Update recommender_bayesian.sql, recommender_geographic.sql,...

Update recommender_bayesian.sql, recommender_geographic.sql, recommender_popularity.sql, recommender_preferences.sql, recommender_tags.sql

recommender_bayesian.sql

+-- Created on: 16/01/2023
+-- @author: Andoni Aranguren Ubierna
+-- Adaptations: 02/2023 @author: Sergio Campos
+
+DELIMITER $$
+CREATE DEFINER=`root`@`localhost` PROCEDURE `recommender_bayesian_action_id`(
+	IN action_id_array VARCHAR(255), 
+	IN kpi INT)
+BEGIN 
+	-- Temporary table so we can parse an set of ids
+	DROP TEMPORARY TABLE IF EXISTS temp_table;
+	CREATE TEMPORARY TABLE temp_table(action_id INT);
+    
+	INSERT INTO temp_table(action_id)
+	SELECT action_id
+	FROM action p
+	WHERE FIND_IN_SET(p.action_id, action_id_array);   
+	
+    -- Amount of actions in input set. Important to check if P(B/A) checks all input actions
+	SET @values_in_temp = (SELECT COUNT(*) FROM temp_table);
+
+	-- P(B|A)
+	DROP TEMPORARY TABLE IF EXISTS probabilidad_b_sub_a;
+	CREATE TEMPORARY TABLE probabilidad_b_sub_a(action_id INT, probability_b_a FLOAT);
+    
+	INSERT INTO probabilidad_b_sub_a(action_id, probability_b_a)
+	SELECT distinct_t.action_id, COUNT(distinct_t.action_id) as probability_b_a FROM CHOSEN_action distinct_t 
+	INNER JOIN (SELECT t.plan_id, t.time_slot, COUNT(t.action_id) as contador, t.action_id FROM CHOSEN_action t
+				WHERE action_id IN (SELECT * FROM temp_table) 
+				GROUP BY t.plan_id, t.time_slot) as t 
+				ON distinct_t.plan_id = t.plan_id
+	WHERE t.contador = @values_in_temp AND NOT FIND_IN_SET(distinct_t.action_id, action_id_array)
+	GROUP BY distinct_t.action_id
+	ORDER BY probability_b_a DESC;
+
+	SET @total_rutas_b = (SELECT COUNT(*) FROM probabilidad_b_sub_a);
+	-- At this actionnt probability is just a count so probability = count / total
+	UPDATE probabilidad_b_sub_a SET probability_b_a = probability_b_a / @total_rutas_b;
+
+	-- P(A)
+	SET @action_total = (SELECT Count(*) FROM CHOSEN_action);
+    
+	DROP TEMPORARY TABLE IF EXISTS probability_a;
+	CREATE TEMPORARY TABLE probability_a(action_id INT, probability_A Float);
+    
+	INSERT INTO probability_a(action_id, probability_A)
+	SELECT p.action_id, p.popularity/@action_total FROM probabilidad_b_sub_a p_b_a
+    INNER JOIN action p WHERE p.action_id = p_b_a.action_id;
+
+	-- P(B)
+	SET @counter_b = (SELECT COUNT(r.plan_id) as Count FROM plan_detail r 
+		INNER JOIN (SELECT t.plan_id, t.time_slot, COUNT(t.action_id) as contador, t.action_id FROM CHOSEN_action t
+					WHERE action_id IN (SELECT * FROM temp_table) 
+					GROUP BY t.plan_id, t.time_slot) as t 
+					ON r.plan_id = t.plan_id
+		WHERE t.contador = @values_in_temp);
+
+	SET @probability_b =  IF(@counter_b is null, 0, @counter_b/(SELECT Count(*) FROM plan_detail));
+    
+    -- P(A/B)
+    IF kpi is null THEN
+		SELECT 
+			p_a.action_id, action.action_name, action.lat, action.lon,
+			--  This lines are so we can get the list of kpis vvvvvvvvvv
+				GROUP_CONCAT(distinct t_list.kpi_id) as kpi_id_list,
+				GROUP_CONCAT(distinct t_list.kpi_name) as kpi_name_list,
+			--  This lines are so we can get the list of kpis ^^^^^^^^^^  
+			probability_a*probability_b_a/@probability_b as bayesian_probability, 
+			probability_a, probability_b_a, @probability_b as probability_b
+		FROM probability_a p_a
+		INNER JOIN probabilidad_b_sub_a p_b_a ON p_b_a.action_id = p_a.action_id
+		LEFT OUTER JOIN action action ON action.action_id = p_a.action_id
+        -- This lines are so we can get the list of kpis vvvvvvvvvv
+			INNER JOIN kpi_action t_p_list ON t_p_list.action_id = p_a.action_id 
+			INNER JOIN kpi t_list ON t_list.kpi_id = t_p_list.kpi_id
+		--  This lines are so we can get the list of kpis ^^^^^^^^^^
+        GROUP BY p_a.action_id
+		ORDER BY bayesian_probability DESC
+		LIMIT 100;
+	ELSE
+		SELECT 
+			p_a.action_id, action.action_name, action.lat, action.lon,
+			--  This lines are so we can get the list of kpis vvvvvvvvvv
+				GROUP_CONCAT(distinct t_list.kpi_id) as kpi_id_list,
+				GROUP_CONCAT(distinct t_list.kpi_name) as kpi_name_list,
+			--  This lines are so we can get the list of kpis ^^^^^^^^^^ 
+			probability_a*probability_b_a/@probability_b as bayesian_probability, 
+			probability_a, probability_b_a, @probability_b as probability_b
+		FROM probability_a p_a
+		INNER JOIN probabilidad_b_sub_a p_b_a ON p_b_a.action_id = p_a.action_id
+		LEFT OUTER JOIN action action ON action.action_id = p_a.action_id
+        INNER JOIN kpi_action t ON t.action_id = action.action_id
+        -- This lines are so we can get the list of kpis vvvvvvvvvv
+			INNER JOIN kpi_action t_p_list ON t_p_list.action_id = p_a.action_id 
+			INNER JOIN kpi t_list ON t_list.kpi_id = t_p_list.kpi_id
+		--  This lines are so we can get the list of kpis ^^^^^^^^^^
+        WHERE t.kpi_id = kpi
+        GROUP BY p_a.action_id
+		ORDER BY bayesian_probability DESC
+		LIMIT 100;
+	END IF;		
+END$$
+DELIMITER ;

