udm_load_ispell_data() loads
ispell data. Returns TRUE on
success, FALSE on error.
agent - agent link
identifier, received after call to
udm_alloc_agent().
var - parameter,
indicating the source for ispell data. May have the following
values:
After using this function to free memory allocated for
ispell data, please use
udm_free_ispell_data(), even if you use
UDM_ISPELL_TYPE_SERVER mode.
The fastest mode is UDM_ISPELL_TYPE_SERVER.
UDM_ISPELL_TYPE_TEXT is slower and UDM_ISPELL_TYPE_DB is the
slowest. The above pattern is
TRUE for mnoGoSearch 3.1.10 - 3.1.11. It is planned
to speed up DB mode in future versions and it is going to be
faster than TEXT mode.
UDM_ISPELL_TYPE_DB - indicates that ispell data
should be loaded from SQL. In this case, parameters val1 and val2 are ignored and should be
left blank. flag should
be equal to 1.
注: flag indicates that after loading ispell data from defined source it sould be sorted (it is necessary for correct functioning of ispell). In case of loading ispell data from files there may be several calls to udm_load_ispell_data(), and there is no sense to sort data after every call, but only after the last one. Since in db mode all the data is loaded by one call, this parameter should have the value 1. In this mode in case of error, e.g. if ispell tables are absent, the function will return FALSE and code and error message will be accessible through udm_error() and udm_errno().
Example:
UDM_ISPELL_TYPE_AFFIX - indicates that ispell data
should be loaded from file and initiates loading affixes
file. In this case val1
defines double letter language code for which affixes are
loaded, and val2 - file
path. Please note, that if a relative path entered, the
module looks for the file not in UDM_CONF_DIR, but in
relation to current path, i.e. to the path where the
script is executed. In case of error in this mode, e.g.
if file is absent, the function will return FALSE, and an error message will
be displayed. Error message text cannot be accessed
through udm_error() and
udm_errno(), since those functions can only
return messages associated with SQL. Please, see flag parameter description
in UDM_ISPELL_TYPE_DB.
Example:
if ((! udm_load_ispell_data($udm,UDM_ISPELL_TYPE_AFFIX,'en','/opt/ispell/en.aff',0)) || (! udm_load_ispell_data($udm,UDM_ISPELL_TYPE_AFFIX,'ru','/opt/ispell/ru.aff',0)) || (! udm_load_ispell_data($udm,UDM_ISPELL_TYPE_SPELL,'en','/opt/ispell/en.dict',0)) || (! udm_load_ispell_data($udm,UDM_ISPELL_TYPE_SPELL,'ru','/opt/ispell/ru.dict',1))) { exit; } |
注: flag is equal to 1 only in the last call.
UDM_ISPELL_TYPE_SPELL - indicates that ispell data
should be loaded from file and initiates loading of
ispell dictionary file. In this case val1 defines double letter
language code for which affixes are loaded, and val2 - file path. Please
note, that if a relative path entered, the module looks
for the file not in UDM_CONF_DIR, but in relation to
current path, i.e. to the path where the script is
executed. In case of error in this mode, e.g. if file is
absent, the function will return
FALSE, and an error message will be displayed.
Error message text cannot be accessed through
udm_error() and
udm_errno(), since those functions can only
return messages associated with SQL. Please, see flag parameter description
in UDM_ISPELL_TYPE_DB.
if ((! Udm_Load_Ispell_Data($udm,UDM_ISPELL_TYPE_AFFIX,'en','/opt/ispell/en.aff',0)) || (! Udm_Load_Ispell_Data($udm,UDM_ISPELL_TYPE_AFFIX,'ru','/opt/ispell/ru.aff',0)) || (! Udm_Load_Ispell_Data($udm,UDM_ISPELL_TYPE_SPELL,'en','/opt/ispell/en.dict',0)) || (! Udm_Load_Ispell_Data($udm,UDM_ISPELL_TYPE_SPELL,'ru','/opt/ispell/ru.dict',1))) { exit; } |
注: flag is equal to 1 only in the last call.
UDM_ISPELL_TYPE_SERVER - enables spell server
support. val1 parameter
indicates address of the host running spell server. val2 ` is not used yet, but
in future releases it is going to indicate number of port
used by spell server.
flag parameter in this case is not needed since
ispell data is stored on spellserver already sorted.
Spelld server reads spell-data from a separate
configuration file
(/usr/local/mnogosearch/etc/spelld.conf by default),
sorts it and stores in memory. With clients server
communicates in two ways: to indexer all the data is
transferred (so that indexer starts faster), from
search.cgi server receives word to normalize and then
passes over to client (search.cgi) list of normalized
word forms. This allows fastest, compared to db and text
modes processing of search queries (by omitting loading
and sorting all the spell data).
udm_load_ispell_data()
function in UDM_ISPELL_TYPE_SERVER mode does not actually
load ispell data, but only defines server address. In
fact, server is automatically used by
udm_find() function when performing search. In
case of errors, e.g. if spellserver is not running or
invalid host indicated, there are no messages returned
and ispell conversion does not work.
注: This function is available in mnoGoSearch 3.1.12 or later.
Example: