After my Twitter sentiment article, those past couple of days, I’ve been playing again with the Cloud Natural Language API. This time, I wanted to make a little demo analyzing the text of speeches and remarks published by the press office of the White House. It’s interesting to see how speeches alternate negative and positive sequences, to reinforce the argument being exposed.

As usual, for my cloud demos, my weapons of choice for rapid development are Apache Groovy, with Glide & Gaelyk on Google App Engine! But for this demo, I needed two things:

What’s the mood on Twitter today? Looking at my little twitter demo from a few weeks ago (using Glide & Gaelyk on Google App Engine), I thought I could enrich the visualization with some sentiment analysis to give more color to those tweets. Fortunately, there’s a new API in Google-town, the Cloud Natural Language API (some more info in the announcement and a great post showing textual analysis of Harry Potter and New York Times)!

The brand-new Cloud Natural Language API provides three key services:

Back in 2009, I created Gaelyk, a lightweight toolkit for developing Google App Engineapps using the Apache Groovyprogramming language. I even had the chance to speak at Google I/O 2009about it! Good times, good times… Vladimír Oranýlater joined me in maintaining and evolving Gaelyk, and Kunal Dabircreated the fun Glide project, which is a thin wrapper around Gaelyk to further streamline the development of small to mid-sized apps for Google App Engine.

Today, I want to share with you a quick start guide to develop a little app, that shows some tweets from selected accounts with the Twitter API(thanks to Twitter4J), and using the Material Design Light template for the look’n feel (I used the “dashboard” template). I won’t list all the exact steps, all the precise changes made to the templates, etc, but I want to give you the keys for having a productive experience with Glide and Gaelyk on App Engine. And here’s a screenshot of what we’ll be building:

Github and Google recently announced and released the Github archive to BigQuery, liberating a huge dataset of source code in multiple programming languages, and making it easier to query it and discover some insights.

Github explained that the dataset comprises over 3 terabytes of data, for 2.8 million repositories, 145 million commits over 2 billion file paths! The Google Cloud Platform blog gave some additional pointers to give hints about what’s possible to do with the querying capabilities of BigQuery. Also, you can have a look at the getting started guide with the steps to follow to have fun yourself with the dataset.

As I recently joined Google’s developer advocacy team for Google Cloud Platform, I thought I could have a little bit of fun with combining my passion for Apache Groovy with some cool cloudy stuff from Google! Incidentally, Paolo Di Tommaso tweeted about his own experiments with using Groovy with Apache Spark, and shared his code on Github:

I thought that would be a nice fun first little project to try to use Groovy to run a Spark job on Google Cloud Dataproc! Dataproc manages Hadoop & Spark for you: it’s a service that provides managed Apache Hadoop, Apache Spark, Apache Pig and Apache Hive. You can easily process big datasets at low cost, control those costs by quickly creating managed clusters of any size and turning them off where you’re done. In addition, you can obviously use all the other Google Cloud Platform services and products from Dataproc (ie. store the big datasets in Google Cloud Storage, on HDFS, through BigQuery, etc.)

The cat is out the bag: I’m joining Google on June 6th, as a Developer Advocate for the Google Cloud Platform team!

My Groovy friends will likely remember when I launched Gaelyk, a lightweight toolkit for developing Groovy apps on Google App Engine? Since then, I’ve always been a big fan of the Google Cloud Platform (although it wasn’t called that way then) and followed the latest developments of the whole platform. And wohhh, so many new services and products have seen the light of day since my early experiments with App Engine! So there will be a lot to learn, a lot to do, and thus, a lot to advocate! 

At our disposal, we have so many ways to interact with an API: from a mobile on iOS or Android, from a web application, or from other services or microservices. And all of them have different needs: one wants only a shallow overview of the data, while the other desires a detailed view of a certain resource and all its sub-resources. It’s becoming difficult to design an API that caters to the needs of those varied consumers.

I had the pleasure of writing an article for Nordic APIs on Web API definition languages.

If you’re into the world of Web APIs, you’ve probably heard of formats like Swagger, RAML or API Blueprint. They allow developers to define the contract of the API, with its endpoints, its resources, its representations, allowed methods, the kind of payloads it understands, the status codes returned, and more.

With the contract of your Web API, you can generate code for your backend implementation or client kits, documentation for publishing the details of your API for your API consumers. This contract becomes a key element of your API strategy: a contract between the frontend team and backend team to be sure to work on the same ground, between your tech team implementing the public API of your company and all the API consumers that will interact with it.

The most common API definition languages we spot in the wild are Swagger / OpenAPI Spec, RAML and API Blueprint. All three let you define your endpoints, your resources, your query or path parameters, your headers, status codes, security schemes, and more.

In a nutshell, these definition languages define the structure of your API, and allow you to describe many elements. As standards in the API industry evolve, however, their purpose and design are under continuous scrutiny. Specifically, the extensibility of API specifications with additional elements and feature sets comes into question.

An interesting trend is emerging in the world of Web APIs, with various engineers and companies advocating for dedicated APIs for each consumer with particular needs. Imagine a world where your system needs to expose not only one API for iOS, one API for Android, one for the website, and one for the AngularJS app front end, but also APIs for various set-top boxes and exotic mobile platforms or for third-party companies that call your API. Beyond any ideal design of your API, reality strikes back with the concrete and differing concerns of varied API consumers. You might need to optimize your API accordingly.