SYNOPSIS
use SWISH::API::Common;
my $swish = SWISH::API::Common->new();
# Index all files in a directory and its subdirectories
$swish->index("/usr/local/share/doc");
# After indexing once (it's persistent), fire up as many
# queries as you like:
# Search documents containing both "swish" and "install"
for my $hit ($swish->search("swish AND install")) {
print $hit->path(), "\n";
}
DESCRIPTION
"SWISH::API::Common" offers an easy interface to the Swish index engine. While SWISH::API offers a complete API, "SWISH::API::Common" focusses on ease of use.THIS MODULE IS CURRENTLY UNDER DEVELOPMENT. THE API MIGHT CHANGE AT ANY TIME.
Currently, "SWISH::API::Common" just allows for indexing documents in a single directory and any of its subdirectories. Also, don't run index() and search() in parallel yet.
INSTALLATION
"SWISH::API::Common" requires "SWISH::API" and the swish engine to be installed. Please download the latest release from
http://swish-e.org/distribution/swish-e-2.4.3.tar.gz
and untar it, type
./configure make make install
and then install SWISH::API which is contained in the distribution:
cd perl perl Makefile.PL make make install
METHODS
- $sw = SWISH::API::Common->new()
-
Constructor. Takes many options, but the defaults are usually fine.
Available options and their defaults:
# Where SWISH::API::Common stores index files etc. swish_adm_dir "$ENV{HOME}/.swish-common" # The path to swish-e, relative is OK swish_exe "swish-e" # Swish index file swish_idx_file "$self->{swish_adm_dir}/default.idx" # Swish configuration file swish_cnf_file "$self->{swish_adm_dir}/default.cnf" # SWISH Stemming swish_fuzzy_indexing_mode => "Stemming_en" # Maximum amount of data (in bytes) extracted # from a single file file_len_max 100_000 # Preserve every indexed file's atime atime_preserve
- $sw->index($dir, ...)
-
Generate a new index of all text documents under directory $dir. One
or more directories can be specified. - $sw->search("foo AND bar");
-
Searches the index, using the given search expression. Returns a list
hits, which can be asked for their path:
# Search documents containing # both "foo" and "bar" for my $hit ($swish->search("foo AND bar")) { print $hit->path(), "\n"; }
- index_remove
- Permanently delete the current index.
TODO List
* More than one index directory * Remove documents from index * Iterator for search hits