recommender_geographic.sql

+-- Created on: 16/01/2023
+-- @author: Andoni Aranguren Ubierna
+-- Adaptations: 02/2023 @author: Sergio Campos
+
+DELIMITER $$
+CREATE DEFINER=`root`@`localhost` PROCEDURE `recommender_geographic_action_id`(IN action_id_in INT, IN kpi INT)
+BEGIN
+	SET @deg_to_km = 111.2;
+	IF kpi is null THEN
+		SELECT 
+			p.action_id, p.action_name, p.lon, p.lat, 
+			-- This lines are so we can get the list of kpis vvvvvvvvvv
+				GROUP_CONCAT(distinct t_list.kpi_id) as kpi_id_list,
+				GROUP_CONCAT(distinct t_list.kpi_name) as kpi_name_list,
+			-- This lines are so we can get the list of kpis ^^^^^^^^^^
+			POW(@deg_to_km * (p.lat - p_t.lat), 2) +
+			POW(@deg_to_km * (p_t.lon - p.lon) * COS(p.lat / 57.3), 2) AS distance_squared
+		FROM action p 
+		LEFT JOIN action p_t ON p_t.action_id = action_id_in
+        -- This lines are so we can get the list of kpis vvvvvvvvvv
+			INNER JOIN kpi_action t_p_list ON t_p_list.action_id = p.action_id 
+			INNER JOIN kpi t_list ON t_list.kpi_id = t_p_list.kpi_id
+        -- This lines are so we can get the list of kpis ^^^^^^^^^^
+		WHERE p_t.action_id != p.action_id
+        GROUP BY p.action_id
+		ORDER BY distance_squared ASC 
+		LIMIT 100;
+	ELSE
+		SELECT 
+			p.action_id, p.action_name, p.lon, p.lat, 
+			-- This lines are so we can get the list of kpis vvvvvvvvvv
+				GROUP_CONCAT(distinct t_list.kpi_id) as kpi_id_list,
+				GROUP_CONCAT(distinct t_list.kpi_name) as kpi_name_list,
+			-- This lines are so we can get the list of kpis ^^^^^^^^^^
+			POW(@deg_to_km * (p.lat - p_t.lat), 2) +
+			POW(@deg_to_km * (p_t.lon - p.lon) * COS(p.lat / 57.3), 2) AS distance_squared
+		FROM action p 
+		LEFT JOIN action p_t ON p_t.action_id = action_id_in
+        INNER JOIN kpi_action t ON t.action_id = p.action_id
+        -- This lines are so we can get the list of kpis vvvvvvvvvv
+			INNER JOIN kpi_action t_p_list ON t_p_list.action_id = p.action_id 
+			INNER JOIN kpi t_list ON t_list.kpi_id = t_p_list.kpi_id
+        -- This lines are so we can get the list of kpis ^^^^^^^^^^
+		WHERE p_t.action_id != p.action_id AND t.kpi_id = kpi
+        GROUP BY p.action_id
+		ORDER BY distance_squared ASC 
+		LIMIT 100;
+	END IF;
+END$$
+DELIMITER ;

