Indexing and Search Engine: g-MIR


Principle

g-MIR is a “full-text” engine allowing the indexing of several hundreds of thousands of documents from a file system (Microsoft or Unix), http server or Content Management System.

The instances of g-MIR can be multiplied to increase the performance and the capacity of the system ten fold.

It recognises all the principal document formats and can, by the use of additional filters, adapt easily to all new formats.

It also allows the labeling of documents by a variety of different attributes. The metadata is used as much for the selection/filtering of the documents as it is for retrieval of the information they contain.


Interrogation

The AMI engine is XCQL compatible, XML language of interrogation is standard CQL (Common Query Language).

The interrogation service is an http service.

It respects all levels of security associated with users and individual documents.

g-MIL

AMI’s engine natively implements g-MIL. That means that it indexes documents not only on the words which they contain, but also on the most significant expressions. Thus, a search on the “new economy” firstly shows documents which “speak about” the new economy, and not simply those which quote this expression incidentally or which simply contain these two words, whatever their frequency of occurrence may be.


Real Time Indexing

The AMI engine allows the active indexing of documents by crawling the corpus and automatically creating the index files.

It allows various indexing processes to be applied, for example, when there is a preset amount of new information to be indexed or such that it updates the index in real time regardless.

The indexing service is an http service.


Applications


The applications for a “full-text” indexing engine are numerous; for example :

·  The Search function of a website or Portal, external or internal.
·  Access to documentary information.
·  Applications for intelligence gathering; competitive, economic, technology…
·  Access to an online product catalogue.