google-api-wrapper 1.0.0a5

Simple wrapper for Google APIs

Overview

This package provides a simple wrapper around Google APIs such as Cloud Compute, IAM, Deployment Manager, and other cloud-related APIs.

It is used internally by Infolinks for provisioning and deployments, and has proven useful, so we’ve decided to share it! Note that this is still work in progress - our internal version is currently more comprehensive in its coverage of Google’s APIs than this version; we will slowly transition more & more APIs from our internal version to this open version, so stay tuned!

NOTE: Google APIs wrapper requires Python 3.5 and above.

Installation

Installing this package from PyPy is simple, using pip:

pip install google-api-wrapper

You can add the --upgrade or -U flag to upgrade it if you already have it installed.

Building

Google APIs wrapper requires Python 3.5 and above. This is a hard requirement unfortunately. That aside, however, building it is fairly simple - you just use pyvenv and pip install -e.

We recommend reading Packaging Guide and the Virtual Environments Guide, but we’ve compiled a simple & quick how-to for your convenience:

1. Clone the repository:

   git clone [e-mail]:infolinks/google-api-wrapper.git

2. Set up a virtual environment:

   pyvenv <virtual-env-directory>

3. Activate the virtual environment (once for every shell/terminal you will be using to build or develop in it):

   source <virtual-env-directory>/bin/activate

4. Install required libraries:

   pip install wheel google-api-python-client twine

5. Install the cloned copy of the source code using:

   pip install -e <git-checkout-of-google-api-wrapper>

6. You can now modify the source code freely.

7. When you want to test it, simple run your Python program which will import the googleapiwrapper package inside it.

Releasing

To release a new version, ensure you are able to successfully build using the above section’s instructions (including registration to PyPi and PyPi testing. Then, do the following:

1. Update the version in setup.py

2. Commit & push the changes to Git:

   git add setup.py
   git commit -m "preparing version X.Y.Z"
   git push
   git tag -a "X.Y.Z" -m "Version X.Y.Z"
   git push origin "X.Y.Z"

3. Build the package:

   rm -rf build dist
   python setup.py sdist bdist_wheel

4. Now that we’ve created the distributions, we need to ensure we are able to upload our package to PyPi. To do this, ensure you have an account over at PyPi as well as at PyPi Testing. We will be using Twine to perform the upload.

5. Ensure you’ve created the ~/.pypirc file, which should look like this (don’t forget to fill-in your credentials inside it!):

   [distutils]
   index-servers=
       pypi
       pypitest
   
   [pypitest]
   repository = https://testpypi.python.org/pypi
   username = <username>
   password = <password>
   
   [pypi]
   repository = https://pypi.python.org/pypi
   username = <username>
   password = <password>

   Note that your own file might contain additional repositories - that’s ok. You will only be using pypitest and pypi when working with this project though.

6. The project needs to be registered with PyPi (to reserve the package name). Note that this only needs to be done once in the project’s lifetime, and has already been done for you! You DO NOT need to run this. But for documentational purposes, here is the command to do it:

   twine register --repository pypitest dist/google_api_wrapper-<version>-py3-none-any.whl
   twine register --repository pypi dist/google_api_wrapper-<version>-py3-none-any.whl

7. Now we can upload the package. We recommend first uploading it to the PyPi Testing package index in order to verify it’s working properly. Here’s how:

   twine upload --repository pypitest dist/*

   Once you’re satisified with the results, you can upload it to the public repository, like this:

   twine upload --repository pypi dist/*

8. Restore the version in setup.py to a dev version (eg. 1.1.0dev)

APIs

The package’s entry point to the APIs is the Cloud module. This module contains a single class - the Cloud class, which you create an instance of using a Google Cloud JSON credentials file (usually generated from a service account) and one or more scopes (scopes are essentially the list of permissions you are requesting to use). By default, if you do not provide any scopes, the framework will ask for the https://www.googleapis.com/auth/cloud-platform scope which means full permissions. Narrow it down to request more specific permissions instead.

Once you have an instance of the Cloud class, it can provide you with more specific Google Cloud API wrappers:

• Compute: this API contains Google Compute APIs such as creating disks & instances, modifying instance group members, etc.

• IAM: this API is used for managing user and service accounts. Not implemented yet.

• Deployment Manager: this API is used for managing deployments using the Google Deployment manager service. Not implemented yet.

• More to come (Network, Pub/Sub, etc)