Inside Intelligence

The Grapeshot API has been designed as a small footprint of code to be inserted in other developers' applications.

It provides the personalization, summarization, categorization features of the iSA Product Set as a single SDK environment, written in C.

The API makes no underlying assumptions about the context in which the API is be used. It is extremely flexible to embed in your own software system. Customers have wrapped the API in java, C++ and other types of class library.

There is a rich set of features which you can add to your own applications - enhancing your "search" function and improving user satisfaction. The Grapeshot API has a small footprint - 300k of code.

• Multiple Operating Systems - Linux, Windows, Solaris etc
• Callable at the command line - good for debugging mode
• Core code is written in C
• Performing 65 transactions per second on single Java JNI thread
• Callable via XML scripts called Grapescript - good for interoperability needs
• Simultaneous read and write capability on Grapeshot index/database files
• Unicode & UTF8 compliant - working in Chinese Arabic and other full character sets

Multi-level API

• First, there is a the API, defined as a set of functions in C.
• Second, there is the command language, made up of a small number of built-in commands, and a larger collection of system macro commands, which may be indefinitely extended with user defined macro commands.
• Third, there are grapes, which are small programs written in an internal scripting language, and which can be compiled and run interpretively by the Grapeshot system.

The three levels closely connect: grapes can be compiled and run through the command language or through the API. Any Grapeshot command can be called as an API function using X_obey. Similarly a grape can cause any Grapeshot string to be obeyed as a Grapeshot command with the grape command obey. Within grapes, access to system resources is provided though a system of handles, which can be opened, closed and operated on, and these roughly correspond to the resources created, processed and released through the API. Often therefore a single task can be plausibly done in a number of different ways. This is especially true of record building, where, at the API level alone, there are still a number of different approaches.

Using the API to call up commands and command sequences provides an easy route to software development. Grapeshot is used at the command line level to build up a set of commands that provide the necessary functionality for some application, and then the commands are called up through the API when Grapeshot is later integrated with a range of other systems.

Grapes can be created in XML, thereby exploiting one key feature of Grapeshot: that it can go beyond using XML as source data, and instead build a complete XML system for querying protocols, indexing scripts and data management file types. Even the Porter Stemmers can be run as XML commands! See XML Platform.

The Grapeshot SDK is a software toolkit designed specifically for OEM usage with a comprehensive API for delivering Intelligent Search Architecture inside your own application. Built for a distributed, networked architecture, Grapeshot is small, robust and XML-based.