Stupa is an associative search engine. You can search related documents with high performance and high precision. Since document data and inverted indexes are kept in memory, stupa reflects updates of documents in search results in real time.
A server implementation of Stupa is possible by using Thrift.
% ./configure
% make
% make check (googletest required)
% sudo make install
% stpctl search [-b][-f] file [invsize]
% stpctl save [-b] infile outfile [invsize]
-b read a binary format file
-f search by feature strings
(default: search by document identifier strings)
invsize maximum size of inverted indexes (default:100)
document_id1 \t key1-1 \t key1-2 \t key1-3 \t ...\n
document_id2 \t key2-1 \t key2-2 \t key2-3 \t ...\n
...
- document_id : string
- key : string
- C++ compiler with STL (Standard Template Library)
- google-sparsehash
- If google-sparsehash not installed, this clustering tool uses "gnu_cxx::hash_map" or "std::map"
GPL2 (Gnu General Public License Version 2)
Mizuki Fujisawa <fujisawa@bayon.cc>