KaiHa / hunt

A flexible, lightweight search platform

Hunt

Hunt is a flexible, lightweight search platform written in Haskell.

The default server implementation provides a powerful JSON-API and a simple web interface.

Features

• Powerful query language
• Schema support (numeric data, dates, geospatial data)
• Granular ranking capabilities
• JSON API
• Extensible architecture

Installation

Dependencies

• GHC: The Glasgow Haskell Compiler
• Cabal: Haskell package management tool

Hunt Installation

The easiest way to get the setup started is to install the current Haskell Platform.

Linux

---

If you're using linux, you are able to use our Makefile for the build.

git clone https://github.com/hunt-framework/hunt.git
cd hunt
make sandbox install

Windows

---

If you're using Windows, you can use cabal for the build. If you like to use sandboxes on Windows, you can copy the neccessary cabal commands from our Makefile.

git clone https://github.com/hunt-framework/hunt.git
cd hunt/hunt-server
cabal install

Getting Started

The following line starts the default server. The web interface is available at http://localhost:3000/.

make startServer

A small sample data set can be inserted with:

make insertJokes

FAQ

Can i run Hunt on a 32Bit machine.

This would not be a good idea. First of all we are using 64Bit Hashes for our internal DocIds. Despite this source of error the 32Bit machine limits the available memory to 4GB.

Why is the CPU usage in idle so high?

GHC performs a a major garbage collection every 0.3 seconds in idle, which can be computationally expensive on a big index. This can be disabled with the GHC RTS option -I0.

Development / History

Hunt was started in 2013 by Ulf Sauer and Chris Reumann to improve and extend the existing Holumbus framework. Holumbus was developed in 2008-2009 by Timo B. Kranz and Sebastian M. Gauck and powers the current Haskell API search Hayoo!. We decided to rebrand, because Hunt represents a major rewrite and breaks compatibility.

A new Hayoo implementation is currently under development by Sebastian Philipp.

Both projects were developed at the FH Wedel under supervision and active support of Prof. Dr. Uwe Schmidt.