OasisLMF Complex Model Integration of HailAUS 7

Run the complex model using the API & UI

The following deployment instruction assumes that a linux system has been deployed with a Debian based distribution (preferably Ubuntu 18.04).

Azure: you will need to create inbound rules for port 22 (SSH) and 8080 (HTTP). The former will be used to connect to the instance via SSH for deployment, file transfer and configuration. The later is used by the Oasis UI to serve the flamingo web interface. The rules can be added in the Networking setting page on the Azure web portal. This page will also show the public IP of your instance.

Azure: you will need to use putty (or a similar connection tool) to connect to the Azure instance via ssh and run the following commands.

Connect to the deployment machine then install p7zip, git, docker and docker-compose

For example on an Ubuntu/Debian based Linux system use:

sudo apt update && sudo apt install p7zip-full git docker docker-compose -y

Add $USER to the docker group and switch user to obtain a new shell

sudo usermod -aG docker $USER
sudo su - $USER

Azure: you may need to perform the following in /mnt as we require at least 35GB of storage for the hazard and a lot more for temporary storage (see the oasis_integration.pdf for more information about the hardware requirement and an example Azure specification). You can also work in a more persistent storage. In the following step, double check that /mnt has enough free space (e.g. df -h /mnt)

Azure: by default, the drive containing the Operating System is 30GB. You will need to expand this drive for larger analysis. See the documentation for more information.

Create working directory and clone the latest Oasis integration release

sudo mkdir /mnt/oasis
sudo chown $USER: /mnt/oasis
cd /mnt/oasis
git clone https://github.com/risk-frontiers/OasisIntegration.git
cd OasisIntegration

Transfer model_data.7z into the OasisIntegration folder. You can use WinSCP or pscp to transfer files from windows to linux. Then extract and remove the compressed archive:

7z x model_data_x.y.7z
rm model_data_x.y.7z

Transfer the license.txt into OasisIntegration/model_data/x.y folder. The folder structure should be as follows:

OasisIntegration/
├── api_evaluation_notebook 
├── complex_model <---------------------- this contains Risk Frontiers' executables
├── conf.ini
├── docker-compose.yml
├── Dockerfile.custom_model_worker
├── install.sh   <---------------------- installation script
├── jenkins
├── model_data   <---------------------- model_data contains revisions of Risk Frontiers' data
│   └── x.y  <---------------------- each version folder contains license.txt
├── model_resource.json
├── README.md
├── requirements.txt
├── reset.sh    <------------------------ removes all containers and, optionaly, stored analysis data
├── setup.py
├── tasks.py
├── tests
└── tools

Optional: Log in using a credential that has access to coreoasis/rf_hail:tagname

docker login

Run the installation script and follow the instructions.

chmod +x install.sh
./install.sh

You will be asked to confirm the value for KTOOLS_NUM_PROCESSES, model data path and license file. You can directly retrieve the docker image from Oasis' docker image repository provided you have run docker login above.

Access via the accessible IP (Public IP for Azure) using the default user: admin pass: password

OasisUI Interface - localhost:8080
API Swagger UI - localhost:8000
API Admin Panel - localhost:8000/admin

Azure: you will also need to add an inbound firewall rule for port 8000 if you need direct access to the Oasis API

To update/upgrade the framework, do

git pull
chmod +x install.sh
./install.sh

Notes:

A valid Risk Frontiers license is required to run the integrated model. Please contact Risk Frontiers for more information.