# Category Import / Export API

> **Prefix:** `/api/v1/admin/categories/{bundle}`

Bulk JSON export and import for the admin **«دسته‌بندی‌ها»** page (`CategoriesPage`),
backed by `App\Category\Controller\CategoryImportController`.

`{bundle}` is one of: `provinces`, `cities`, `specialties`, `doctor_services`,
`insurances`, `tags`.

**Permission:** `ROLE_ADMIN` (whole controller).

Each bundle maps to one table:

| bundle | table | has `slug` | has `weight` | references |
|---|---|:---:|:---:|---|
| `provinces` | `provinces` | ❌ | ✅ | — |
| `cities` | `cities` | ❌ | ✅ | `province_id` → provinces, `representation_id` → representations |
| `specialties` | `specialties` | ✅ | ✅ | `parent_id` → specialties (same file) |
| `doctor_services` | `doctor_services` | ✅ | ✅ | `specialty_id` → specialties |
| `insurances` | `insurances` | ❌ | ❌ | `type` ∈ {`basic`,`supplementary`} |
| `tags` | `tags` | ✅ | ❌ | — |

---

## GET `/api/v1/admin/categories/{bundle}/export`

Complete export of a bundle — **every** row, no pagination cap. Returns the raw
table columns (the same keys an import expects), so export → import is lossless.

> ⚠️ Do **not** use the paginated admin list endpoints (e.g. `/api/v1/admin/specialties?limit=9999`)
> as an export source for import: those cap at 100 rows, and feeding a truncated
> file into the wipe+replace import below deletes every row beyond the first 100.

### Response `200`
```json
{
  "success": true,
  "data": [
    { "id": 1, "uuid": "…", "name": "قلب", "slug": "ghalb", "status": 1, "weight": 0, "parent_id": null }
  ]
}
```

For `cities` the export rows carry every column plus `province_name`, and
`social_media` is returned as a JSON **object** — the exact shape
`nobat724_front/data/city.json` uses, so the export file can be dropped into the
frontend (and re-imported) without any transformation.

### Errors
| Code | HTTP | Description |
|------|------|-------------|
| `ERR_AUTH_001` | 401 | Missing token |
| `ERR_AUTH_006` | 403 | Not admin |
| `ERR_BUNDLE_UNKNOWN` | 404 | Unknown bundle |

---

## POST `/api/v1/admin/categories/{bundle}/import`

**Wipe + full replace** of the bundle's table from the uploaded JSON, inside a
single transaction (FK checks disabled during the swap). Raw `id` values from the
file are preserved so relations stay intact.

Validation is **strict and two-phase**: the whole file is validated first; if
**any** row is invalid, nothing is written and every error is returned. Unknown
extra keys (e.g. `province_name` produced by export) are ignored.

### Request Body (`application/json`)
A non-empty array of records, or `{ "items": [ … ] }`. Max **5000** rows.

```json
[
  { "id": 1, "name": "قلب",  "slug": "ghalb", "status": 1, "weight": 0, "parent_id": null },
  { "id": 2, "name": "آریتمی", "slug": "arr", "status": 1, "weight": 1, "parent_id": 1 }
]
```

### Per-row validation rules
| Field | Rule |
|---|---|
| `id` | required, positive integer, unique within the file |
| `name` | required, non-empty string ≤ 255 |
| `status` | `0` or `1` (default `1`) |
| `weight` | integer ≥ 0 — bundles with a weight column only |
| `slug` | required, ≤ 255, unique within file — `specialties`/`doctor_services`/`tags` |
| `type` | `basic` or `supplementary` — `insurances` only |
| `uuid` | optional; generated if missing |
| `parent_id` | `specialties`: null, or an `id` present **in the same file** |
| `specialty_id` | `doctor_services`: null, or an existing `specialties.id` |
| `province_id` / `representation_id` | `cities`: null, or an existing id in the referenced table |
| `title`, `site_name`, `slogan`, `domain`, `keywords`, `description`, `footer_description`, `contact_phone`, `email`, `logo_url` | `cities`: optional SEO/branding strings (`title` = meta title of the city website) |
| `social_media` | `cities`: optional; JSON object (e.g. `{"instagram": "…"}`) or its string form — stored as JSON |

### Response `200`
```json
{ "success": true, "data": { "imported": 128 } }
```

### Response `422` (validation failed — nothing written)
```json
{
  "success": false,
  "data": null,
  "errors": [
    { "row": 2, "field": "parent_id", "message": "ردیف 2: والد با شناسه 999 در همین فایل وجود ندارد" }
  ]
}
```

### Errors
| Code | HTTP | Description |
|------|------|-------------|
| `ERR_AUTH_001` | 401 | Missing token |
| `ERR_AUTH_006` | 403 | Not admin |
| `ERR_BUNDLE_UNKNOWN` | 404 | Unknown bundle |
| `ERR_IMPORT_FORMAT` | 422 | Body is not a non-empty array of records |
| `ERR_IMPORT_TOO_LARGE` | 422 | More than 5000 rows |
| (field errors) | 422 | One entry per failed row/field; no rows written |
| `ERR_IMPORT_FAILED` | 500 | DB error during replace (transaction rolled back) |

---

## Seeding a from-scratch database

Canonical category data ships in the repo at `data/seed/*.json`
(`provinces.json`, `cities.json`, `specialties.json`, `doctor_services.json`).
After creating an empty schema, populate it with:

```bash
ddev exec php bin/console app:seed-categories
```

The command (`App\Category\Command\SeedCategoriesCommand`) reads those files and
runs the same strict validation + wipe-and-replace as the HTTP import, in
dependency order (provinces → cities, specialties → doctor_services). The shared
logic lives in `App\Category\Service\CategoryImporter`.

Full rebuild:
```bash
ddev exec php bin/console doctrine:database:drop --force
ddev exec php bin/console doctrine:database:create
ddev exec php bin/console doctrine:migrations:migrate --no-interaction
ddev exec php bin/console app:seed-categories
ddev exec php bin/console app:create-admin 09100000001 'Test@1234'
```