ENCODE Metadata Database

Running the application locally using Docker

Install

Download and install Docker.
Start Docker
Open Docker preferences, find the Advanced tab under Resources. Make sure the engine has at least 8GB of memory.

Build

All the following commands should be run in the root of this repository.

Clean up possible previous build artifacts.

$ make clean

Build the docker image (first time you run this it will take up to 15 minutes):

$ docker build -t encoded-devcontainer:latest -f .devcontainer/Dockerfile .

Start the container with the appropriate ports forwarded, and this directory mounted on /workspaces/encoded in the container.

$ docker run --rm -it -p 6378:6378 -p 6543:6543 -p 8000:8000 -p 9201:9201 -v $(pwd):/workspaces/encoded --workdir /workspaces/encoded --name encode-container encoded-devcontainer:latest bash

In the shell that opens within the container you started in step 3. run the following commands:

$ make devcontainer
$ dev-servers development.ini --app-name app --clear --init --load

In other terminal open a shell in the running container:

$ docker exec -it encode-container bash

In the shell that you opened in step 5. run:

$ pserve development.ini

Browse the app at localhost:6543
Closing both terminals will cause the container to exit. You do not need to do anything else to stop the app.

Running the application in Github Codespaces

In this repository, click the green Code button, choose the Codespaces tab and then click the ... to create a new Codespace.
In the options you can choose the branch (you can also check out your branch later), and the machine size (the second smallest with 4 cores and 8GB of memory is enough).
Click Create codespace
Building the image (and specifically the npm ci command) will take about 15 minutes.
Once the build completes you will be take to a VSCode editor running in your browser. Wait for the postCreateCommand to finish.
Choose the branch you want to run the app from (if you did not do it in the step 2.)
In the terminal run dev-servers development.ini --app-name app --clear --init --load
Open a new terminal tab (the button with a + -symbol).
Run pserve development.ini.
You can now browse the app via the pop-up, or the address shown next to the pserve(6543) Local Address column in ports tab above the terminal window.

Deploying an AWS demo

Building the application is not necessary to deploy a demo. All you need is a python virtual environment with boto3 package installed. In the root of this repository run:

$ python src/encoded/commands/deploy.py <options>

System Installation (Deprecated method, using Docker or Codespaces recommended)

See Snovault OSX System Installation. ENCODE installs Snovault as it is a dependency. The System Installation is the same for both. However, you do not need to set up a running Snovault instance yourself.

Application Installation

For issues see Snovault OSX App Installation first.

Create a virtual env in your work directory.

This example uses the python module venv. Other options would also work, like conda or pyenv. Please note that older versions of pip may cause issues when updating the application. On MacOS pip 21.0.1 is known to work.
```
$ cd your-work-dir
$ python3 -m venv encoded-venv
$ source encoded-venv/bin/activate
$ pip install -U pip==21.0.1
```

Clone the repo and cd into it

git clone git@github.com:ENCODE-DCC/encoded.git
cd encoded

Build Application
```
make clean && make install
```
If you need to develop snovault side by side you can use the following commands instead, assuming encoded and snovault are present at the same level in your filesystem and virtual environment is activated.
```
$ cd .. && pip install -e snovault
$ cd encoded && make clean && make install
```

Run Application

$ dev-servers development.ini --app-name app --clear --init --load
# In a separate terminal, make sure you are in the encoded-venv
$ pserve development.ini

Browse to the interface at http://localhost:6543

Run Tests

# Make sure you are in the encoded-venv
./circle-tests.sh bdd
./circle-tests.sh indexing
./circle-tests.sh indexer
./circle-tests.sh not-bdd-non-indexing
./circle-tests.sh npm

You can also invoke pytest directly if you need more granular control over which Python tests to run.

# Make sure you are in your venv
# Run a specific test in a specific file
$ pytest TEST_FILE_PATH::TEST_NAME
# Run tests with the given mark
$ pytest -m $PYTEST_MARK

Working on the Pyramid configuration

The Pyramid INI files are templated out using Jsonnet. To update these configurations, install the jsonnet executable with brew install jsonnet. Running make config will generate the new configuration and format the jsonnet files, make sure to run this before pushing or CircleCI will fail.

The Jsonnet files and generated config are located in conf/pyramid/. The file sections.libsonnet is a library of functions that each returns a representation of a single section of an INI file. The file config.jsonnet assembles these sections and outputs a concrete INI file.