-
Notifications
You must be signed in to change notification settings - Fork 5
Installation on macOS
The installation on macOS needs some additional tools which are provided by Homebrew for example. Get Homebrew using this command:
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
Then install the additional tools:
brew install node openssl postgresql redis vips
Check the status of the postgresql
and redis
services:
brew services
Both services should have the status "started", otherwise start them now:
brew services start postgresql
brew services start redis
and check the status again.
Get the latest source code from GitLab and install it in your HOME
directory:
git clone https://gitlab.com/scripta/escriptorium.git $HOME/escriptorium
macOS already provides Python 3.9 which can be used for eScriptorium as part of the Command Line Developer Tools.
If those tools are not already installed (/usr/bin/python3 --version
shows error message), those tools should be installed first:
xcode-select --install
If Python 3.9 still does not work, running this command should help:
sudo xcode-select --switch /Library/Developer/CommandLineTools
Then create a virtual Python environment and activate it:
/usr/bin/python3 -m venv $HOME/venv3.9
source $HOME/venv3.9/bin/activate
Update / install some basic Python packages:
pip install -U pip setuptools wheel
Maybe the following hack is required before the required Python packages can be installed:
ln -s /Library/Developer/CommandLineTools/Library/Frameworks/Python3.framework/Versions/3.9/lib/python3.9/config-3.9-darwin $VIRTUAL_ENV/lib/python3.9
Install the required Python packages:
export CPPFLAGS="-I$(brew --prefix)/opt/openssl@3/include"
export LDFLAGS="-L$(brew --prefix)/opt/openssl@3/lib"
# The scipy version in requirements.txt cannot be installed.
pip install $(cat $HOME/escriptorium/app/requirements.txt|grep -v "^#"|grep -v scipy)
Unless the development mode is disabled later in the local settings, some additional packages are required:
pip install -r $HOME/escriptorium/app/requirements-dev.txt
Get the Node modules and build the local code using these commands:
cd $HOME/escriptorium/front
npm install
# Optionally run the following command to fix security issues.
npm audit fix
npm run build
createdb escriptorium
cd $HOME/escriptorium/app
# Copy the local settings. Optionally edit them for example to disable the development mode.
cp escriptorium/local_settings.py{.example,}
python manage.py check --settings escriptorium.local_settings
python manage.py migrate --settings escriptorium.local_settings
Activate the virtual Python environment if it is not still active:
source $HOME/venv3.9/bin/activate
Run the worker process and the web application:
export DJANGO_SETTINGS_MODULE=escriptorium.local_settings
cd $HOME/escriptorium/app
celery -A escriptorium worker -l DEBUG &
python manage.py runserver --settings escriptorium.local_settings
Now connect your web browser to http://localhost:8000/ and login with username and password admin
.
Latest pytorch should support accelerated training, see https://pytorch.org/blog/introducing-accelerated-pytorch-training-on-mac/. This still has to be tested with kraken. Updated installation instructions will follow.
Run Debian with podman:
podman run -ti --name escriptorium docker.io/debian:bookworm-slim
Upgrade installed packages and install package sudo
:
apt update
apt upgrade
apt install sudo
Continue the installation by following the instructions for Linux.