recommender_popularity.sql

+-- Created on: 16/01/2023
+-- @author: Andoni Aranguren Ubierna
+-- Adaptations: 02/2023 @author: Sergio Campos
+
+recommender_kpis_action_idDELIMITER $$
+CREATE DEFINER=`root`@`localhost` PROCEDURE `recommender_popularity`(IN kpi INT)
+BEGIN
+	IF kpi is null THEN
+		SELECT p.action_id, p.action_name, p.lat, p.lon,
+		-- This lines are so we can get the list of kpis vvvvvvvvvv
+			GROUP_CONCAT(distinct t_list.kpi_id) as kpi_id_list,
+			GROUP_CONCAT(distinct t_list.kpi_name) as kpi_name_list,
+		-- This lines are so we can get the list of kpis ^^^^^^^^^^
+		p.popularity as popularity
+		FROM action p
+        -- This lines are so we can get the list of kpis vvvvvvvvvv
+			INNER JOIN kpi_action t_p_list ON t_p_list.action_id = p.action_id 
+			INNER JOIN kpi t_list ON t_list.kpi_id = t_p_list.kpi_id
+        -- This lines are so we can get the list of kpis ^^^^^^^^^^
+        GROUP BY p.action_id
+		ORDER BY popularity DESC
+		LIMIT 100;
+	ELSE
+		SELECT p.action_id, p.action_name, p.lat, p.lon,
+		-- This lines are so we can get the list of kpis vvvvvvvvvv
+			GROUP_CONCAT(distinct t_list.kpi_id) as kpi_id_list,
+			GROUP_CONCAT(distinct t_list.kpi_name) as kpi_name_list,
+		-- This lines are so we can get the list of kpis ^^^^^^^^^^
+		p.popularity as popularity
+		FROM action p
+        INNER JOIN kpi_action t ON t.action_id = p.action_id
+        -- This lines are so we can get the list of kpis vvvvvvvvvv
+			INNER JOIN kpi_action t_p_list ON t_p_list.action_id = p.action_id 
+			INNER JOIN kpi t_list ON t_list.kpi_id = t_p_list.kpi_id
+        -- This lines are so we can get the list of kpis ^^^^^^^^^^
+        WHERE t.kpi_id = kpi
+        GROUP BY p.action_id
+		ORDER BY popularity DESC
+		LIMIT 100;
+	END IF;
+END$$
+DELIMITER ;
+DELIMITER $$
+CREATE DEFINER=`root`@`localhost` PROCEDURE `recommender_geographic_action_id`(IN action_id_in INT, IN kpi INT)
+BEGIN
+	SET @deg_to_km = 111.2;
+	IF kpi is null THEN
+		SELECT 
+			p.action_id, p.action_name, p.lon, p.lat, 
+			-- This lines are so we can get the list of kpis vvvvvvvvvv
+				GROUP_CONCAT(distinct t_list.kpi_id) as kpi_id_list,
+				GROUP_CONCAT(distinct t_list.kpi_name) as kpi_name_list,
+			-- This lines are so we can get the list of kpis ^^^^^^^^^^
+			POW(@deg_to_km * (p.lat - p_t.lat), 2) +
+			POW(@deg_to_km * (p_t.lon - p.lon) * COS(p.lat / 57.3), 2) AS distance_squared
+		FROM action p 
+		LEFT JOIN action p_t ON p_t.action_id = action_id_in
+        -- This lines are so we can get the list of kpis vvvvvvvvvv
+			INNER JOIN kpi_action t_p_list ON t_p_list.action_id = p.action_id 
+			INNER JOIN kpi t_list ON t_list.kpi_id = t_p_list.kpi_id
+        -- This lines are so we can get the list of kpis ^^^^^^^^^^
+		WHERE p_t.action_id != p.action_id
+        GROUP BY p.action_id
+		ORDER BY distance_squared ASC 
+		LIMIT 100;
+	ELSE
+		SELECT 
+			p.action_id, p.action_name, p.lon, p.lat, 
+			-- This lines are so we can get the list of kpis vvvvvvvvvv
+				GROUP_CONCAT(distinct t_list.kpi_id) as kpi_id_list,
+				GROUP_CONCAT(distinct t_list.kpi_name) as kpi_name_list,
+			-- This lines are so we can get the list of kpis ^^^^^^^^^^
+			POW(@deg_to_km * (p.lat - p_t.lat), 2) +
+			POW(@deg_to_km * (p_t.lon - p.lon) * COS(p.lat / 57.3), 2) AS distance_squared
+		FROM action p 
+		LEFT JOIN action p_t ON p_t.action_id = action_id_in
+        INNER JOIN kpi_action t ON t.action_id = p.action_id
+        -- This lines are so we can get the list of kpis vvvvvvvvvv
+			INNER JOIN kpi_action t_p_list ON t_p_list.action_id = p.action_id 
+			INNER JOIN kpi t_list ON t_list.kpi_id = t_p_list.kpi_id
+        -- This lines are so we can get the list of kpis ^^^^^^^^^^
+		WHERE p_t.action_id != p.action_id AND t.kpi_id = kpi
+        GROUP BY p.action_id
+		ORDER BY distance_squared ASC 
+		LIMIT 100;
+	END IF;
+END$$
+DELIMITER ;
\ No newline at end of file

