Skip to content

Latest commit

 

History

History
148 lines (104 loc) · 4.04 KB

INSTALL.md

File metadata and controls

148 lines (104 loc) · 4.04 KB
Raw

Installation

There are several ways to install portal_client:

  1. Using easy_install
  2. Using pip
  3. Using VirtualEnv
  4. Using Docker

The portal client requires Python 3, the Boto library, and two Google python libraries:

Using easy_install

Download or clone the portal_client code from github.

  $ cd portal_client

Then one can simply use easy_install. If you have root privileges:

  # easy_install .

or with sudo:

  $ sudo easy_install .

If you are performing a non-root installation, you can still use easy_install. First, pick an installation directory. In this example we'll use /tmp. Then add the installation directory to your PYTHONPATH environment variable if it isn't already there:

  $ export PYTHONPATH=$PYTHONPATH:/tmp

Then invoke easy_install with the --install-dir option. Note the final '.', which tells easy_install where to look for the setup.py script.

  $ easy_install --install-dir /tmp .

Using pip

Another tool that is commonly used to install Python modules is pip. To use pip to install portal_client, download the source code as shown above, then invoke pip as root or using sudo:

  $ cd portal_client
  
  $ sudo pip3 install .

Using VirtualEnv

An easy way to install portal_client and the necessary dependencies is to use VirtualEnv and pip (or pip3 on some systems). The following commands assume you already have VirtualEnv installed on your system and the portal_client software downloaded.

  1. Create a virtual environment
virtualenv /path/to/venvs/portal_client
  1. Activate the virtual environment
source /path/to/venvs/portal_client/bin/activate
  1. Install the portal_client into the virtual environment

Switch back to your download of portal_client and execute pip3 install as follows:

pip3 install .

This will retrieve and install the dependencies as well.

Using Docker

The portal_client code comes with a Dockerfile, which, when used, will build a docker image with Python 3.6 as well as the dependencies specific to the portal client. One can then use this Docker image to execute the client using the following steps:

  1. Build the image. Change to the directory containing the Dockerfile and execute:
docker build -t portal_client .
  1. Use the built image to start a container and execute the client:
docker run -ti --rm portal_client portal_client --help
  1. Test the container by downloading a few small files. The command below should download two files to your current directory ($PWD). This works because we have mapped your current working directory to the /tmp directory in the container with the -v option, and we are executing the client in the /tmp directory with the use of the -w option.
docker run -v "$PWD:/tmp" -w /tmp -ti --rm portal_client portal_client \
    --url=https://raw.githubusercontent.com/IGS/portal_client/master/example_manifests/example_manifest.tsv
  • If running on EC2, this will automatically be detected and S3 will be the preferred endpoint. Example:
docker run -ti --rm -v "$PWD":/tmp -w /tmp portal_client portal_client \
    --url=https://raw.githubusercontent.com/IGS/portal_client/master/example_manifests/example_manifest.tsv
  • If you wish to control which protocol/endpoint to prioritize, you can pass a single endpoint or a comma-separated list (e.g. 'HTTP' or 'HTTP,S3,FTP'). For example, to override the S3 prioritized endpoint on an AWS EC2 instance with the HTTP endpoint:
docker run -ti --rm -v "$PWD:/tmp" -w /tmp portal_client portal_client \
     --endpoint-priority=HTTP \
    --url=https://raw.githubusercontent.com/IGS/portal_client/master/example_manifests/example_manifest.tsv