Merge pull request #117 from gabriel-curtino/master

Use rowid instead id, allowing custom primary keys

Author: oldmoe

lib/litestack/litesearch/index.rb

@@ -103,21 +103,21 @@ def search(term, options = {})
     generate_results(rs)
   end
 
-  def similar(id, limit = 10)
+  def similar(rowid, limit = 10)
     #   pp term = @db.execute(@schema.sql_for(:similarity_query), id)
     rs = if @schema.schema[:tokenizer] == :trigram
       # just use the normal similarity approach for now
       # need to recondisder that for trigram indexes later
-      @stmts[:similar].execute(id, limit) # standard:disable Style/IdenticalConditionalBranches
+      @stmts[:similar].execute(rowid, limit) # standard:disable Style/IdenticalConditionalBranches
     else
-      @stmts[:similar].execute(id, limit) # standard:disable Style/IdenticalConditionalBranches
+      @stmts[:similar].execute(rowid, limit) # standard:disable Style/IdenticalConditionalBranches
     end
 
     generate_results(rs)
   end
 
   def clear!
-    @stmts[:delete_all].execute!(id)
+    @stmts[:delete_all].execute!(rowid)
   end
 
   def drop!

lib/litestack/litesearch/model.rb

@@ -24,24 +24,18 @@ def search_models
   end
 
   module InstanceMethods
+    # rowid = id by default
+    def rowid
+      id
+    end
+
     def similar(limit = 10)
-      conn = self.class.get_connection
-      idx = conn.search_index(self.class.send(:index_name))
-      r_a_h = conn.results_as_hash
-      conn.results_as_hash = true
-      rs = idx.similar(id, limit)
-      conn.results_as_hash = r_a_h
-      result = []
-      rs.each do |row|
-        obj = self.class.fetch_row(row["id"])
-        obj.search_rank = row["search_rank"]
-        result << obj
-      end
-      result
+      self.class.similar(rowid, limit)
     end
   end
 
   module ClassMethods
+
     def litesearch
       # it is possible that this code is running when there is no table created yet
       if !defined?(ActiveRecord::Base).nil? && ancestors.include?(ActiveRecord::Base)
@@ -90,6 +84,22 @@ def drop_index!
       get_connection.search_index(index_name).drop!
     end
 
+    def similar(rowid, limit = 10)
+      conn = get_connection
+      idx = conn.search_index(send(:index_name))
+      r_a_h = conn.results_as_hash
+      conn.results_as_hash = true
+      rs = idx.similar(rowid, limit)
+      conn.results_as_hash = r_a_h
+      result = []
+      rs.each do |row|
+        obj = fetch_row(row["rowid"])
+        obj.search_rank = row["search_rank"]
+        result << obj
+      end
+      result
+    end
+
     def search_all(term, options = {})
       options[:offset] ||= 0
       options[:limit] ||= 25
@@ -136,6 +146,7 @@ def create_instance(row)
   end
 
   module ActiveRecordSchemaMethods
+
     attr_accessor :model_class
 
     def field(name, attributes = {})
@@ -161,17 +172,26 @@ def field(name, attributes = {})
     def allowed_attributes
       super + [:polymorphic, :as, :action_text]
     end
+
   end
 
-  module ActiveRecordInstanceMethods; end
+  module ActiveRecordInstanceMethods
+    def rowid
+      self.class.rowid(id)
+    end
+  end
 
   module ActiveRecordClassMethods
     def get_connection
       connection.raw_connection
     end
 
-    def fetch_row(id)
-      find(id)
+    def rowid(id)
+      where(primary_key => id).limit(1).pluck(:rowid)&.first
+    end
+
+    def fetch_row(rowid)
+      find_by("rowid = ?", rowid)
     end
 
     def search(term)
@@ -185,7 +205,7 @@ def search(term)
       self.select(
         "#{table_name}.*"
       ).joins(
-        "INNER JOIN #{index_name} ON #{table_name}.id = #{index_name}.rowid AND rank != 0 AND #{index_name} MATCH ", Arel.sql("'#{term}'")
+        "INNER JOIN #{index_name} ON #{table_name}.rowid = #{index_name}.rowid AND rank != 0 AND #{index_name} MATCH ", Arel.sql("'#{term}'")
       ).select(
         "-#{index_name}.rank AS search_rank"
       ).order(
@@ -199,6 +219,10 @@ def create_instance(row)
   end
 
   module SequelInstanceMethods
+    def rowid
+      self.class.rowid(id)
+    end
+
     def search_rank
       @values[:search_rank]
     end
@@ -209,8 +233,12 @@ def search_rank=(rank)
   end
 
   module SequelClassMethods
-    def fetch_row(id)
-      self[id]
+    def rowid(id)
+      where(primary_key => id).get(:rowid)
+    end
+
+    def fetch_row(rowid)
+      self[rowid]  # where(Sequel.lit("rowid = ?", rowid)).first
     end
 
     def get_connection
@@ -221,7 +249,7 @@ def search(term)
       dataset.select(
         Sequel.lit("#{table_name}.*, -#{index_name}.rank AS search_rank")
       ).inner_join(
-        Sequel.lit("#{index_name}(:term) ON #{table_name}.id = #{index_name}.rowid AND rank != 0", {term: term})
+        Sequel.lit("#{index_name}(:term) ON #{table_name}.rowid = #{index_name}.rowid AND rank != 0", {term: term})
       ).order(
         Sequel.lit("rank")
       )

lib/litestack/litesearch/schema.rb

@@ -52,6 +52,10 @@ def table(table_name)
     @schema[:table] = table_name
   end
 
+  def primary_key(new_primary_key)
+    @schema[:primary_key] = new_primary_key
+  end
+
   def fields(field_names)
     field_names.each { |f| field f }
   end
@@ -185,7 +189,7 @@ def clean
   end
 
   def allowed_attributes
-    [:weight, :col, :target, :source, :conditions, :reference]
+    [:weight, :col, :target, :source, :conditions, :reference, :primary_key]
   end
 end
 

lib/litestack/litesearch/schema_adapters/backed_adapter.rb

@@ -46,7 +46,7 @@ def create_primary_triggers_sql(active = false)
     DELETE FROM #{name} WHERE rowid = NEW.rowid;
   END;
   CREATE TRIGGER #{name}_delete AFTER DELETE ON #{table} BEGIN
-    DELETE FROM #{name} WHERE rowid = OLD.id;
+    DELETE FROM #{name} WHERE rowid = OLD.rowid;
   END;
     SQL
   end
@@ -59,10 +59,10 @@ def drop_secondary_trigger_poly_sql(target_table, target_col, col)
     "DROP TRIGGER IF EXISTS #{target_table}_#{target_col}_#{name}_update;"
   end
 
-  def create_secondary_trigger_sql(target_table, target_col, col)
+  def create_secondary_trigger_sql(target_table, target_col, col, primary_key)
     <<~SQL
       CREATE TRIGGER IF NOT EXISTS #{target_table}_#{target_col}_#{col}_#{name}_update AFTER UPDATE OF #{target_col} ON #{target_table} BEGIN
-        #{rebuild_sql} AND #{table}.#{col} = NEW.id;
+        #{rebuild_sql} AND #{table}.#{col} = NEW.#{primary_key};
       END;
     SQL
   end
@@ -98,7 +98,7 @@ def create_secondary_triggers_sql
     @schema[:fields].each do |name, field|
       if field[:trigger_sql]
         if field[:col]
-          sql << create_secondary_trigger_sql(field[:target_table], field[:target_col], field[:col])
+          sql << create_secondary_trigger_sql(field[:target_table], field[:target_col], field[:col], field[:primary_key])
         elsif field[:source]
           sql << create_secondary_trigger_poly_sql(field[:target_table], field[:target_col], name, field[:conditions])
         end
@@ -108,18 +108,19 @@ def create_secondary_triggers_sql
   end
 
   def rebuild_sql
-    "INSERT OR REPLACE INTO #{name}(rowid, #{active_field_names.join(", ")}) SELECT #{table}.id, #{select_cols_sql} FROM #{joins_sql} #{filter_sql}"
+    "INSERT OR REPLACE INTO #{name}(rowid, #{active_field_names.join(", ")}) SELECT #{table}.rowid, #{select_cols_sql} FROM #{joins_sql} #{filter_sql}"
   end
 
   def enrich_schema
     @schema[:fields].each do |name, field|
       if field[:target] && !field[:target].start_with?("#{table}.")
         field[:target] = field[:target].downcase
         target_table, target_col = field[:target].split(".")
+        field[:primary_key] = :id unless field[:primary_key]
         field[:col] = :"#{name}_id" unless field[:col]
         field[:target_table] = target_table.to_sym
         field[:target_col] = target_col.to_sym
-        field[:sql] = "(SELECT #{field[:target_col]} FROM #{field[:target_table]} WHERE id = NEW.#{field[:col]})"
+        field[:sql] = "(SELECT #{field[:target_col]} FROM #{field[:target_table]} WHERE #{field[:primary_key]} = NEW.#{field[:col]})"
         field[:trigger_sql] = true # create_secondary_trigger_sql(field[:target_table], field[:target_col], field[:col])
         field[:target_table_alias] = "#{field[:target_table]}_#{name}"
       elsif field[:source]
@@ -167,9 +168,9 @@ def joins_sql
         join_table = +""
         join_table << "#{field[:target_table]} AS #{field[:target_table_alias]} ON "
         if field[:col]
-          join_table << "#{field[:target_table_alias]}.id = #{@schema[:table]}.#{field[:col]}" if field[:col]
+          join_table << "#{field[:target_table_alias]}.#{field[:primary_key]} = #{@schema[:table]}.#{field[:col]}" if field[:col]
         elsif field[:source]
-          join_table << "#{field[:target_table_alias]}.#{field[:reference]} = #{@schema[:table]}.id"
+          join_table << "#{field[:target_table_alias]}.#{field[:reference]} = #{@schema[:table]}.rowid"
           if field[:conditions]
             join_table << " AND "
             join_table << field[:conditions].collect { |k, v| "#{field[:target_table_alias]}.#{k} = '#{v}'" }.join(" AND ")

