The current deployment uses the existing VM configuration and basically is
(a) downloading the current version from GitHub,
(b) creating a link to /opt/config.py and
(c) linking from /opt/embeddings.
- Login
kinit wilke
(or another upb user)ssh wilke@embeddings.cs.upb.de
- Source code to /opt
emb=/opt/embeddings.cc-2022-06-30
(or another directory)cd /tmp
wget https://github.com/dice-group/embeddings.cc/archive/refs/heads/master.zip
unzip master.zip ; rm master.zip
mv embeddings.cc-master $emb
- Link config
cd $emb
mkdir instance
ln -s /opt/config.py instance/
- Link new instance
sudo systemctl stop embeddings.service
unlink /opt/embeddings
ln -s $emb /opt/embeddings
sudo systemctl start embeddings.service
- Eventually delete old deloyment directories
screen -S webservice-index
. /opt/bashrc.sh
cd /opt/embeddings
./scripts/run-webservice-index-vm.sh
- Test: http://embeddings.cs.uni-paderborn.de:8008/ping
- Elasticsearch API
- Elasticsearch 8.3.1
- Debug webservice-public
- Use run-webservice-public-local.sh to start another instance
- Access at http://embeddings.cs.upb.de:1337
Use deployment commands in screen session. Before, activate environment:
. /opt/anaconda3/etc/profile.d/conda.sh
conda activate embeddings