DataparkSearch Engine

DataparkSearch Engine is a full-featured open sources web-based search engine released under the GNU General Public License and designed to organize search within a website, group of websites, intranet or local system.

Key features

• Support for http, https, ftp, nntp and news URL schemes.
• htdb virtual URL scheme for indexing SQL databases.
• text/html, text/xml, text/plain, audio/mpeg (MP3) and image/gif mime types built-in support.
• External parsers support for other document types.
• Ability to index multilingual sites using content negotiation.
• Searching all of the word forms using ispell affixes and dictionaries.
• Fuzzy searching based on acronyms and abbreviations.
• Stop-words, synonyms and acronyms lists.
• Boolean query language support.
• Popularity Rank based on neural network model.
• Results sorting by relevance, popularity rank, last modified time and by importance (a multiplication of relevance and popularity rank).
• Various character sets support.
• Accent insensitive search.
• Phrase segmenting for Chinese, Japanese, Korean and Thai languages.
• mod_dpsearch - search module for Apache web server.
• Internationalized Domain Names support.
• The Summary Extraction Algorithm.

Documentation

DataparkSearch documentation is enclosed in release or snapshot distribution in doc subdirectory. And it's also available on-line in English (PDF, 1,417,401 bytes) and in Russian.

You can use our forum to ask about DataparkSearch. Or you may subscribe to DataparkSearch group at Yahoo! Groups: tech.groups.yahoo.com/group/dataparksearch/. And you can share your DataparkSearch experience with all the World at our wiki.

Previous DataparkSearch versions and ChangeLog, (As a RSS feed).

DataparkSearch Engine PAD file.

DataparkSearch's TODO.

Download

Latest DataparkSearch version released: dpsearch-4.49.tar.gz, 2,493,884 bytes, 13.02.2008, 13:21 MSK

You may try latest snapshot: dpsearch-4.50-20042008.tar.gz, 2,517,043 bytes, 21.04.2008, 01:58 MSK

dpsearch-spell-ja.tgz, 68,705 bytes, 09.11.2004, 01:06 MSK - Quasi-ispell data for Japanese. THIS IS NOT VALID ISPELL DATA. Can be used only with DataparkSearch 4.27 or later version. All data are in EUC-JP charset.

Additional Data

Frequency dictionaries

Traditional Chinese, 730,641 bytes, 05.03.2005, 03:07 MSK

Mandarin, 394,634 bytes, 05.03.2005, 03:07 MSK

Korean, 30,624 bytes, 05.03.2005, 03:07 MSK

Korean, EUC-KR charset, 246,038 bytes

Thai, 126,572 bytes, 05.03.2005, 03:07 MSK

Synonym lists

English, 774,663 bytes, 05.03.2005, 03:07 MSK

German, 131,880 bytes, 05.03.2005, 03:07 MSK

Italian, 166,684 bytes, 05.03.2005, 03:07 MSK

Polish, 92,158 bytes, 05.03.2005, 03:07 MSK

Russian, 73,968 bytes, 06.03.2005, 21:09 MSK

Acronym and abbreviation lists

English biomedical acronyms and abbreviations, 7,084 bytes, 30.07.2005, 01:42 MSK

Other code

Mirrors

• dataparksearch.newsa.ru (Russia, Moscow).

Bugs

You may see all open or new bug reports or post your bug reports in our bug system.

Sample sites

• 43°N 39°E (PgSQL, cache mode, searchd is used. SQL-server: PIII 670MHz, 512M RAM, IDE SATA 120Gb. Searching PC: Celeron 2.25GHz, 1G RAM, IDE UDMA100. 1'157'668 pages, 1'273'860 sites, 27.2 Gbytes indexed. ). Test search in Chinese, Test search in Japanese, Test search in Korean, Test search in Thai.
• All Sochi's Internet (as subsection of 43°N 39°E)
• News Lookup Service (MySQL, cache mode, searchd not used)
• DataparkSearch Engine usage location map

Donate

If you use DataparkSearch and found it useful, or want to encourage further development, feel free to make a donation (at Kagi) to support this project. Any amount is greatfully appreciated.