AdvaS Advanced Search News

With full documentation in both English and German. No interface changes.

RC2 extends the English documentation. No interface changes.

Advas is completely rewritten, and comes with a new interface as a Python class. The RC1 is available for further testing. The documentation is in German, and lacks a full English translation at the moment. For the available code you may also have a look at the AdvaS git repository.

The git module is enabled, and the source code can be retrieved from there, too. As the first steps the older advas version 0.2.3 was added as well as the unreleased 0.2.4, and the current source code for 0.2.5.

In contrast to the earlier releases, version 0.2.5 comes as an OOP class. This simplifies the usage. Currently, the German version of the documentation is nearly complete, the English version is adapted and expected to be finished within the next week.

Fixed bugs and improved both soundex and metaphone code transformation. Available as RPM for Mandrake 9.2, 10.1 and
Fedora Core 2.

Split advas into several method-specific modules. Added
advas integrity test (for basic functions).
Fixed bug in is_comment and idf (wrong internal variable type).
Added another phonetic algorithm: caverphone 2.0.

Available as tar.gz and RPM package for Mandrake 9.2, 10.1 and for Fedora Core 2.

Added library init script (fixed bug).

Now available as a RPM package for Mandrake 9.2 and for Fedora Core 2.

Stable release. Now available as a tar.gz archive and a RPM package for Mandrake 9.2.

Now available in two formats - as a conventional tar.gz archive, and as an RPM file for Mandrake 9.2 .

Added the phonetic algorithm NYSIIS. Fixed spelling mistakes in the documentation.

Code improvements (comp_descriptors, comp_ngrams, get_ngrams, metaphone, ngram_stemmer) that result in a time reduction up to 90%. Corrections in the documentation.

Added functions for determining synonyms (words with the same meaning) with the use of the OpenThesaurus (http://thesaurus.kdenews.org) (text version). Added IR references.

Added functions for document classification (categories). Code optimization in various routines. Extended language detection (by keywords) - see homepage for more.

Added Knuth-Morris-Pratt algorithm. Improved documentation.

Added documentation (HTML, German).

Extended AdvaS class (seems to be stable). Added a simple ranking algorithm.

Extended AdvaS class (test version). Fixed bug in comp_descriptors. Optimized code.

Added AdvaS as a python class (unfinished test version).

Added k-nearest neighbour algorithm (kNN).

Fixed bugs in I/O error handling in several functions (calc_succ_variety, is_language, table_lookup_stemmer). Optimized functions (split_line, merge_lists, merge_lists_idf, remove_items).

Added a detailed user guide coming as a set of HTML documents.

Extended wordlists for language detection.

Fixed bug in the metaphone implementation and added a method for deriving the used language in a given text.

Advas 0.0.5 was extended with an implementation of the metaphone algorithm. Metaphone works better than soundex.

Added a successor variety stemmer and rsv
function.

paninaro