Sphinx находит не все посты и записи блогов (3.4.6)

[siv1987 2628]

Для начала выяснить если блоги вообще поддерживают поиск по сфинксу.

[Mac 1]

Да вроде поддерживают:

################################# --- BLOG --- ##############################

source ipb_blog_search_main : ipb_ipb_source_config
{
   # Set our forum PID counter
   sql_query_pre   = REPLACE INTO ipb_cache_store VALUES( 'sphinx_blog_counter', (SELECT max(entry_id) FROM ipb_blog_entries), 0, UNIX_TIMESTAMP(), 0 )

   # Query posts for the main source
   sql_query       = SELECT e.entry_id, e.entry_id as search_id, e.entry_name, e.entry_name as tordinal, e.entry, e.entry_date, e.entry_author_id, e.entry_num_comments, e.blog_id, \
                            b.blog_owner_only, b.blog_private, b.blog_disabled, \
                           CASE WHEN e.entry_status='published' THEN 0 ELSE e.entry_author_id END AS entry_not_published, \
                           CASE WHEN b.blog_owner_only=0 THEN 0 ELSE b.member_id END AS blog_owner_id, \
                           CASE WHEN b.blog_authorized_users IS NULL THEN 0 ELSE 1 END AS authorized_users \
                       FROM ipb_blog_entries e \
                     LEFT JOIN ipb_blog_blogs b ON ( b.blog_id=e.blog_id )

   # Fields    
   sql_attr_uint           = search_id
   sql_attr_uint           = blog_owner_only
   sql_attr_timestamp      = entry_date
   sql_attr_string         = tordinal
   sql_attr_uint           = entry_author_id
   sql_attr_uint           = entry_num_comments
   sql_attr_uint           = blog_id
   sql_attr_uint           = blog_private
   sql_attr_uint           = blog_disabled
   sql_attr_uint           = entry_not_published
   sql_attr_uint           = blog_owner_id
   sql_attr_uint           = authorized_users
   sql_attr_multi          = uint tag_id from query; SELECT tag_meta_id, tag_id FROM ipb_core_tags WHERE tag_meta_app='blog' AND tag_meta_area='entries'
   sql_ranged_throttle = 0
}

source ipb_blog_search_delta : ipb_blog_search_main
{
   # Override the base sql_query_pre
   sql_query_pre = 

   # Query posts for the delta source
   sql_query       = SELECT e.entry_id, e.entry_id as search_id, e.entry_name, e.entry_name as tordinal, e.entry, e.entry_date, e.entry_author_id, e.entry_num_comments, e.blog_id, b.blog_owner_only, b.blog_private, b.blog_disabled, \
                           CASE WHEN e.entry_status='published' THEN 0 ELSE e.entry_author_id END AS entry_not_published, \
                           CASE WHEN b.blog_owner_only=0 THEN 0 ELSE b.member_id END as blog_owner_id, \
                           CASE WHEN b.blog_authorized_users IS NULL THEN 0 ELSE 1 END AS authorized_users \
                       FROM ipb_blog_entries e \
                     LEFT JOIN ipb_blog_blogs b ON ( b.blog_id=e.blog_id ) \
                     WHERE e.entry_id > ( SELECT cs_value FROM ipb_cache_store WHERE cs_key='sphinx_blog_counter' )    
}

index ipb_blog_search_main
{
   source          = ipb_blog_search_main
   path            = /var/lib/sphinx/mysite_ipb/ipb_blog_search_main

   docinfo         = extern
   mlock           = 0
   morphology      = none
   min_word_len    = 2
   charset_type    = utf-8
   html_strip      = 0
   #infix_fields    = entry_name, entry
   #min_infix_len   = 3
   #enable_star     = 1
}

index ipb_blog_search_delta : ipb_blog_search_main
{
  source           = ipb_blog_search_delta
  path             = /var/lib/sphinx/mysite_ipb/ipb_blog_search_delta
}

source ipb_blog_comments_main : ipb_ipb_source_config
{
   # Set our forum PID counter
   sql_query_pre   = REPLACE INTO ipb_cache_store VALUES( 'sphinx_blog_comments_counter', (SELECT max(comment_id) FROM ipb_blog_comments), 0, UNIX_TIMESTAMP(), 0 )

