Explore Google’s Latest Open‑Source Gems: From Puppeteer to TensorFlow Fold

Google has long been a model for open‑source contribution, with about 900 active GitHub users in 2017 and roughly 1,100 top repositories.

Google open‑sourced many valuable projects such as the popular machine‑learning system TensorFlow, the web‑frontend framework Angular, and Kubernetes. Below is a look at the notable projects released in the past year.

Puppeteer

Puppeteer is a Node.js API for controlling headless Chrome (or full Chrome). It enables tasks such as generating screenshots and PDFs, retrieving SPA content for pre‑rendering (SSR), and crawling websites.

Guetzli

Guetzli is a JPEG encoder that aims for excellent compression at high visual quality, producing images 20‑30% smaller than libjpeg while offering faster decompression. It only generates non‑progressive JPEGs.

Facets

Facets is a Google‑open‑sourced visualization tool that helps understand, analyze, and debug machine‑learning datasets. It consists of Facets Overview for feature‑level visualizations and Facets Dive for exploring individual data points.

Grumpy

Grumpy is a Python‑to‑Go source code translator and runtime that replaces CPython 2.7. It compiles Python code to Go source, which is then compiled to native code, eliminating the need for a virtual machine.

Python Fire

Python Fire is a Google open‑source library that automatically generates command‑line interfaces (CLI) from any Python code, simplifying CLI creation, debugging, and exploration of existing code.

DeepLearn.js

DeepLearn.js (now TensorFlow.js) is a hardware‑accelerated open‑source JavaScript library for machine intelligence in browsers. It provides APIs for building differentiable data‑flow graphs and a suite of mathematical functions, enabling training and inference of neural networks directly in the browser.

Sonnet

Sonnet is a high‑level framework for quickly building neural‑network modules on TensorFlow. It follows an object‑oriented design similar to Torch/NN, allowing modules to be called with input tensors, automatically handling variable sharing.

Istio

Istio is an open‑source service mesh that provides load balancing, service authentication, and monitoring for microservices, primarily deployed on Kubernetes.

Abseil

Abseil is a collection of C++ library extensions developed at Google for over a decade. It supports projects such as Protocol Buffers, gRPC, and TensorFlow, offering widely tested and production‑ready utilities.

Kubeflow

Kubeflow is a Google‑released machine‑learning toolkit that simplifies running ML workloads on Kubernetes, focusing on providing easy access to the best open‑source solutions rather than rebuilding services.

BOARDGAME.IO

BOARDGAME.IO is a Google open‑source game framework that lets developers express game rules as simple functions describing state transitions, handling networking and backend concerns automatically.

container‑diff

container‑diff is a Google open‑source tool for analyzing and comparing Docker images. It can examine image history, filesystem layers, and package managers (apt, pip, npm) to reveal differences between one or two images.

TensorFlow Fold

TensorFlow Fold is a library for building TensorFlow models that operate on structured data, where the computation graph adapts to the input’s structure. It supports dynamic batching for efficient execution.

OpenFermion

OpenFermion is an open‑source project for compiling and analyzing quantum algorithms that simulate fermionic systems, providing data structures and tools to manipulate fermionic and qubit Hamiltonians.