-- Data model to support content repository of the ArsDigita Community -- System -- Copyright (C) 1999-2000 ArsDigita Corporation -- Author: Karl Goldstein (karlg@arsdigita.com) -- $Id: content-create.sql,v 1.36.2.6 2003/03/01 21:21:20 donb Exp $ -- This is free software distributed under the terms of the GNU Public -- License. Full text of the license is available from the GNU Project: -- http://www.fsf.org/copyleft/gpl.html ---------------------------------- -- CMS datatypes ---------------------------------- -- create ats datatypes for cms create function inline_0 () returns integer as ' begin insert into acs_datatypes (datatype, max_n_values) values (''text'', null); insert into acs_datatypes (datatype, max_n_values) values (''keyword'', 1); return 0; end;' language 'plpgsql'; select inline_0 (); drop function inline_0 (); -------------------------------------------------------------- -- MIME TYPES -------------------------------------------------------------- -- Mime data for the following table is in mime-type-data.sql create table cr_mime_types ( label varchar(200), mime_type varchar(200) constraint cr_mime_types_pk primary key, file_extension varchar(200) ); comment on table cr_mime_types is ' Standard MIME types recognized by the content management system. '; comment on table cr_mime_types is ' file_extension is not used to recognize MIME types, but to associate a file extension to the file after its MIME type is specified. '; -- Currently file_extension is the pk although it seems likely someone -- will want to support multiple mime types with the same extension. -- Would need UI work however create table cr_extension_mime_type_map ( extension varchar(200) constraint cr_mime_type_extension_map_pk primary key, mime_type varchar(200) constraint cr_mime_ext_map_mime_type_ref references cr_mime_types ); create index cr_extension_mime_type_map_idx on cr_extension_mime_type_map(mime_type); comment on table cr_extension_mime_type_map is ' a mapping table for extension to mime_type in db version of ns_guesstype data '; -- Load the mime type data. \i ../common/mime-type-data.sql create table cr_content_mime_type_map ( content_type varchar(100) constraint cr_content_mime_map_ctyp_fk references acs_object_types, mime_type varchar(200) constraint cr_content_mime_map_typ_fk references cr_mime_types, constraint cr_content_mime_map_pk primary key (content_type, mime_type) ); comment on table cr_content_mime_type_map is ' A mapping table that restricts the MIME types associated with a content type. '; -- RI Index -- fairly static, could probably omit this one. create index cr_cont_mimetypmap_mimetyp_idx ON cr_content_mime_type_map(mime_type); -------------------------------------------------------------- -- LOCALES -------------------------------------------------------------- create table cr_locales ( locale varchar(4) constraint cr_locale_abbrev_pk primary key, label varchar(200) constraint cr_locale_name_nil not null constraint cr_locale_name_unq unique, nls_language varchar(30) constraint cr_locale_nls_lang_nil not null, nls_territory varchar(30), nls_charset varchar(30) ); comment on table cr_locales is ' Locale definitions in Oracle consist of a language, and optionally territory and character set. (Languages are associated with default territories and character sets when not defined). The formats for numbers, currency, dates, etc. are determined by the territory. '; insert into cr_locales ( locale, label, nls_language, nls_territory, nls_charset ) values ( 'us', 'American', 'AMERICAN', 'AMERICA', 'WE8ISO8859P1' ); -------------------------------------------------------------- -- CONTENT TYPES -------------------------------------------------------------- create table cr_type_children ( parent_type varchar(100) constraint cr_type_children_parent_fk references acs_object_types, child_type varchar(100) constraint cr_type_children_child_fk references acs_object_types, relation_tag varchar(100), min_n integer, max_n integer, constraint cr_type_children_pk primary key (parent_type, child_type, relation_tag) ); comment on table cr_type_children is ' Constrains the allowable content types which a content type may contain. '; -- RI Indexes create index cr_type_children_chld_type_idx ON cr_type_children(child_type); create table cr_type_relations ( content_type varchar(100) constraint cr_type_relations_parent_fk references acs_object_types, target_type varchar(100) constraint cr_type_relations_child_fk references acs_object_types, relation_tag varchar(100), min_n integer, max_n integer, constraint cr_type_relations_pk primary key (content_type, target_type, relation_tag) ); comment on table cr_type_relations is ' Constrains the allowable object types to which a content type may relate (see above). '; -- RI Indexes create index cr_type_relations_tgt_typ_idx ON cr_type_relations(target_type); -------------------------------------------------------------- -- CONTENT ITEMS -------------------------------------------------------------- -- Define the cr_items table create table cr_items ( item_id integer constraint cr_items_item_id_fk references acs_objects on delete cascade constraint cr_items_pk primary key, parent_id integer constraint cr_items_parent_id_nil not null constraint cr_items_parent_id_fk references acs_objects on delete cascade, name varchar(400) constraint cr_items_name_nil not null, locale varchar(4) constraint cr_items_locale_fk references cr_locales, live_revision integer, latest_revision integer, publish_status varchar(40) constraint cr_items_pub_status_chk check (publish_status in ('production', 'ready', 'live', 'expired') ), content_type varchar(100) constraint cr_items_rev_type_fk references acs_object_types, storage_type varchar(10) default 'text' not null constraint cr_items_storage_type check (storage_type in ('lob','text','file')), storage_area_key varchar(100) default 'CR_FILES' not null, tree_sortkey varbit ); create index cr_items_by_locale on cr_items(locale); create index cr_items_by_content_type on cr_items(content_type); create unique index cr_items_by_live_revision on cr_items(live_revision); create unique index cr_items_by_latest_revision on cr_items(latest_revision); create unique index cr_items_unique_name on cr_items(parent_id, name); create unique index cr_items_unique_id on cr_items(parent_id, item_id); create index cr_items_by_parent_id on cr_items(parent_id); create index cr_sortkey_idx on cr_items(tree_sortkey); -- content-create.sql patch -- -- adds standard mechanism for deleting revisions from the file-system -- -- Walter McGinnis (wtem@olywa.net), 2001-09-23 -- based on original photo-album package code by Tom Baginski -- create table cr_files_to_delete ( path varchar(250), storage_area_key varchar(100) ); comment on table cr_files_to_delete is ' Table to store files to be deleted by a scheduled sweep. Since binaries are stored in filesystem and attributes in database, need a way to delete both atomically. So any process to delete file-system cr_revisions, copies the file path to this table as part of the delete transaction. Sweep run later to remove the files from filesystem once database info is successfully deleted. '; -- DCW, this can't be defined in the apm_package_versions table defintion, -- because cr_items is created afterwards. alter table apm_package_versions add constraint apm_package_ver_item_id_fk foreign key (item_id) references cr_items(item_id); create function cr_items_get_tree_sortkey(integer) returns varbit as ' declare p_item_id alias for $1; begin return tree_sortkey from cr_items where item_id = p_item_id; end;' language 'plpgsql'; create function cr_items_tree_insert_tr () returns opaque as ' declare v_parent_sk varbit default null; v_max_value integer; begin if new.parent_id is null then select max(tree_leaf_key_to_int(tree_sortkey)) into v_max_value from cr_items where parent_id is null; else select max(tree_leaf_key_to_int(tree_sortkey)) into v_max_value from cr_items where parent_id = new.parent_id; select tree_sortkey into v_parent_sk from cr_items where item_id = new.parent_id; end if; new.tree_sortkey := tree_next_key(v_parent_sk, v_max_value); return new; end;' language 'plpgsql'; create trigger cr_items_tree_insert_tr before insert on cr_items for each row execute procedure cr_items_tree_insert_tr (); create function cr_items_tree_update_tr () returns opaque as ' declare v_parent_sk varbit default null; v_max_value integer; p_id integer; v_rec record; clr_keys_p boolean default ''t''; begin if new.item_id = old.item_id and ((new.parent_id = old.parent_id) or (new.parent_id is null and old.parent_id is null)) then return new; end if; for v_rec in select item_id from cr_items where tree_sortkey between new.tree_sortkey and tree_right(new.tree_sortkey) order by tree_sortkey LOOP if clr_keys_p then update cr_items set tree_sortkey = null where tree_sortkey between new.tree_sortkey and tree_right(new.tree_sortkey); clr_keys_p := ''f''; end if; select parent_id into p_id from cr_items where item_id = v_rec.item_id; if p_id is null then select max(tree_leaf_key_to_int(tree_sortkey)) into v_max_value from cr_items where parent_id is null; else select max(tree_leaf_key_to_int(tree_sortkey)) into v_max_value from cr_items where parent_id = p_id; select tree_sortkey into v_parent_sk from cr_items where item_id = p_id; end if; update cr_items set tree_sortkey = tree_next_key(v_parent_sk, v_max_value) where item_id = v_rec.item_id; end LOOP; return new; end;' language 'plpgsql'; create trigger cr_items_tree_update_tr after update on cr_items for each row execute procedure cr_items_tree_update_tr (); comment on table cr_items is ' Each content item has a row in this table, as well as a row in the acs_objects table. The parent_id is used to place an item in a directory or place it within another container item. '; comment on column cr_items.content_type is ' The content type constrains the type of revision that may be added to this item (an item should have revisions of only one type). If null, then no revisions should be allowed. '; create table cr_child_rels ( rel_id integer constraint cr_child_rels_rel_pk primary key constraint cr_child_rels_rel_fk references acs_objects, parent_id integer constraint cr_child_rels_parent_nil not null, child_id integer constraint cr_child_rels_child_nil not null, relation_tag varchar(100), order_n integer ); create index cr_child_rels_by_parent on cr_child_rels(parent_id); create unique index cr_child_rels_unq_id on cr_child_rels(parent_id, child_id); comment on table cr_child_rels is ' Provides for richer parent-child relationships than the simple link encapsulated in the primary table. May be subclassed to provide additional attributes. '; create table cr_item_rels ( rel_id integer constraint cr_item_rels_pk primary key constraint cr_item_rels_fk references acs_objects, item_id integer constraint cr_item_rels_item_fk references cr_items, related_object_id integer constraint cr_item_rels_rel_obj__fk references acs_objects, relation_tag varchar(100), order_n integer ); create unique index cr_item_rel_unq on cr_item_rels ( item_id, related_object_id, relation_tag ); -- RI Indexes create index cr_item_rels_rel_obj_id_idx ON cr_item_rels(related_object_id); comment on table cr_item_rels is ' Describes all relations from one item to any number of other objects. '; comment on column cr_item_rels.relation_tag is ' A token for lightweight classification of item relationships. If additional attributes are required, then a subtype of item_rel may be created. '; comment on column cr_item_rels.order_n is ' Optional column for specifying a sort order. Note that the ordering method is application-dependent (it may be by relation type or across multiple relation types). '; -------------------------------------------------------------- -- CONTENT REVISIONS -------------------------------------------------------------- -- Define the cr_revisions table create table cr_revisions ( revision_id integer constraint cr_revisions_rev_id_fk references acs_objects (object_id) on delete cascade constraint cr_revisions_pk primary key, item_id integer constraint cr_revisions_item_id_nil not null constraint cr_revisions_item_id_fk references cr_items on delete cascade, title varchar(1000), description text, publish_date timestamptz, mime_type varchar(200) default 'text/plain' constraint cr_revisions_mime_type_ref references cr_mime_types, nls_language varchar(50), -- lob_id if storage_type = lob. lob integer references lobs, -- content holds the file name if storage type = file -- otherwise it holds the text data if storage_type = text. content text, content_length integer ); -- RI Indexes create index cr_revisions_lob_idx ON cr_revisions(lob); create index cr_revisions_item_id_idx ON cr_revisions(item_id); create trigger cr_revisions_lob_trig before delete or update or insert on cr_revisions for each row execute procedure on_lob_ref(); create index cr_revisions_by_mime_type on cr_revisions(mime_type); create index cr_revisions_title_idx on cr_revisions(title); -- create index cr_revisions_lower_title_idx on cr_revisions(lower(title)); -- create index cr_revisions_title_ltr_idx on cr_revisions(substr(lower(title), 1, 1)); comment on table cr_revisions is ' Each content item may be associated with any number of revisions. The item_id is used to associate revisions with an item. '; comment on column cr_revisions.nls_language is ' NLS_LANGUAGE is required in the same table as the content column for multi-lingual searching in Intermedia. '; -- postgresql RI bug causes multiple failures with regards to deletion of -- content_revisions (DanW dcwickstrom@earthlink.net) -- alter table cr_items add constraint cr_items_live_fk -- foreign key (live_revision) references cr_revisions(revision_id); -- alter table cr_items add constraint cr_items_latest_fk -- foreign key (latest_revision) references cr_revisions(revision_id); create function cr_revision_del_ri_trg() returns opaque as ' declare dummy integer; v_latest integer; v_live integer; begin select 1 into dummy from cr_revisions where revision_id = old.live_revision; if FOUND then raise EXCEPTION ''Referential Integrity: live_revision still exists: %'', old.live_revision; end if; select 1 into dummy from cr_revisions where revision_id = old.latest_revision; if FOUND then raise EXCEPTION ''Referential Integrity: latest_revision still exists: %'', old.latest_revision; end if; return old; end;' language 'plpgsql'; create function cr_revision_ins_ri_trg() returns opaque as ' declare dummy integer; v_latest integer; v_live integer; begin select 1 into dummy from cr_revisions where revision_id = new.live_revision; if NOT FOUND and new.live_revision is NOT NULL then raise EXCEPTION ''Referential Integrity: live_revision does not exist: %'', new.live_revision; end if; select 1 into dummy from cr_revisions where revision_id = new.latest_revision; if NOT FOUND and new.latest_revision is NOT NULL then raise EXCEPTION ''Referential Integrity: latest_revision does not exist: %'', new.latest_revision; end if; return new; end;' language 'plpgsql'; create function cr_revision_up_ri_trg() returns opaque as ' declare dummy integer; v_latest integer; v_live integer; begin select 1 into dummy from cr_revisions where revision_id = new.live_revision; if NOT FOUND and new.live_revision <> old.live_revision and new.live_revision is NOT NULL then raise EXCEPTION ''Referential Integrity: live_revision does not exist: %'', new.live_revision; end if; select 1 into dummy from cr_revisions where revision_id = new.latest_revision; if NOT FOUND and new.latest_revision <> old.latest_revision and new.latest_revision is NOT NULL then raise EXCEPTION ''Referential Integrity: latest_revision does not exist: %'', new.latest_revision; end if; return new; end;' language 'plpgsql'; create function cr_revision_del_rev_ri_trg() returns opaque as ' declare dummy integer; begin select 1 into dummy from cr_items where item_id = old.item_id and live_revision = old.revision_id; if FOUND then raise EXCEPTION ''Referential Integrity: attempting to delete live_revision: %'', old.revision_id; end if; select 1 into dummy from cr_items where item_id = old.item_id and latest_revision = old.revision_id; if FOUND then raise EXCEPTION ''Referential Integrity: attempting to delete latest_revision: %'', old.revision_id; end if; return old; end;' language 'plpgsql'; -- reimplementation of RI triggers. (DanW dcwickstrom@earthlink.net) create trigger cr_revision_del_ri_trg after delete on cr_items for each row execute procedure cr_revision_del_ri_trg(); create trigger cr_revision_up_ri_trg after update on cr_items for each row execute procedure cr_revision_up_ri_trg(); create trigger cr_revision_ins_ri_trg after insert on cr_items for each row execute procedure cr_revision_ins_ri_trg(); create trigger cr_revision_del_rev_ri_trg after delete on cr_revisions for each row execute procedure cr_revision_del_rev_ri_trg(); -- (DanW - OpenACS) Added cleanup trigger to log file items that need -- to be cleaned up from the CR. create function cr_cleanup_cr_files_del_trg() returns opaque as ' declare begin insert into cr_files_to_delete select r.content as path, i.storage_area_key from cr_items i, cr_revisions r where i.item_id = r.item_id and r.revision_id = old.revision_id and i.storage_type = ''file''; return old; end;' language 'plpgsql'; create trigger cr_cleanup_cr_files_del_trg before delete on cr_revisions for each row execute procedure cr_cleanup_cr_files_del_trg(); create table cr_revision_attributes ( revision_id integer constraint cr_revision_attributes_pk primary key constraint cr_revision_attributes_fk references cr_revisions, attributes text ); comment on column cr_revision_attributes.attributes is ' An XML document representing the compiled attributes for a revision '; -- create global temporary table cr_content_text ( -- revision_id integer primary key, -- content CLOB -- ) on commit delete rows; create table cr_content_text ( revision_id integer primary key, content integer ); comment on table cr_content_text is ' A temporary table for holding text extracted from the content blob. Provides a workaround for the fact that blob_to_string(content) has 4000 character limit. '; -------------------------------------------------------------- -- CONTENT PUBLISHING -------------------------------------------------------------- create table cr_item_publish_audit ( item_id integer, old_revision integer, new_revision integer, old_status varchar(40), new_status varchar(40), publish_date timestamptz constraint cr_item_publish_audit_date_nil not null ); create index cr_item_publish_audit_idx on cr_item_publish_audit(item_id); comment on table cr_item_publish_audit is ' An audit table (populated by a trigger on cr_items.live_revision) that is used to keep track of the publication history of an item. '; create table cr_release_periods ( item_id integer constraint cr_release_periods_fk references cr_items constraint cr_release_periods_pk primary key, start_when timestamptz default current_timestamp, end_when timestamptz default current_timestamp + interval '20 years' ); create table cr_scheduled_release_log ( exec_date timestamptz default current_timestamp not null, items_released integer not null, items_expired integer not null, err_num integer, err_msg varchar(500) ); comment on table cr_scheduled_release_log is ' Maintains a record, including any exceptions that may have aborted processing, for each scheduled update of live content. '; create table cr_scheduled_release_job ( job_id integer, last_exec timestamptz ); comment on table cr_scheduled_release_job is ' One-row table to track job ID of scheduled release update. '; insert into cr_scheduled_release_job values (NULL, now()); -------------------------------------------------------------- -- CONTENT FOLDERS -------------------------------------------------------------- create table cr_folders ( folder_id integer -- removed due to postgresql RI bug which causes deletion failures. -- replace with user triggers -- DanW (dcwickstrom@earthlink.net) -- constraint cr_folder_id_fk references -- cr_items on delete cascade constraint cr_folders_pk primary key, label varchar(1000), description text, has_child_folders boolean default 'f', has_child_symlinks boolean default 'f', package_id integer constraint cr_fldr_pkg_id_fk references apm_packages ); comment on table cr_folders is ' Folders are used to support a virtual file system within the content repository. '; --RI Indexes create index cr_folders_package_id_idx ON cr_folders(package_id); create function cr_folder_ins_up_ri_trg() returns opaque as ' declare dummy integer; v_latest integer; v_live integer; begin select 1 into dummy from cr_items where item_id = new.folder_id; if NOT FOUND then raise EXCEPTION ''Referential Integrity: folder_id does not exist in cr_items: %'', new.folder_id; end if; return new; end;' language 'plpgsql'; create function cr_folder_del_ri_trg() returns opaque as ' declare dummy integer; begin delete from cr_folders where folder_id = old.item_id; return old; end;' language 'plpgsql'; -- reimplementation of RI triggers. (DanW dcwickstrom@earthlink.net) create trigger cr_folder_ins_up_ri_trg before insert or update on cr_folders for each row execute procedure cr_folder_ins_up_ri_trg(); create trigger cr_folder_del_ri_trg before delete on cr_items for each row execute procedure cr_folder_del_ri_trg(); create table cr_folder_type_map ( folder_id integer constraint cr_folder_type_map_fldr_fk references