   # Query posts for the main source
   sql_query       = SELECT c.comment_id, c.comment_id as search_id, c.member_id as comment_member_id, c.comment_date, c.comment_approved, c.comment_text, \
                            e.entry_id, e.entry_name as tordinal, e.entry_date, e.entry_author_id, e.entry_num_comments, e.blog_id, \
                            b.blog_owner_only, b.blog_private, b.blog_disabled, \
                           CASE WHEN e.entry_status='published' THEN 0 ELSE e.entry_author_id END AS entry_not_published, \
                           CASE WHEN b.blog_owner_only=0 THEN 0 ELSE b.member_id END AS blog_owner_id, \
                           CASE WHEN b.blog_authorized_users IS NULL THEN 0 ELSE 1 END AS authorized_users, \
                           CONCAT(e.entry_last_comment_date, '.', e.entry_id ) as last_post_group \
                       FROM ipb_blog_comments c \
                     LEFT JOIN ipb_blog_entries e ON ( c.entry_id=e.entry_id ) \
                     LEFT JOIN ipb_blog_blogs b ON ( b.blog_id=e.blog_id )

   # Fields    
   sql_attr_uint           = search_id
   sql_attr_uint           = entry_id
   sql_attr_uint           = blog_owner_only
   sql_attr_timestamp      = entry_date
   sql_attr_string         = tordinal
   sql_attr_uint           = entry_author_id
   sql_attr_uint           = entry_num_comments
   sql_attr_uint           = blog_id
   sql_attr_uint           = blog_private
   sql_attr_uint           = blog_disabled
   sql_attr_uint           = entry_not_published
   sql_attr_uint           = blog_owner_id
   sql_attr_uint           = authorized_users
   sql_attr_uint           = last_post_group
   sql_attr_timestamp      = comment_date
   sql_attr_uint           = comment_member_id
   sql_attr_uint           = comment_approved
   sql_ranged_throttle = 0
}

source ipb_blog_comments_delta : ipb_blog_comments_main
{
   # Override the base sql_query_pre
   sql_query_pre = 

   # Query posts for the delta source
   sql_query       = SELECT c.comment_id, c.comment_id as search_id, c.member_id as comment_member_id, c.comment_date, c.comment_approved, c.comment_text, \
                            e.entry_id, e.entry_name as tordinal, e.entry_date, e.entry_author_id, e.entry_num_comments, e.blog_id, \
                            b.blog_owner_only, b.blog_private, b.blog_disabled, \
                           CASE WHEN e.entry_status='published' THEN 0 ELSE e.entry_author_id END AS entry_not_published, \
                           CASE WHEN b.blog_owner_only=0 THEN 0 ELSE b.member_id END AS blog_owner_id, \
                           CASE WHEN b.blog_authorized_users IS NULL THEN 0 ELSE 1 END AS authorized_users, \
                           CONCAT(e.entry_last_comment_date, '.', e.entry_id ) as last_post_group \
                       FROM ipb_blog_comments c \
                     LEFT JOIN ipb_blog_entries e ON ( c.entry_id=e.entry_id ) \
                     LEFT JOIN ipb_blog_blogs b ON ( b.blog_id=e.blog_id ) \
                     WHERE c.comment_id <= ( SELECT cs_value FROM ipb_cache_store WHERE cs_key='sphinx_blog_comments_counter' )    
}

index ipb_blog_comments_main
{
   source          = ipb_blog_comments_main
   path            = /var/lib/sphinx/mysite_ipb/ipb_blog_comments_main

   docinfo         = extern
   mlock           = 0
   morphology      = none
   min_word_len    = 2
   charset_type    = utf-8
   html_strip      = 0
   #infix_fields    = comment_text
   #min_infix_len   = 3
   #enable_star     = 1
}

index ipb_blog_comments_delta : ipb_blog_comments_main
{
  source           = ipb_blog_comments_delta
  path             = /var/lib/sphinx/mysite_ipb/ipb_blog_comments_delta
}

[siv1987 2628]

Тогда должен работать.

[Mac 1]

Как я и подозревал, это был баг. Добил ТП, исправили. Всем, у кого стоит сфинкс, советую внести этот багфикс.

 

А вам, siv1987, большое человеческое спасибо за участие! А то в других сообществах IPB просто глухая стена.

Изменено 12 сентября 2014 пользователем Mac

[Trotor 27]

Поиск Сфинкс конечно отличный - ищет мгновенно.

 

Однако не решил проблему как у автора топика - максимальное кол-во страниц 40, как не меняй разные параметры в файлах :(

 

И ещё. Индексация всего контента раз в сутки (не нового, новый контент индексируеться каждые десять минут за микросекунды) занимает почти две минуты, и всё время это форум простаивает. Можно ли как то оптимизировать индексацию?