There are a few requirements for your environment:
- Docker and Docker Compose is used as a run-time environment; All of the services are packaged as containers and we use Docker Compose as a simple way for you to get started (in a production setting we use OpenShift, Kubernetes, and any cloud-variant of Kubernetes)
- If you are using the CLI (and you probably will), then Python must be available, preferably in a virtual environment (venv)
Docker images are currently not available via OS-Climate, so we will use the original Docker images available on DockerHub. The images are available under the username: "brodagroupsoftware".
If you wish to build the system and create your own Docker images then follow the Buiding the Ecosystem Guide.
We provide several scripts that make it easier to build and run the Ecosystem Platform components. They do, however, expect a small set of environment variables:
- HOME_DIR: This points to your workspace top level directory; Components will be cloned into this directory, and all component scripts expect this to be in inplace fr build and run scripts to work
- DOCKERHUB_USERNAME: Your docker user name which is used to specify your Docker images
- DOCKERHUB_TOKEN: A valid Docker Hub token to publish your images (note that this is OPTIONAL and is not required unless you wish to publish your docker images to a public image repository)
- SAMPLES_DIR: This points to the directory containing sample data; This is only required for a few components and will be note where needed
So, set "HOME_DIR" environment variable (typically in your personal enviornment script) to your workspace top level.
For example, I use bash/zsh as my shell environment/tool and I put the following into my "~/.zshrc" (as shown below, under my personal home directory, I have a "Development/scratch" directory, but you can change this to anything you wish):
export HOME_DIR="~/Development/scratch"
The Ecosystem Platform can be started using the following commands which will pull all necessary Docker images (under the "brodagroupsoftware" username).
First, clone the platform run-time components "osc-dm-mesh-srv".
git clone https://github.com/os-climate/osc-dm-mesh-srv.git
Next, change your current working directory to the created directory.
cd osc-dm-mesh-srv
Now, setup your environment:
source ./bin/environment.sh
Note that we have already setup Docker images under the Docker username "brodagroupsoftware" which should allow you to pull images. If you choose to build the system (see Building the Platform) then you will need to set your own Docker Hub username.
After setting up your environment, you will see setup several environment variables that are used by the platform (something like the following).
USER_NAME: ericbroda
HOME_DIR: /Users/ericbroda/Development/scratch
DOCKERHUB_USERNAME: brodagroupsoftware
DOCKERHUB_TOKEN: ****
ROOT_DIR: /Users/ericbroda/Development/scratch
PROJECT: osc-dm-mesh-srv
PROJECT_DIR: /Users/ericbroda/Development/scratch/osc-dm-mesh-srv
PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION: python
Finally, start the platform (make sure Docker is running):
$PROJECT_DIR/app/startd.sh
This will get the core components running and available. You should see several Docker images running:
docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
42ab62c65dd7 brodagroupsoftware/osc-dm-search-srv:latest "python3 /app/server…" 4 seconds ago Up 2 seconds 0.0.0.0:23000->8000/tcp docker-osc-dm-search-srv-1
1716d89280c8 brodagroupsoftware/osc-dm-registrar-srv:latest "python3 /app/server…" 4 seconds ago Up 2 seconds 0.0.0.0:21000->8000/tcp docker-osc-dm-registrar-srv-1
df7f9487cd67 brodagroupsoftware/osc-dm-proxy-srv:latest "python3 /app/server…" 5 seconds ago Up 2 seconds 0.0.0.0:20000->8000/tcp docker-osc-dm-proxy-srv-1
06476365d0d5 brodagroupsoftware/osc-dm-marketplace-ux:latest "docker-entrypoint.s…" 5 seconds ago Up 2 seconds 0.0.0.0:3000->3000/tcp docker-osc-dm-marketplace-ux-1
a0448815f89b quay.io/coreos/etcd:v3.5.0 "/usr/local/bin/etcd" 5 seconds ago Up 3 seconds 0.0.0.0:2379-2380->2379-2380/tcp docker-etcd1-1
5c34be81afa3 quay.io/coreos/etcd:v3.5.0 "/usr/local/bin/etcd" 5 seconds ago Up 3 seconds 2379-2380/tcp docker-etcd2-1
3f6ad9e74091 quay.io/coreos/etcd:v3.5.0 "/usr/local/bin/etcd" 5 seconds ago Up 3 seconds 2379-2380/tcp docker-etcd3-1
Next, we will bootstrap a data product.
First, clone the data product run-time components "osc-dm-meshdp-srv" (note the name of the repo is close, but different from the previous repo).
git clone https://github.com/os-climate/osc-dm-meshdp-srv.git
Next, change your current working directory to the recently created directory.
cd osc-dm-meshdp-srv
Now, setup your environment:
source ./bin/environment.sh
This will setup several environment variables that are used by the platform:
--- Environment ---
USER_NAME: ericbroda
HOME_DIR: /Users/ericbroda/Development/scratch
DOCKERHUB_USERNAME: brodagroupsoftware
DOCKERHUB_TOKEN: ****
ROOT_DIR: /Users/ericbroda/Development/scratch
PROJECT: osc-dm-meshdp-srv
PROJECT_DIR: /Users/ericbroda/Development/scratch/osc-dm-meshdp-srv
DATA_DIR: /Users/ericbroda/Development/scratch/osc-dm-samples-dat
Data products are define in directory defined by the "DATA_DIR" environment variable. Instances are used to uniquely identify a data product, and the data product identifier is used to specify the configuration in DATA_DIR.
