-- Run this once on an existing MySQL database before deploying the
-- single-version SOP application. It is intentionally explicit: the Go
-- service only performs the data-preserving backfill and does not drop data.

SET @sql = (
  SELECT IF(COUNT(*) = 0,
    'ALTER TABLE sops ADD COLUMN start_node_key VARCHAR(64) NOT NULL DEFAULT ''start''',
    'SELECT 1')
  FROM information_schema.columns
  WHERE table_schema = DATABASE() AND table_name = 'sops' AND column_name = 'start_node_key'
);
PREPARE stmt FROM @sql;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;

UPDATE sops s
JOIN (
  SELECT sop_id, MAX(id) AS version_id
  FROM sop_versions
  WHERE status = 'published'
  GROUP BY sop_id
) current ON current.sop_id = s.id
JOIN sop_versions sv ON sv.id = current.version_id
SET s.start_node_key = COALESCE(NULLIF(sv.start_node_key, ''), NULLIF(s.start_node_key, ''), 'start');

SET @sql = (
  SELECT IF(COUNT(*) = 0,
    'ALTER TABLE sop_nodes ADD COLUMN sop_id BIGINT UNSIGNED NULL',
    'SELECT 1')
  FROM information_schema.columns
  WHERE table_schema = DATABASE() AND table_name = 'sop_nodes' AND column_name = 'sop_id'
);
PREPARE stmt FROM @sql;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;

SET @sql = (
  SELECT IF(COUNT(*) = 0,
    'ALTER TABLE sop_edges ADD COLUMN sop_id BIGINT UNSIGNED NULL',
    'SELECT 1')
  FROM information_schema.columns
  WHERE table_schema = DATABASE() AND table_name = 'sop_edges' AND column_name = 'sop_id'
);
PREPARE stmt FROM @sql;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;

SET @sql = (
  SELECT IF(COUNT(*) = 0,
    'ALTER TABLE sop_runs ADD COLUMN sop_id BIGINT UNSIGNED NULL',
    'SELECT 1')
  FROM information_schema.columns
  WHERE table_schema = DATABASE() AND table_name = 'sop_runs' AND column_name = 'sop_id'
);
PREPARE stmt FROM @sql;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;

UPDATE sop_nodes n JOIN sop_versions sv ON sv.id = n.sop_version_id
SET n.sop_id = sv.sop_id WHERE n.sop_id IS NULL;
UPDATE sop_edges e JOIN sop_versions sv ON sv.id = e.sop_version_id
SET e.sop_id = sv.sop_id WHERE e.sop_id IS NULL;
UPDATE sop_runs r JOIN sop_versions sv ON sv.id = r.sop_version_id
SET r.sop_id = sv.sop_id WHERE r.sop_id IS NULL;

-- A single-version SOP can only have one graph. Keep the latest published
-- version's nodes and edges; remove historical graph rows after the backup.
DELETE n
FROM sop_nodes n
JOIN sop_versions sv ON sv.id = n.sop_version_id
JOIN (
  SELECT sop_id, MAX(id) AS version_id
  FROM sop_versions
  WHERE status = 'published'
  GROUP BY sop_id
) current ON current.sop_id = sv.sop_id
WHERE sv.id <> current.version_id;

DELETE e
FROM sop_edges e
JOIN sop_versions sv ON sv.id = e.sop_version_id
JOIN (
  SELECT sop_id, MAX(id) AS version_id
  FROM sop_versions
  WHERE status = 'published'
  GROUP BY sop_id
) current ON current.sop_id = sv.sop_id
WHERE sv.id <> current.version_id;

ALTER TABLE sop_nodes MODIFY COLUMN sop_id BIGINT UNSIGNED NOT NULL;
ALTER TABLE sop_edges MODIFY COLUMN sop_id BIGINT UNSIGNED NOT NULL;
ALTER TABLE sop_runs MODIFY COLUMN sop_id BIGINT UNSIGNED NOT NULL;

SET @sql = (
  SELECT IF(COUNT(*) = 0,
    'ALTER TABLE sop_nodes ADD INDEX idx_sop_nodes_sop_id (sop_id)',
    'SELECT 1')
  FROM information_schema.statistics
  WHERE table_schema = DATABASE() AND table_name = 'sop_nodes' AND index_name = 'idx_sop_nodes_sop_id'
);
PREPARE stmt FROM @sql;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;

SET @sql = (
  SELECT IF(COUNT(*) = 0,
    'ALTER TABLE sop_edges ADD INDEX idx_sop_edges_sop_id (sop_id)',
    'SELECT 1')
  FROM information_schema.statistics
  WHERE table_schema = DATABASE() AND table_name = 'sop_edges' AND index_name = 'idx_sop_edges_sop_id'
);
PREPARE stmt FROM @sql;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;

SET @sql = (
  SELECT IF(COUNT(*) = 0,
    'ALTER TABLE sop_runs ADD INDEX idx_sop_runs_sop_id (sop_id)',
    'SELECT 1')
  FROM information_schema.statistics
  WHERE table_schema = DATABASE() AND table_name = 'sop_runs' AND index_name = 'idx_sop_runs_sop_id'
);
PREPARE stmt FROM @sql;
EXECUTE stmt;
DEALLOCATE PREPARE stmt;

-- Verify the backfill before the destructive cleanup below.
SELECT COUNT(*) AS nodes_without_sop FROM sop_nodes WHERE sop_id IS NULL;
SELECT COUNT(*) AS edges_without_sop FROM sop_edges WHERE sop_id IS NULL;
SELECT COUNT(*) AS runs_without_sop FROM sop_runs WHERE sop_id IS NULL;

-- After the three counts above are all zero, remove the old version fields.
-- The old columns/table are intentionally retained by this backfill step.
-- Run 20260819_drop_sop_versions.sql only after all three counts above are 0.