rajch_ales/product-import-ai
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Prepare scrape image storage schema

Browse Source
This commit is contained in:
rajch_ales
2026-08-05 23:47:57 +02:00
parent aa138b62ad
commit 05338a062a
3 changed files with 108 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
sql/local-scrape-db.sql
+75
View File
@@ -127,6 +127,81 @@ CREATE TABLE IF NOT EXISTS `scrape_assets` (
CONSTRAINT `fk_scrape_assets_product` FOREIGN KEY (`scrape_product_id`) REFERENCES `scrape_products` (`id`)
) ENGINE=InnoDB;
CREATE TABLE IF NOT EXISTS `scrape_image_blobs` (
`id` BIGINT UNSIGNED NOT NULL AUTO_INCREMENT,
`sha256` CHAR(64) NOT NULL,
`perceptual_hash` CHAR(16) NULL,
`mime_type` VARCHAR(64) NOT NULL,
`width` INT UNSIGNED NULL,
`height` INT UNSIGNED NULL,
`size_bytes` INT UNSIGNED NOT NULL,
`image_data` LONGBLOB NOT NULL,
`created_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
PRIMARY KEY (`id`),
UNIQUE KEY `uq_scrape_image_blobs_sha256` (`sha256`),
KEY `idx_scrape_image_blobs_phash` (`perceptual_hash`),
KEY `idx_scrape_image_blobs_size` (`width`, `height`, `size_bytes`)
) ENGINE=InnoDB;
CREATE TABLE IF NOT EXISTS `scrape_picture_candidates` (
`id` BIGINT UNSIGNED NOT NULL AUTO_INCREMENT,
`image_blob_id` BIGINT UNSIGNED NOT NULL,
`source_id` BIGINT UNSIGNED NOT NULL,
`scrape_product_id` BIGINT UNSIGNED NULL,
`manufacturer_id` INT UNSIGNED NULL,
`id_product_catalog` BIGINT UNSIGNED NULL,
`id_product` BIGINT UNSIGNED NULL,
`supplier_reference` VARCHAR(255) NULL,
`ean13` VARCHAR(32) NULL,
`color` VARCHAR(255) NULL,
`size_value` VARCHAR(128) NULL,
`combi` VARCHAR(255) NULL,
`source_page_url` VARCHAR(2048) NULL,
`source_image_url` VARCHAR(2048) NOT NULL,
`source_position` INT UNSIGNED NOT NULL DEFAULT 0,
`target_strategy` ENUM('unknown', 'color', 'size', 'combination', 'manual') NOT NULL DEFAULT 'unknown',
`target_color` VARCHAR(255) NULL,
`target_size` VARCHAR(128) NULL,
`target_combi` VARCHAR(255) NULL,
`target_ean13` VARCHAR(32) NULL,
`target_id_product_catalog` BIGINT UNSIGNED NULL,
`match_confidence` DECIMAL(5,2) NOT NULL DEFAULT 0.00,
`duplicate_status` ENUM('unique', 'exact_duplicate', 'visual_duplicate', 'review') NOT NULL DEFAULT 'review',
`duplicate_of_candidate_id` BIGINT UNSIGNED NULL,
`status` ENUM('draft', 'approved', 'rejected') NOT NULL DEFAULT 'draft',
`review_note` TEXT NULL,
`created_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
`updated_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
PRIMARY KEY (`id`),
UNIQUE KEY `uq_scrape_picture_candidate_source_image` (`source_id`, `source_image_url`(512)),
KEY `idx_scrape_picture_candidates_blob` (`image_blob_id`),
KEY `idx_scrape_picture_candidates_source` (`source_id`),
KEY `idx_scrape_picture_candidates_product` (`manufacturer_id`, `supplier_reference`, `ean13`),
KEY `idx_scrape_picture_candidates_target` (`target_strategy`, `target_color`, `target_size`, `target_combi`),
KEY `idx_scrape_picture_candidates_duplicate` (`duplicate_status`, `duplicate_of_candidate_id`),
CONSTRAINT `fk_scrape_picture_candidates_blob` FOREIGN KEY (`image_blob_id`) REFERENCES `scrape_image_blobs` (`id`),
CONSTRAINT `fk_scrape_picture_candidates_source` FOREIGN KEY (`source_id`) REFERENCES `scrape_sources` (`id`),
CONSTRAINT `fk_scrape_picture_candidates_product` FOREIGN KEY (`scrape_product_id`) REFERENCES `scrape_products` (`id`)
) ENGINE=InnoDB;
CREATE TABLE IF NOT EXISTS `scrape_picture_target_rules` (
`id` BIGINT UNSIGNED NOT NULL AUTO_INCREMENT,
`manufacturer_id` INT UNSIGNED NULL,
`source_id` BIGINT UNSIGNED NULL,
`id_category_default` INT UNSIGNED NULL,
`catalog_category` VARCHAR(255) NULL,
`product_name_pattern` VARCHAR(255) NULL,
`target_strategy` ENUM('color', 'size', 'combination', 'manual') NOT NULL,
`confidence` DECIMAL(5,2) NOT NULL DEFAULT 0.00,
`examples_count` INT UNSIGNED NOT NULL DEFAULT 0,
`note` TEXT NULL,
`created_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
`updated_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
PRIMARY KEY (`id`),
KEY `idx_scrape_picture_target_rules_lookup` (`manufacturer_id`, `source_id`, `id_category_default`),
CONSTRAINT `fk_scrape_picture_target_rules_source` FOREIGN KEY (`source_id`) REFERENCES `scrape_sources` (`id`)
) ENGINE=InnoDB;
CREATE TABLE IF NOT EXISTS `scrape_runs` (
`id` BIGINT UNSIGNED NOT NULL AUTO_INCREMENT,
`source_id` BIGINT UNSIGNED NULL,