recommender_preferences.sql

+-- Created on: 16/01/2023
+-- @author: Andoni Aranguren Ubierna
+-- Adaptations: 02/2023 @author: Sergio Campos
+
+DELIMITER $$
+CREATE DEFINER=`root`@`localhost` PROCEDURE `recommender_preferences`(
+	IN emission_total FLOAT,
+    IN noise_total FLOAT,
+    IN kpi INT)
+BEGIN 
+	IF kpi IS NULL THEN
+		SELECT c_action.action_id, p.action_name, p.lat, p.lon,
+		--  This lines are so we can get the list of kpis vvvvvvvvvv
+			GROUP_CONCAT(distinct t_list.kpi_id) as kpi_id_list,
+			GROUP_CONCAT(distinct t_list.kpi_name) as kpi_name_list,
+		--  This lines are so we can get the list of kpis ^^^^^^^^^^
+		COUNT(c_action.action_id) as Count
+		FROM CHOSEN_action c_action
+		INNER JOIN           
+			(SELECT r.plan_id, 
+					r.time_slot, 
+					POWER(r.emission_total-emission_total,2) + POWER(r.noise_total-noise_total,2) as distance
+			FROM plan_detail r
+			ORDER BY distance ASC
+			LIMIT 5) as r ON c_action.plan_id = r.plan_id AND c_action.time_slot = r.time_slot
+		INNER JOIN action p ON c_action.action_id = p.action_id 
+        -- This lines are so we can get the list of kpis vvvvvvvvvv
+			INNER JOIN kpi_action t_p_list ON t_p_list.action_id = p.action_id 
+			INNER JOIN kpi t_list ON t_list.kpi_id = t_p_list.kpi_id
+		--  This lines are so we can get the list of kpis ^^^^^^^^^^
+		GROUP BY c_action.action_id
+		ORDER BY Count DESC
+		LIMIT 100;
+	ELSE
+		SELECT c_action.action_id, p.action_name, p.lat, p.lon,
+		--  This lines are so we can get the list of kpis vvvvvvvvvv
+			GROUP_CONCAT(distinct t_list.kpi_id) as kpi_id_list,
+			GROUP_CONCAT(distinct t_list.kpi_name) as kpi_name_list,
+		--  This lines are so we can get the list of kpis ^^^^^^^^^^
+		COUNT(c_action.action_id) as Count
+		FROM CHOSEN_action c_action
+		INNER JOIN           
+			(SELECT r.plan_id, 
+					r.time_slot, 
+					POWER(r.emission_total-emission_total,2) + POWER(r.noise_total-noise_total,2) as distance
+			FROM plan_detail r
+			ORDER BY distance ASC
+			LIMIT 5) as r ON c_action.plan_id = r.plan_id AND c_action.time_slot = r.time_slot
+		INNER JOIN action p ON c_action.action_id = p.action_id 
+        INNER JOIN kpi_action t ON t.action_id = p.action_id
+        -- This lines are so we can get the list of kpis vvvvvvvvvv
+			INNER JOIN kpi_action t_p_list ON t_p_list.action_id = p.action_id 
+			INNER JOIN kpi t_list ON t_list.kpi_id = t_p_list.kpi_id
+		--  This lines are so we can get the list of kpis ^^^^^^^^^^
+        WHERE t.kpi_id = kpi
+		GROUP BY c_action.action_id
+		ORDER BY Count DESC
+		LIMIT 100;
+	END IF;
+END$$
+DELIMITER ;