lib/litestack/litesearch/schema_adapters/basic_adapter.rb

@@ -14,6 +14,10 @@ def table
     @schema[:table]
   end
 
+  def primary_key
+    @schema[:primary_key] || :id
+  end
+
   def fields
     @schema[:fields]
   end
@@ -78,8 +82,8 @@ def sql_for(method, *args)
   def generate_sql
     @sql[:create_index] = :create_index_sql
     @sql[:create_vocab_tables] = :create_vocab_tables_sql
-    @sql[:insert] = "INSERT OR REPLACE INTO #{name}(rowid, #{active_col_names_sql}) VALUES (:id, #{active_col_names_var_sql}) RETURNING rowid"
-    @sql[:delete] = "DELETE FROM #{name} WHERE rowid = :id"
+    @sql[:insert] = "INSERT OR REPLACE INTO #{name}(rowid, #{active_col_names_sql}) VALUES (:rowid, #{active_col_names_var_sql}) RETURNING rowid"
+    @sql[:delete] = "DELETE FROM #{name} WHERE rowid = :rowid"
     @sql[:count] = "SELECT count(*) FROM #{name}(:term)"
     @sql[:count_all] = "SELECT count(*) FROM #{name}"
     @sql[:delete_all] = "DELETE FROM #{name}"
@@ -89,11 +93,11 @@ def generate_sql
     @sql[:ranks] = :ranks_sql
     @sql[:set_config_value] = "INSERT OR REPLACE INTO #{name}_config(k, v) VALUES (:key, :value)"
     @sql[:get_config_value] = "SELECT v FROM #{name}_config WHERE k = :key"
-    @sql[:search] = "SELECT rowid AS id, -rank AS search_rank FROM #{name}(:term) WHERE rank !=0 ORDER BY rank LIMIT :limit OFFSET :offset"
-    @sql[:similarity_terms] = "SELECT DISTINCT term FROM #{name}_instance WHERE doc = :id AND FLOOR(term) IS NULL AND LENGTH(term) > 2 AND NOT instr(term, ' ') AND NOT instr(term, '-') AND NOT instr(term, ':') AND NOT instr(term, '#') AND NOT instr(term, '_') LIMIT 15"
-    @sql[:similarity_query] = "SELECT group_concat('\"' || term || '\"', ' OR ') FROM #{name}_row WHERE term IN (#{@sql[:similarity_terms]})"
-    @sql[:similarity_search] = "SELECT rowid AS id, -rank AS search_rank FROM #{name}(:term) WHERE rowid != :id ORDER BY rank LIMIT :limit"
-    @sql[:similar] = "SELECT rowid AS id, -rank AS search_rank FROM #{name} WHERE #{name} = (#{@sql[:similarity_query]}) AND rowid != :id ORDER BY rank LIMIT :limit"
+    @sql[:search] = "SELECT rowid, -rank AS search_rank FROM #{name}(:term) WHERE rank !=0 ORDER BY rank LIMIT :limit OFFSET :offset"
+    @sql[:similarity_terms] = "SELECT DISTINCT term FROM #{name}_instance WHERE doc = :rowid AND FLOOR(term) IS NULL AND LENGTH(term) > 2 AND NOT instr(term, ' ') AND NOT instr(term, '-') AND NOT instr(term, ':') AND NOT instr(term, '#') AND NOT instr(term, '_') LIMIT 15"
+    @sql[:similarity_query] = "SELECT group_concat(term, ' OR ') FROM #{name}_row WHERE term IN (#{@sql[:similarity_terms]})"
+    @sql[:similarity_search] = "SELECT rowid, -rank AS search_rank FROM #{name}(:term) WHERE rowid != :rowid ORDER BY rank LIMIT :limit"
+    @sql[:similar] = "SELECT rowid, -rank AS search_rank FROM #{name} WHERE #{name} = (#{@sql[:similarity_query]}) AND rowid != :rowid ORDER BY rank LIMIT :limit"
     @sql[:update_index] = "UPDATE sqlite_schema SET sql = :sql WHERE name = '#{name}'"
     @sql[:update_content_table] = "UPDATE sqlite_schema SET sql = :sql WHERE name = '#{name}_content'"
   end

lib/litestack/litesearch/schema_adapters/standalone_adapter.rb

@@ -5,11 +5,11 @@ def generate_sql
     @sql[:adjust_temp_content] = "UPDATE sqlite_schema SET sql (SELECT sql FROM sqlite_schema WHERE name = '#{name}_content') WHERE name = #{name}_content_temp"
     @sql[:restore_content] = "ALTER TABLE #{name}_content_temp RENAME TO #{name}_content"
     @sql[:rebuild] = "INSERT INTO #{name}(#{name}) VALUES ('rebuild')"
-    @sql[:similar] = "SELECT rowid AS id, *, -rank AS search_rank FROM #{name} WHERE #{name} = (#{@sql[:similarity_query]}) AND rowid != :id ORDER BY rank LIMIT :limit"
+    @sql[:similar] = "SELECT rowid, *, -rank AS search_rank FROM #{name} WHERE #{name} = (#{@sql[:similarity_query]}) AND rowid != :rowid ORDER BY rank LIMIT :limit"
     @sql[:drop_content_table] = "DROP TABLE #{name}_content"
     @sql[:drop_content_col] = :drop_content_col_sql
     @sql[:create_content_table] = :create_content_table_sql
-    @sql[:search] = "SELECT rowid AS id, *, -rank AS search_rank FROM #{name}(:term) WHERE rank !=0 ORDER BY rank LIMIT :limit OFFSET :offset"
+    @sql[:search] = "SELECT rowid, *, -rank AS search_rank FROM #{name}(:term) WHERE rank !=0 ORDER BY rank LIMIT :limit OFFSET :offset"
   end
 
   private
@@ -26,6 +26,6 @@ def drop_content_col_sql(col_index)
   def create_content_table_sql(count)
     cols = []
     count.times { |i| cols << "c#{i}" }
-    "CREATE TABLE #{name}_content(id INTEGER PRIMARY KEY, #{cols.join(", ")})"
+    "CREATE TABLE #{name}_content(rowid INTEGER PRIMARY KEY, #{cols.join(", ")})"
   end
 end