107 lines
5.7 KiB
SQL
107 lines
5.7 KiB
SQL
-- Data-quality rule evolution evidence. Additive and safe to run repeatedly on MySQL 8.
|
|
-- This migration does not activate rules or change any asset lifecycle state.
|
|
|
|
CREATE TABLE IF NOT EXISTS `aihr_processing_rule` (
|
|
`id` bigint NOT NULL AUTO_INCREMENT,
|
|
`tenant_id` varchar(20) NOT NULL,
|
|
`rule_code` varchar(64) NOT NULL,
|
|
`version_no` int NOT NULL,
|
|
`stage` varchar(30) NOT NULL,
|
|
`risk_class` varchar(20) NOT NULL,
|
|
`action` varchar(20) NOT NULL,
|
|
`implementation_type` varchar(30) NOT NULL,
|
|
`scope_json` json DEFAULT NULL,
|
|
`config_json` json NOT NULL,
|
|
`status` varchar(20) NOT NULL DEFAULT 'DRAFT',
|
|
`owner_id` bigint NOT NULL,
|
|
`supersedes_rule_id` bigint DEFAULT NULL,
|
|
`shadow_started_time` datetime DEFAULT NULL,
|
|
`retired_time` datetime DEFAULT NULL,
|
|
`create_time` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP,
|
|
`update_time` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
|
|
PRIMARY KEY (`id`),
|
|
UNIQUE KEY `uk_aihr_processing_rule_version` (`tenant_id`, `rule_code`, `version_no`),
|
|
KEY `idx_aihr_processing_rule_status` (`tenant_id`, `status`, `stage`, `update_time`),
|
|
CONSTRAINT `ck_aihr_processing_rule_stage` CHECK (`stage` IN
|
|
('INGESTION','PARSING','NORMALIZATION','STRUCTURING','DEDUPLICATION','PRIVACY','DOMAIN','QUALITY','PUBLISHING','INDEXING')),
|
|
CONSTRAINT `ck_aihr_processing_rule_risk` CHECK (`risk_class` IN ('LOW','MEDIUM','HIGH','CRITICAL')),
|
|
CONSTRAINT `ck_aihr_processing_rule_action` CHECK (`action` IN ('FLAG','REVIEW','QUARANTINE')),
|
|
CONSTRAINT `ck_aihr_processing_rule_implementation` CHECK (`implementation_type` IN
|
|
('DETERMINISTIC','STATISTICAL','LLM_ASSISTED')),
|
|
CONSTRAINT `ck_aihr_processing_rule_status` CHECK (`status` IN
|
|
('DRAFT','SHADOW','CANARY','ACTIVE','PAUSED','RETIRED'))
|
|
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci COMMENT='Versioned data-quality processing rules';
|
|
|
|
CREATE TABLE IF NOT EXISTS `aihr_processing_rule_transition` (
|
|
`id` bigint NOT NULL AUTO_INCREMENT,
|
|
`tenant_id` varchar(20) NOT NULL,
|
|
`rule_id` bigint NOT NULL,
|
|
`from_status` varchar(20) NOT NULL,
|
|
`to_status` varchar(20) NOT NULL,
|
|
`reason` varchar(500) NOT NULL,
|
|
`operator_id` bigint NOT NULL,
|
|
`create_time` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP,
|
|
PRIMARY KEY (`id`),
|
|
KEY `idx_aihr_rule_transition_rule` (`tenant_id`, `rule_id`, `create_time`),
|
|
CONSTRAINT `ck_aihr_rule_transition_from` CHECK (`from_status` IN
|
|
('DRAFT','SHADOW','CANARY','ACTIVE','PAUSED','RETIRED')),
|
|
CONSTRAINT `ck_aihr_rule_transition_to` CHECK (`to_status` IN
|
|
('DRAFT','SHADOW','CANARY','ACTIVE','PAUSED','RETIRED'))
|
|
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci COMMENT='Audited processing-rule status transitions';
|
|
|
|
CREATE TABLE IF NOT EXISTS `aihr_rule_evaluation` (
|
|
`id` bigint NOT NULL AUTO_INCREMENT,
|
|
`tenant_id` varchar(20) NOT NULL,
|
|
`rule_id` bigint NOT NULL,
|
|
