Add idempotent migration to repair SAG email threading schema

migrations/159_repair_sag_email_threading_schema.sql

@@ -0,0 +1,104 @@
+-- Migration 159: Repair SAG email threading schema (idempotent)
+-- Purpose:
+-- 1) Ensure sag_emails link table exists and is constrained correctly.
+-- 2) Ensure email_messages has threading columns used by SAG email tab.
+-- 3) Backfill thread_key where missing.
+
+-- Ensure link table exists
+CREATE TABLE IF NOT EXISTS sag_emails (
+    sag_id INTEGER REFERENCES sag_sager(id) ON DELETE CASCADE,
+    email_id INTEGER REFERENCES email_messages(id) ON DELETE CASCADE,
+    created_at TIMESTAMPTZ DEFAULT NOW()
+);
+
+-- Ensure required columns on email_messages exist
+ALTER TABLE email_messages
+    ADD COLUMN IF NOT EXISTS in_reply_to VARCHAR(500),
+    ADD COLUMN IF NOT EXISTS email_references TEXT,
+    ADD COLUMN IF NOT EXISTS thread_key VARCHAR(500);
+
+-- Cleanup duplicates before adding unique constraint/PK
+WITH ranked AS (
+    SELECT ctid,
+           ROW_NUMBER() OVER (PARTITION BY sag_id, email_id ORDER BY created_at NULLS LAST, ctid) AS rn
+    FROM sag_emails
+)
+DELETE FROM sag_emails se
+USING ranked r
+WHERE se.ctid = r.ctid
+  AND r.rn > 1;
+
+-- Ensure PK exists
+DO $$
+BEGIN
+    IF NOT EXISTS (
+        SELECT 1
+        FROM pg_constraint
+        WHERE conname = 'sag_emails_pkey'
+          AND conrelid = 'sag_emails'::regclass
+    ) THEN
+        ALTER TABLE sag_emails
+            ADD CONSTRAINT sag_emails_pkey PRIMARY KEY (sag_id, email_id);
+    END IF;
+END $$;
+
+-- Ensure FKs exist
+DO $$
+BEGIN
+    IF NOT EXISTS (
+        SELECT 1
+        FROM pg_constraint
+        WHERE conname = 'sag_emails_sag_id_fkey'
+          AND conrelid = 'sag_emails'::regclass
+    ) THEN
+        ALTER TABLE sag_emails
+            ADD CONSTRAINT sag_emails_sag_id_fkey
+            FOREIGN KEY (sag_id) REFERENCES sag_sager(id) ON DELETE CASCADE;
+    END IF;
+
+    IF NOT EXISTS (
+        SELECT 1
+        FROM pg_constraint
+        WHERE conname = 'sag_emails_email_id_fkey'
+          AND conrelid = 'sag_emails'::regclass
+    ) THEN
+        ALTER TABLE sag_emails
+            ADD CONSTRAINT sag_emails_email_id_fkey
+            FOREIGN KEY (email_id) REFERENCES email_messages(id) ON DELETE CASCADE;
+    END IF;
+END $$;
+
+-- Helpful indexes for case email tab
+CREATE INDEX IF NOT EXISTS idx_sag_emails_sag_id ON sag_emails(sag_id);
+CREATE INDEX IF NOT EXISTS idx_sag_emails_email_id ON sag_emails(email_id);
+CREATE INDEX IF NOT EXISTS idx_email_messages_thread_key ON email_messages(thread_key) WHERE thread_key IS NOT NULL;
+CREATE INDEX IF NOT EXISTS idx_email_messages_in_reply_to ON email_messages(in_reply_to) WHERE in_reply_to IS NOT NULL;
+
+-- Backfill thread_key for rows where it is missing
+UPDATE email_messages
+SET thread_key = LOWER(
+    REGEXP_REPLACE(
+        COALESCE(
+            NULLIF(
+                SPLIT_PART(
+                    REGEXP_REPLACE(COALESCE(email_references, ''), '^[\s<>,]+', ''),
+                    ' ',
+                    1
+                ),
+                ''
+            ),
+            NULLIF(in_reply_to, ''),
+            NULLIF(message_id, '')
+        ),
+        '[<>\s]',
+        '',
+        'g'
+    )
+)
+WHERE (thread_key IS NULL OR TRIM(thread_key) = '')
+  AND COALESCE(NULLIF(email_references, ''), NULLIF(in_reply_to, ''), NULLIF(message_id, '')) IS NOT NULL;
+
+COMMENT ON TABLE sag_emails IS 'Emails linked to the Case (SAG).';
+COMMENT ON COLUMN email_messages.in_reply_to IS 'Raw In-Reply-To header used for SAG threading lookup';
+COMMENT ON COLUMN email_messages.email_references IS 'Raw References header used for SAG threading lookup';
+COMMENT ON COLUMN email_messages.thread_key IS 'Stable normalized thread key for grouping email conversations';