From f10f0b21f93739f18ed865023e78645a791a7231 Mon Sep 17 00:00:00 2001
From: icarus <eurfelux@gmail.com>
Date: Mon, 20 Oct 2025 05:08:24 +0800
Subject: [PATCH] Revert "refactor(db): remove unused ocr provider schema
 table"

This reverts commit 9c740f82ad58dc25d23a8ad688ab3d1cbb615dab.
---
 src/main/data/db/schemas/ocr/provider.ts | 50 ++++++++++++++++++++++++
 1 file changed, 50 insertions(+)
 create mode 100644 src/main/data/db/schemas/ocr/provider.ts

diff --git a/src/main/data/db/schemas/ocr/provider.ts b/src/main/data/db/schemas/ocr/provider.ts
new file mode 100644
index 0000000000..75b2281ec2
--- /dev/null
+++ b/src/main/data/db/schemas/ocr/provider.ts
@@ -0,0 +1,50 @@
+import type { OcrProviderCapabilityRecord, OcrProviderConfig } from '@types'
+import { index, sqliteTable, text } from 'drizzle-orm/sqlite-core'
+
+import { createUpdateTimestamps } from '../columnHelpers'
+
+export const ocrProviderTable = sqliteTable(
+  'ocr_provider',
+  {
+    /**
+     * Unique identifier for the provider.
+     * For built-in providers, it's 'tesseract', 'system', etc.
+     * For custom providers, it can be any unique string (we typically use UUID v4).
+     * As the primary key, it ensures the uniqueness of each provider.
+     */
+    id: text('id').primaryKey(),
+
+    /**
+     * Display name of the provider, e.g., "Tesseract OCR".
+     * For built-in providers, this value is used internally and is not exposed to users; the display name shown in the UI is locale-based by i18n.
+     * Cannot be null.
+     */
+    name: text('name').notNull(),
+
+    /**
+     * Object describing the provider's capabilities, e.g., { image: true }.
+     * Stored as JSON in a text column. Drizzle's `mode: 'json'` handles
+     * serialization and deserialization automatically. `$type` provides strong typing.
+     * Cannot be null; should store an empty object `{}` even if no specific capabilities.
+     */
+    capabilities: text('capabilities', { mode: 'json' }).$type<OcrProviderCapabilityRecord>().notNull(),
+
+    /**
+     * Provider-specific configuration. This is a polymorphic field, its structure varies by provider type.
+     * For example, Tesseract's configuration is entirely different from PaddleOCR's.
+     * Storing it as JSON is the most flexible approach to accommodate any configuration structure.
+     * Since this is a polymorphic field, both frontend and backend must validate
+     * that the structure matches the expected schema for the corresponding provider type
+     * before saving.
+     * This field is nullable because `config` in the `OcrProvider` type is optional.
+     */
+    config: text('config', { mode: 'json' }).$type<OcrProviderConfig>(),
+
+    /** Timestamps. May not useful. */
+    ...createUpdateTimestamps
+  },
+  (t) => [index('name').on(t.name)]
+)
+
+export type OcrProviderInsert = typeof ocrProviderTable.$inferInsert
+export type OcrProviderSelect = typeof ocrProviderTable.$inferSelect