cr_folders, content_type varchar(100) constraint cr_folder_type_map_typ_fk references acs_object_types, constraint cr_folder_type_map_pk primary key (folder_id, content_type) ); comment on table cr_folder_type_map is ' A one-to-many mapping table of content folders to content types. Basically, this table restricts the content types a folder may contain. Future releases will add numeric and tagged constraints similar to thos available for content types. '; -- RI Indexes create index cr_folder_typ_map_cont_typ_idx ON cr_folder_type_map(content_type); -------------------------------------------------------------- -- CONTENT TEMPLATES -------------------------------------------------------------- create table cr_templates ( template_id integer constraint cr_template_id_fk references cr_items on delete cascade constraint cr_templates_pk primary key ); comment on table cr_templates is ' Templates are a special class of text objects that are used for specifying the layout of a content item. They may be mapped to content types for defaults, or may be mapped to individual content items. '; create table cr_template_use_contexts ( use_context varchar(100) constraint cr_template_use_contexts_pk primary key ); comment on table cr_template_use_contexts is ' A simple table (for now) for constraining template use contexts. '; insert into cr_template_use_contexts values ('admin'); insert into cr_template_use_contexts values ('public'); create table cr_type_template_map ( content_type varchar(100) constraint cr_type_template_map_typ_fk references acs_object_types constraint cr_type_template_map_typ_nil not null, template_id integer constraint cr_type_template_map_tmpl_fk references cr_templates, use_context varchar(100) constraint cr_type_template_map_ctx_nil not null constraint cr_type_template_map_ctx_fk references cr_template_use_contexts, is_default boolean default 'f', constraint cr_type_template_map_pk primary key (content_type, template_id, use_context) ); create index cr_ttmap_by_content_type on cr_type_template_map(content_type); create index cr_ttmap_by_template_id on cr_type_template_map(template_id); create index cr_ttmap_by_use_context on cr_type_template_map(use_context); comment on table cr_type_template_map is ' A simple mapping template among content types and templates. Used to determine the default template to use in any particular context, as well as for building any UI that allows publishers to choose from a palette of templates. '; comment on column cr_type_template_map.use_context is ' A token to indicate the context in which a template is appropriate, such as admin or public. Should be constrained when it becomes clearer how this will be used. '; create table cr_item_template_map ( item_id integer constraint cr_item_template_map_item_fk references cr_items constraint cr_item_template_map_item_nil not null, template_id integer constraint cr_item_template_map_tmpl_fk references cr_templates constraint cr_item_template_map_tmpl_nil not null, use_context varchar(100) constraint cr_item_template_map_ctx_nil not null constraint cr_item_template_map_ctx_fk references cr_template_use_contexts, constraint cr_item_template_map_pk primary key (item_id, template_id, use_context) ); create index cr_itmap_by_item_id on cr_item_template_map(item_id); create index cr_itmap_by_template_id on cr_item_template_map(template_id); create index cr_itmap_by_use_context on cr_item_template_map(use_context); comment on table cr_item_template_map is ' Allows a template to be assigned to a specific item. '; -------------------------------------------------------------- -- CONTENT SYMLINKS -------------------------------------------------------------- create table cr_symlinks ( symlink_id integer constraint cr_symlink_id_fk references cr_items on delete cascade constraint cr_symlinks_pk primary key, target_id integer constraint cr_symlink_target_id_fk references cr_items constraint cr_symlink_target_id_nil not null, label varchar(1000) ); create index cr_symlinks_by_target_id on cr_symlinks(target_id); comment on table cr_symlinks is ' Symlinks are pointers to items within the content repository. '; -------------------------------------------------------------- -- CONTENT EXTLINKS -------------------------------------------------------------- create table