Ad-libs.js: A light weight polyfill and dependency free common utilities library

Ad-libs.js is a collection of front end CommonJS modules. It provides front end services such as browser and OS detection, XMLHttpRequests, and performance markers. It allows the user to safely execute various Javascript methods across all browsers.

See the full wiki documentation for more information.

| Source Code

Conversant ConcurrentQueue and DisruptorBlockingQueue

Disruptor is the highest performing intra-thread transfer mechanism available in Java. Conversant Disruptor is the highest performing implementation of this type of ring buffer queue because it has almost no overhead and it exploits a particularly simple design.

See the full wiki documentation for more information.

Wiki | Source Code

Conversant RTree

RTree is a spatial indexing strategy that involves building a tree of bounding boxes that support arbitrary range searches. RTrees are efficient for geospatial data but can be extended to support any data that is amenable to range search queries.

Conversant RTree is a hyperdimensional implementation of RTree that supports data with arbitrarily large numbers of orthogonal relations.

See the full wiki documentation for more information.

Wiki | Source Code

Greenplum ORC Reader

See the full wiki documentation for more information.

Wiki | Source Code

PostgreSQL HyperLogLog Estimator

This is an implementation of HyperLogLog algorithm as first described in the paper "HyperLogLog: the analysis of near-optimal cardinality estimation algorithm", published by Flajolet, Fusy, Gandouet and Meunier in 2007. Generally it is an improved version of LogLog algorithm with the last step modified, to combine the parts using harmonic means.

Several improvements have been included that are described in "HyperLogLog in Practice: Algorithmic Engineering of a State of The Art Cardinality Estimation Algorithm", published by Stefan Heulem, Marc Nunkesse and Alexander Hall.

See the full wiki documentation for more information.

Source Code

Conversant Spark Profiler

Spark Profiler shows how "events" generated by Spark applications can be analyzed for profiling them. Profiling here means understanding how and where an application spent its time, the amount of processing it did, its memory footprint, etc. Since Apache Spark is a distributed processing framework, this kind of processing helps understand application resource utilization and provides a framework to optimize and tune applications.

See the project repo for more information.

Wiki | Source Code

Mara: Annotations for Mapreduce

Mara is a design and implementation for developing Hadoop MapReduce jobs in Java. Mara grew out of the need to simplify and standardize the process in use at Conversant.

See the full wiki documentation for more information.

Wiki | Source Code