recommender_tags.sql

+-- Created on: 16/01/2023
+-- @author: Andoni Aranguren Ubierna
+-- Adaptations: 02/2023 @author: Sergio Campos
+
+DELIMITER //
+DELIMITER $$
+CREATE DEFINER=`root`@`localhost` PROCEDURE `recommender_kpis_action_id`(IN action_id INT)
+BEGIN 
+	SELECT kpi.kpi_id, kpi_name, count(*) as Count, kpi.popularity as Popularity, count(*)/kpi.popularity as Probability FROM kpi kpi
+		INNER JOIN (
+		SELECT DISTINCT t_p.action_id, t_p.kpi_id, c_p.plan_id, r.time_slot FROM kpi_action t_p
+			INNER JOIN CHOSEN_action c_p ON c_p.action_id = t_p.action_id
+			INNER JOIN plan_detail r ON r.plan_id = c_p.plan_id AND r.time_slot = c_p.time_slot
+			INNER JOIN CHOSEN_action target_p ON r.plan_id = target_p.plan_id AND r.time_slot = target_p.time_slot
+			WHERE target_p.action_id = action_id) as distinct_kpis ON kpi.kpi_id = distinct_kpis.kpi_id
+		GROUP BY kpi.kpi_id
+		ORDER BY Probability DESC;
+END$$
+DELIMITER ;
+
+
+DELIMITER $$
+CREATE DEFINER=`root`@`localhost` PROCEDURE `recommender_kpis_kpi_id`(IN kpi_id INT)
+BEGIN 
+SELECT kpi.kpi_id, kpi_name, count(*) as Count, kpi.popularity as Popularity, count(*)/kpi.popularity as Probability FROM kpi kpi
+	INNER JOIN (
+	SELECT DISTINCT t_p.action_id, t_p.kpi_id, c_p.plan_id, r.time_slot FROM kpi_action t_p
+		INNER JOIN CHOSEN_action c_p ON c_p.action_id = t_p.action_id
+		INNER JOIN plan_detail r ON r.plan_id = c_p.plan_id AND r.time_slot = c_p.time_slot
+		INNER JOIN CHOSEN_action target_p ON r.plan_id = target_p.plan_id AND r.time_slot = target_p.time_slot
+		INNER JOIN kpi_action target_kpi_p ON target_kpi_p.kpi_id != t_p.kpi_id AND target_p.action_id = target_kpi_p.action_id 
+		WHERE target_kpi_p.kpi_id = kpi_id) as distinct_kpis ON kpi.kpi_id = distinct_kpis.kpi_id
+	GROUP BY kpi.kpi_id
+	ORDER BY Probability DESC;
+END$$
+DELIMITER ;