cr_extlinks ( extlink_id integer constraint cr_extlink_id_fk references cr_items on delete cascade constraint cr_extlinks_pk primary key, url varchar(1000) constraint cr_extlink_url_nil not null, label varchar(1000) constraint cr_extlink_label_nil not null, description text ); comment on table cr_extlinks is ' Extlinks are pointers to items anywhere on the web which the publisher wishes to categorize, index and relate to items in the content repository. '; -------------------------------------------------------------- -- CONTENT KEYWORDS -------------------------------------------------------------- create table cr_keywords ( keyword_id integer constraint cr_keywords_pk primary key, parent_id integer constraint cr_keywords_hier references cr_keywords, heading varchar(600) constraint cr_keywords_name_nil not null, description text, has_children boolean, tree_sortkey varbit ); -- RI Indexes create index cr_keywords_parent_id_idx ON cr_keywords(parent_id); create function cr_keywords_get_tree_sortkey(integer) returns varbit as ' declare p_keyword_id alias for $1; begin return tree_sortkey from cr_keywords where keyword_id = p_keyword_id; end;' language 'plpgsql'; create function cr_keywords_tree_insert_tr () returns opaque as ' declare v_parent_sk varbit default null; v_max_value integer; begin if new.parent_id is null then select max(tree_leaf_key_to_int(tree_sortkey)) into v_max_value from cr_keywords where parent_id is null; else select max(tree_leaf_key_to_int(tree_sortkey)) into v_max_value from cr_keywords where parent_id = new.parent_id; select tree_sortkey into v_parent_sk from cr_keywords where keyword_id = new.parent_id; end if; new.tree_sortkey := tree_next_key(v_parent_sk, v_max_value); return new; end;' language 'plpgsql'; create trigger cr_keywords_tree_insert_tr before insert on cr_keywords for each row execute procedure cr_keywords_tree_insert_tr (); create function cr_keywords_tree_update_tr () returns opaque as ' declare v_parent_sk varbit default null; v_max_value integer; p_id integer; v_rec record; clr_keys_p boolean default ''t''; begin if new.keyword_id = old.keyword_id and ((new.parent_id = old.parent_id) or (new.parent_id is null and old.parent_id is null)) THEN return new; end if; for v_rec in select keyword_id from cr_keywords where tree_sortkey between new.tree_sortkey and tree_right(new.tree_sortkey) order by tree_sortkey LOOP if clr_keys_p then update cr_keywords set tree_sortkey = null where tree_sortkey between new.tree_sortkey and tree_right(new.tree_sortkey); clr_keys_p := ''f''; end if; select parent_id into p_id from cr_keywords where keyword_id = v_rec.keyword_id; if p_id is null then select max(tree_leaf_key_to_int(tree_sortkey)) into v_max_value from cr_keywords where parent_id is null; else select max(tree_leaf_key_to_int(tree_sortkey)) into v_max_value from cr_keywords where parent_id = p_id; select tree_sortkey into v_parent_sk from cr_keywords where keyword_id = p_id; end if; update cr_keywords set tree_sortkey = tree_next_key(v_parent_sk, v_max_value) where keyword_id = v_rec.keyword_id; end LOOP; return new; end;' language 'plpgsql'; create trigger cr_keywords_tree_update_tr after update on cr_keywords for each row execute procedure cr_keywords_tree_update_tr (); comment on table cr_keywords is ' Stores a subject taxonomy for classifying content items, analogous to the system used by a library. '; comment on column cr_keywords.heading is ' A subject heading. This will become a message ID in the next release so it should never be referenced directly (only through the API) '; comment on column cr_keywords.description is ' Description of a subject heading. This will be a message ID in the next release so it should never be referenced directly (only through the API) '; create table cr_item_keyword_map ( item_id integer constraint cr_item_keyword_map_item_fk references cr_items constraint cr_item_keyword_map_item_nil not null, keyword_id integer constraint cr_item_keyword_map_kw_fk references cr_keywords constraint cr_item_keyword_map_kw_nil not null, constraint cr_item_keyword_map_pk primary key (item_id, keyword_id) ); -- RI Indexes create index cr_item_keyword_map_kw_id_idx ON cr_item_keyword_map(keyword_id); -------------------------------------------------------------- -- TEXT SUBMISSION -------------------------------------------------------------- create table cr_text ( text_data text ); comment on table cr_text is ' A simple placeholder table for generating input views, so that a complete revision may be added with a single INSERT statement. '; insert into cr_text values (NULL); create function cr_text_tr () returns opaque as ' begin raise EXCEPTION ''-20000: Inserts are not allowed into cr_text.''; return new; end;' language 'plpgsql'; create trigger cr_text_tr before insert on cr_text for each row execute procedure cr_text_tr (); -- show errors -------------------------------------------------------------- -- DOCUMENT SUBMISSION WITH CONVERSION TO HTML -------------------------------------------------------------- -- create global temporary table cr_doc_filter ( -- revision_id integer primary key, -- content BLOB --) on commit delete rows; create table cr_doc_filter ( revision_id integer primary key, -- content BLOB -- need a blob trigger here content integer ); -- Source PL/SQL Definitions. \i content-util.sql \i content-xml.sql -- prompt *** Creating packaged call specs for Java utility methods... \i content-package.sql -- prompt *** Defining and compiling packages... \i packages-create.sql -- prompt *** Creating object types... \i types-create.sql -- this index requires prefs created in content-search -- create index cr_doc_filter_index on cr_doc_filter ( content ) -- indextype is ctxsys.context -- parameters ('FILTER content_filter_pref' ); comment on table cr_doc_filter is ' A temporary table for holding binary documents that are to be converted into HTML (or plain text) prior to insertion into the repository. '; -- prompt *** Compiling documentation package... \i doc-package.sql -- prompt *** Creating image content type... \i content-image.sql -- by default, map all MIME types to 'content_revision' create function inline_1 () returns integer as ' declare v_id integer; begin PERFORM content_type__register_mime_type(''content_revision'', ''text/html''); PERFORM content_type__register_mime_type(''content_revision'', ''text/plain''); PERFORM content_type__register_mime_type(''content_revision'', ''application/rtf''); v_id := content_folder__new ( ''pages'', ''Pages'', ''Site pages go here'', 0, null, content_item__get_root_folder(null), now(), null, null ); PERFORM content_folder__register_content_type( v_id, ''content_revision'', ''t'' ); PERFORM content_folder__register_content_type( v_id, ''content_folder'', ''t'' ); PERFORM content_folder__register_content_type( v_id, ''content_symlink'', ''t'' ); v_id := content_folder__new ( ''templates'', ''Templates'', ''Templates which render the pages go here'', 0, null, content_template__get_root_folder(), now(), null, null ); PERFORM content_folder__register_content_type( v_id, ''content_folder'', ''t'' ); PERFORM content_folder__register_content_type( v_id, ''content_symlink'', ''t'' ); PERFORM content_folder__register_content_type( v_id, ''content_template'', ''t'' ); return 0; end;' language 'plpgsql'; select inline_1 (); drop function inline_1 (); create function inline_2 () returns integer as ' declare v_item_id integer; v_revision_id integer; begin select acs_object_id_seq.nextval into v_item_id; PERFORM content_template__new( ''default_template'', ''-200'', v_item_id, now(), null, null ); v_revision_id := content_revision__new( ''Template'', NULL, now(), ''text/html'', null, '''', v_item_id, NULL, now(), null, null); update cr_revisions set content_length = length(content) where revision_id = v_revision_id; update cr_items set live_revision = v_revision_id where item_id = v_item_id; PERFORM content_type__register_template( ''content_revision'', v_item_id, ''public'', ''t''); PERFORM content_type__register_template( ''image'', v_item_id, ''public'', ''t''); -- testing, this may go away. DanW PERFORM content_type__register_template( ''content_template'', v_item_id, ''public'', ''t''); return 0; end;' language 'plpgsql'; select inline_2 (); drop function inline_2 (); -- this was added for edit-this-page and others -- 05-Nov-2001 Jon Griffin jon@mayuli.com alter table cr_folders add constraint cr_flder_pkg_id_fk foreign key (package_id) references apm_packages (package_id); --constraint cr_fldr_pkg_id_fk -- prompt *** Preparing search indices... \i content-search.sql