribution/#download-section - Python 3.7)
3. Copy hdpa_env --> ..../anaconda3/envs
4. conda activate hdpa_env
5. Make sure if it is correctly set (conda info --envs)

6. Check also ..../anaconda3/envs/hdpa_env/bin/kraken if path in line 1 is properly set. It must be absolute path which is set to the python within hdpa_env (/...../anaconda3/envs/hdpa_env/bin/python).

and check if .../app/libs/recognizer/ocr_engine/kraken_batch_skript.sh is executable

7. Move to the app folder and run "python manage.py runserver "ip_addr:port" (localhost:8000)"
(Due to the correct path policy, it is necessary to be in app folder and run directly manage.py 
without any parent path)

8. Open web browser --> localhost:8000

The individual steps are sequential (1. binarization, 2. deskewing, 3a. Line Segmentation 3b. Block Segmentation). Even it is possible to perform deskewing before binarization, we strongly recommend to take steps in given sequence to avoid possible errors)

Optional step: if you want to use tensorflow-gpu instead of tensorflow, please install tensorflow-gpu.

Kraken ocr: https://github.com/mittagessen/kraken