|
Information-Retrieval System InfoReS-XL
InfoReS-XL (Information Retrieval System - Extended Language) represents
the Information Retrieval System, primary directed toward processing of
full-text & bibliographical documents. The architecture of InfoReS-XL is
close to the classic model of Information Retrieval System with inverted
index, which makes possible to conduct quick interactive search of
information in databases of large scopes.
Like the majority of similar systems, InfoReS-XL has the means for the
following:
- Creation of documentary databases;
- Preparation of inquiries & search in databases;
- Supplement of authentification &users authorization on the level of
databases & their fields;
- Inference of the results in the user defined format.
At present InfoReS-XL allows to accumulate large scopes of information &
accomplish quick access to documentary data of wide range of sizes,
structures & purposes.
InfoReS-XL system (the first version appeared in 1995) preserved the
kernel of functional possibilities of precedent search systems & style
of interactive solution of tasks, accepted in such Information Retrieval
Systems as STAIRS & WAIS. Along with that the system got the development
in the direction of effectiveness of full-text search, integration with
the means of Internet & WWW, & also in minimization of resource
expenditures.
In consequence of that InfoReS-XL system in comparison with other
systems has the following advantages:
- effectiveness & economy of databases creation procedures is supplied;
- multi-base operation is realized, that is simultaneous search in
several databases
- standard for UNIX & WWW means of collection & statistics processing
are connected;
- standard means for users authentification are used;
- flexible tools for system administrating are created.
Search potentialities
InfoReS XL system inquires, used under search condition consist of
operands - search terms or their right truncations & operators.
As operands search terms of natural language or their right truncations,
the length of which exceed 2 letters can be used. By default every
lexeme is perceived as a right truncation. Along with that there's a
possibility of exact indication of term. The system is not perceptible
to the letters registers.
In InfoReS XL system the following set of operators, disposed in
accordance with established priorities are used:
- NOT (AND NOT) - logical NOT;
- AND - logical AND;
- OR - logical OR;
- ADJUST - operator of context proximity.
ADJUST selects pairs of search terms, which are apart from each other
for not more than (defined by a user number of words) in the text of a
document (default - 2), along with that the order of terms. While usage
of operators the use of synonym ranges is accepted (which can be
re-defined by system administrator).
- NOT: "!", "^", "NOT";
- AND: "&", " " (space), "+", "AND";
- OR: "|", ",", ";", "OR";
- ADJUST: "~", "ADJ", "NEAR".
The inquiry can be multi-leveled. Different levels are either defined by
priorities of operators, incoming into them or are defined by round
brackets.
Database preservation environment
As preservation environment of textual part of database the following
can be used:
- textual files, with separator between the documents;
- hierarchy of catalogues, in which every document is in a separate
file;
- news-server (e.g., INN).
All auxiliary sets of data (files of unique words, index & inverse
tables, etc.) are binary files.
Search result report generator
The results of search (listing of relevant documents & the documents
themselves) are depicted by template, which are external toward search
sub-system & can be modified without re-compilation of program part. The
given templates are textual files with HTML - marking, in which
meta-symbols of database fields are inserted where necessary -
accordingly, while list of documents (or the documents themselves)
depicting, they are replaced by real meanings.
Inculcations
InfoReS-XL Information Retrieval System at present is the basis for
construction of information environments as at own servers "ElVisti" &
also at partners' & purchasers' servers (more than 30 implementations),
among which there's
|
