Demo: Image Search with Manticore Search

Publicly available demo - image.manticoresearch.com

Blogpost about this project

Just run the demo locally

To run the project on your local machine, you need to have Docker installed with the compose plugin. Follow these commands to start:

git clone https://github.com/manticoresoftware/manticore-image-search.git
cd manticore-image-search
cd docker/containers/manticore

wget https://github.com/manticoresoftware/manticore-image-search/releases/download/241204/backup.tar.gz && rm -fr backup && tar xzf backup.tar.gz

cd ../..
docker compose down -v
docker compose up

After completing these steps, the project should be accessible at http://localhost/.

Further configuration

The default port for the server is 80, so if you need to change it, update the nginx section in app/config/app.ini.yml.

Preparing for deployment

If you plan to extend this project beyond a simple Manticore Image Search demo and want to run your own version from this repository, we've got you covered with remote server deployment. To get started, first grab yoda and give its documentation a quick read to understand how it works.

Deployment

For deployment, tweak the docker/Envfile with your server details and make sure you have passwordless authentication set up with your SSH key. Then, simply run:

yoda deploy --env=production

Setting up a new server in 5 steps

Please remember, deployment happens from the default branch to the server from the local machine. You can use the --branch flag with the yoda command to deploy from a different branch.

Place your public keys in the docker/.ssh/authorized_keys folder on the local machine.

1. Start by initializing a new server using Rocky Linux 9 as the base operating system.

2. Enter your server's IP address in the Envfile file on your local machine.

3. Execute the following command to prepare the server and also set up SSH keys for you:

   yoda setup --host=server-ip

4. Kick off the deployment process with the next command:

   yoda deploy --host=server-ip

   or

   yoda deploy --env=production --branch=main

5. Wait for it to finish, and then you're good to go!