optimize refine_persons function (#1993)

* optimize refine_persons function

* change fd to upc ( user points of contacts)

Author: malek10xdev

supabase/migrations/20250122185859_optimize_refine_persons.sql

@@ -0,0 +1,120 @@
+DROP FUNCTION private.get_grouped_tags_by_person;
+DROP FUNCTION private.refine_persons;
+CREATE OR REPLACE FUNCTION private.refine_persons(userid uuid)
+RETURNS void
+LANGUAGE plpgsql
+SET search_path = ''
+AS $$
+BEGIN
+    CREATE TEMP TABLE user_points_of_contact AS
+        SELECT
+            poc.person_email,
+            poc.name,
+            poc.plus_address,
+            poc.message_id,
+            poc."from",
+            poc."to",
+            poc.bcc,
+            poc.cc,
+            poc.reply_to,
+            m.date,
+            m.conversation
+        FROM private.pointsofcontact poc
+        JOIN private.messages m 
+            ON poc.message_id = m.message_id
+            AND poc.user_id = m.user_id
+        WHERE poc.user_id = userid;
+
+    CREATE TEMP TABLE grouped_tags AS
+       SELECT
+         person_email AS email,
+         array_agg(name) AS tags
+       FROM private.tags tg
+       WHERE
+        tg.user_id = userid
+        AND reachable IN (1, 2)
+        AND person_email NOT IN (
+            SELECT person_email
+            FROM private.tags tg
+            WHERE tg.user_id = userid
+            GROUP BY person_email
+            HAVING
+                MAX(reachable) = 3
+                  OR (count(DISTINCT reachable) = 1 AND MAX(reachable) = 3)
+                  OR (count(DISTINCT reachable) = 2 AND ARRAY[1, 3] <@ array_agg(DISTINCT reachable))
+          )
+        GROUP BY person_email;
+
+    CREATE TEMP TABLE name_aggregates AS
+        SELECT
+            upc.person_email,
+            upc.name,
+            MAX(upc.date) AS recent_date,
+            COUNT(*) AS total,
+            array_agg(upc.name) OVER (PARTITION BY upc.person_email) AS alternate_name
+        FROM user_points_of_contact upc
+        WHERE upc.name IS NOT NULL
+        GROUP BY upc.person_email, upc.name;
+
+    CREATE TEMP TABLE real_names AS
+        SELECT DISTINCT
+            na.person_email,
+            FIRST_VALUE(na.name) OVER (
+                PARTITION BY na.person_email
+                ORDER BY na.total DESC, na.recent_date DESC
+            ) AS preferred_name,
+            na.alternate_name
+        FROM name_aggregates na;
+
+    CREATE TEMP TABLE email_aggregates AS
+        SELECT
+            upc.person_email,
+            MAX(upc.date) AS recency,
+            MIN(upc.date) AS seniority,
+            private.get_distinct_or_exclude_from_array(array_agg(upc.plus_address)) AS alternate_email,
+            COUNT(*) AS occurrence,
+            COUNT(CASE WHEN upc."from" = true OR upc.reply_to = true THEN 1 END) AS sender,
+            COUNT(CASE WHEN upc."to" = true OR upc.bcc = true OR upc.cc = true THEN 1 END) AS recipient,
+            COUNT(CASE WHEN upc.conversation = true THEN 1 END) AS conversations,
+            COUNT(CASE WHEN upc.conversation = true AND upc."from" = true THEN 1 END) AS replied_conversations
+        FROM user_points_of_contact upc
+        GROUP BY upc.person_email;
+
+    CREATE TEMP TABLE combined_data AS
+        SELECT
+            ea.*,
+            gt.tags as tags,
+            pn.preferred_name AS name,
+            private.get_distinct_or_exclude_from_array(pn.alternate_name, pn.preferred_name) AS alternate_name
+        FROM email_aggregates ea
+        LEFT JOIN real_names pn ON ea.person_email = pn.person_email
+        JOIN grouped_tags gt ON ea.person_email = gt.email;
+
+    INSERT INTO private.refinedpersons (
+        user_id, email, occurrence, recency, seniority,
+        sender, recipient, conversations, replied_conversations, tags
+    )
+    SELECT
+        userid,
+        cd.person_email,
+        cd.occurrence,
+        cd.recency,
+        cd.seniority,
+        cd.sender,
+        cd.recipient,
+        cd.conversations,
+        cd.replied_conversations,
+        cd.tags
+    FROM combined_data cd
+    ON CONFLICT (user_id, email) DO UPDATE
+    SET
+        occurrence = EXCLUDED.occurrence,
+        recency = EXCLUDED.recency,
+        seniority = EXCLUDED.seniority,
+        sender = EXCLUDED.sender,
+        recipient = EXCLUDED.recipient,
+        conversations = EXCLUDED.conversations,
+        replied_conversations = EXCLUDED.replied_conversations,
+        tags = EXCLUDED.tags;
+END;
+$$;