`asset_id` bigint DEFAULT NULL,
|
|
`version_id` bigint DEFAULT NULL,
|
|
`sample_key` varchar(100) NOT NULL,
|
|
`evaluation_mode` varchar(20) NOT NULL DEFAULT 'SHADOW',
|
|
`expected_decision` varchar(20) NOT NULL,
|
|
`actual_decision` varchar(20) NOT NULL,
|
|
`confidence` decimal(6,5) DEFAULT NULL,
|
|
`expected_source` varchar(20) NOT NULL,
|
|
`matched_reason_codes_json` json DEFAULT NULL,
|
|
`false_allow` tinyint(1) NOT NULL DEFAULT 0,
|
|
`false_block` tinyint(1) NOT NULL DEFAULT 0,
|
|
`run_id` varchar(64) NOT NULL,
|
|
`evaluator_id` bigint NOT NULL,
|
|
`create_time` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP,
|
|
PRIMARY KEY (`id`),
|
|
UNIQUE KEY `uk_aihr_rule_evaluation_sample` (`tenant_id`, `rule_id`, `run_id`, `sample_key`),
|
|
KEY `idx_aihr_rule_evaluation_metrics` (`tenant_id`, `rule_id`, `evaluation_mode`, `create_time`),
|
|
KEY `idx_aihr_rule_evaluation_asset` (`tenant_id`, `asset_id`, `version_id`),
|
|
CONSTRAINT `ck_aihr_rule_evaluation_mode` CHECK (`evaluation_mode` IN ('SHADOW','CANARY')),
|
|
CONSTRAINT `ck_aihr_rule_expected_decision` CHECK (`expected_decision` IN ('PASS','REVIEW','BLOCK')),
|
|
CONSTRAINT `ck_aihr_rule_actual_decision` CHECK (`actual_decision` IN ('PASS','REVIEW','BLOCK')),
|
|
CONSTRAINT `ck_aihr_rule_expected_source` CHECK (`expected_source` IN ('HUMAN','GOLDEN')),
|
|
CONSTRAINT `ck_aihr_rule_false_allow` CHECK (`false_allow` IN (0,1)),
|
|
CONSTRAINT `ck_aihr_rule_false_block` CHECK (`false_block` IN (0,1))
|
|
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci COMMENT='Shadow and canary rule evaluations against human or golden labels';
|
|
|
|
CREATE TABLE IF NOT EXISTS `aihr_review_sample` (
|
|
`id` bigint NOT NULL AUTO_INCREMENT,
|
|
`tenant_id` varchar(20) NOT NULL,
|
|
`rule_id` bigint NOT NULL,
|
|
`source_evaluation_id` bigint DEFAULT NULL,
|
|
`asset_id` bigint DEFAULT NULL,
|
|
`version_id` bigint DEFAULT NULL,
|
|
`sample_key` varchar(100) NOT NULL,
|
|
`sampling_strategy` varchar(30) NOT NULL,
|
|
`status` varchar(20) NOT NULL DEFAULT 'PENDING',
|
|
`review_result` varchar(30) DEFAULT NULL,
|
|
`review_note` varchar(1000) DEFAULT NULL,
|
|
`sampled_by` bigint NOT NULL,
|
|
`sampled_time` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP,
|
|
`reviewed_by` bigint DEFAULT NULL,
|
|
`reviewed_time` datetime DEFAULT NULL,
|
|
PRIMARY KEY (`id`),
|
|
UNIQUE KEY `uk_aihr_review_sample` (`tenant_id`, `rule_id`, `sample_key`),
|
|
KEY `idx_aihr_review_sample_queue` (`tenant_id`, `status`, `sampling_strategy`, `sampled_time`),
|
|
CONSTRAINT `ck_aihr_review_sample_strategy` CHECK (`sampling_strategy` IN
|
|
('RANDOM','RISK_WEIGHTED','MISMATCH','MANUAL')),
|
|
CONSTRAINT `ck_aihr_review_sample_status` CHECK (`status` IN ('PENDING','REVIEWED')),
|
|
CONSTRAINT `ck_aihr_review_sample_result` CHECK (`review_result` IS NULL OR `review_result` IN
|
|
('CONFIRMED','FALSE_ALLOW','FALSE_BLOCK','NOT_APPLICABLE'))
|
|
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci COMMENT='Risk-based human review samples for automated rule outcomes';
|