Finally, start the data product as follows, using instance 0 and data product "rmi":
$PROJECT_DIR/app/startd.sh 0 rmi
Now, you may see a few errors in the log for the data product. That is OK. Do not panic! The reason this is happening is that we have bootstrapped services but we have not registered information (user, products etc) that are used to create a healthy system. We will use the test cases (next section) to register information while also testing core system capabilities.
At this point one additional Docker container will be running ("docker-osc-dm-product-srv-1"):
docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
4512a8463445 brodagroupsoftware/osc-dm-product-srv:latest "python3 /app/server…" 5 seconds ago Up 3 seconds 0.0.0.0:24000->8000/tcp docker-osc-dm-product-srv-1
42ab62c65dd7 brodagroupsoftware/osc-dm-search-srv:latest "python3 /app/server…" About a minute ago Up About a minute 0.0.0.0:23000->8000/tcp docker-osc-dm-search-srv-1
1716d89280c8 brodagroupsoftware/osc-dm-registrar-srv:latest "python3 /app/server…" About a minute ago Up About a minute 0.0.0.0:21000->8000/tcp docker-osc-dm-registrar-srv-1
df7f9487cd67 brodagroupsoftware/osc-dm-proxy-srv:latest "python3 /app/server…" About a minute ago Up About a minute 0.0.0.0:20000->8000/tcp docker-osc-dm-proxy-srv-1
06476365d0d5 brodagroupsoftware/osc-dm-marketplace-ux:latest "docker-entrypoint.s…" About a minute ago Up About a minute 0.0.0.0:3000->3000/tcp docker-osc-dm-marketplace-ux-1
a0448815f89b quay.io/coreos/etcd:v3.5.0 "/usr/local/bin/etcd" About a minute ago Up About a minute 0.0.0.0:2379-2380->2379-2380/tcp docker-etcd1-1
5c34be81afa3 quay.io/coreos/etcd:v3.5.0 "/usr/local/bin/etcd" About a minute ago Up About a minute 2379-2380/tcp docker-etcd2-1
3f6ad9e74091 quay.io/coreos/etcd:v3.5.0 "/usr/local/bin/etcd" About a minute ago Up About a minute 2379-2380/tcp docker-etcd3-1
The Ecosystem Platform has a CLI (Command Line Interface) which exposes almost all capabilities from a terminal.
We will use the CLI to get some of our compnents setup, so let's get it installed.
First, clone the CLI osc-dm-mesh-cli.
Next, setup your environment as follows (note that "source" is used)
source ./bin/environment.sh 0 rmi
Next, create a new virtual environment using these convenience scripts, execute the following (this will create a directory called "venv" in your current working directory):
$PROJECT_DIR/bin/venv.sh
Activate the virtual environment:
source $PROJECT_DIR/bin/vactivate.sh
Install the required libraries as follows:
pip install -r requirements.txt
Install pytest:
pip install pytest
The CLI documentation explains how to issue commands to the Ecosystem Platform and its data products. But to make this easy and convenient, we will just use some of the test cases to register users and products (you can run all test cases by following instructions later in this document).
Go to your CLI ("osc-dm-mesh-cli") directory (ensure your environment is setup) and then execute the following test case:
pytest ./tests/test_cli_users.py
After this is run, your data product registration error will go away within a minute, and your data product will be fully registered and ready!
Assuming you have followed the tutorial this far, then the Ecosystem Platform will be running and a sample data product will be registered.
Now, let's see if we can see it in our Bazaar (Marketplace). Open your favorite browser and point to the following link:
http://localhost:3000
You should see a "Login" screen that looks something like the following:
Login with the following credentials:
- Email Address / Username: subscriber.user@brodagroupsoftware.com
- Password: anything you want (we do not have security enabled at this point)
- Role: Subscriber (selected from the dropdown list)
Now, you should see a "Marketplace" screen that looks something like the following:
At this point we only have a limited number of sample data products. Find one of the data product "tiles" that interest you and click on "VIEW" (at the bottom of the tile).
Now, you should see an "Product" screen for the selected product, that looks something like the following:
.
Click on the "VIEW" at the bottom of the first Artifact tile.
Now, you should see an "Artifact" screen for the selected artifact, that looks something like the following:
Click on "ADD TO CART" and you should see a small notification show up at the bottom of the screen indicating that the artifact has been added to your cart.
Click on the "CART" at the top left part of the screen in the header. You should see the contents of your cart containin the artifact that you just added to the cart. Your screen will look something like this:
Now, lets "purchase" the items in the cart (note: there is no "purchase" and nothing is bought. Rather, we use a shopiing cart metaphor to retain useful artifacts in an order for future reference). Click on "ORDER CART" and you should see an empty cart shown, meaning that the order has completed and you have no items in the cart anymore.
Click on "ORDERS" on the left in the header at the top of the screen.
There are several options in the header at the top of the screen that helps navigation:
- MARKETPLACE: presents the Marketplace screen
- DASHBOARD: Not implemented yet, but is intended to be a summary of your products, carts, and activity
- SEARCH: Not implemented yet, but is intended to provide natural language search to find/filter data products
At any point you can logout by clicking "LOGOUT" at the top left of the header at the top of the screen.
Go to your CLI terminal. There are several test cases (in the "tests" directory):
- test_cli_product_generate.py: Generate UUIDs for product and artifacts
- test_cli_users.py: Register and view users
- test_cli_product.py: Register and view products
- test_cli_discovery.py: View information about a data product
- test_cli_carts_orders.py: Register carts and "purchase" orders
Run each of the above test cases (except test_cli_product_generate.py as this is already setup with sample data)