# Aether API V3 Frontend Integration Guide (Svelte/TypeScript)

This guide explains how to update or create frontend functions to interact with the new **Aether API V3 CRUD** endpoints. V3 introduces a nested URL structure, a powerful POST-based search, and improved performance.

---

## 1. Key Differences (V2 vs V3)

| Feature | CRUD V2 (Legacy) | CRUD V3 (Modern) |
| --- | --- | --- |
| **Base Prefix** | `/v2/crud` | `/v3/crud` |
| **List Suffix** | Uses `/list` (e.g., `/v2/crud/journal/list`) | **No suffix** (e.g., `/v3/crud/journal/`) |
| **Nested Path** | Not supported in URL path | **Supported** (e.g., `/v3/crud/journal/{id}/journal_entry/`) |
| **Order By** | Passed in **Headers** (`order_by_li`) | Passed as **Query Parameter** (`order_by_li`) |
| **Complex Search**| Limited to GET `jp` (~2KB limit) | **POST `/search`** (Unlimited size) |
| **Latency Simulation**| Not standardized | `X-Delay-ms` (Header) or `delay_ms` (Query) |

---

## 2. Implementing V3 CRUD Functions

### A. List Top-Level Objects (GET)
Use this for simple lists or filtering by a parent via query parameters.

```ts
export async function get_ae_obj_li_v3({
    api_cfg,
    obj_type,
    for_obj_type,
    for_obj_id,
    enabled = 'enabled',
    hidden = 'not_hidden',
    limit = 100,
    offset = 0,
    order_by_li = null, // e.g., {"created_on": "DESC"}
    delay_ms = 0
}) {
    // 1. Build V3 Endpoint (Note: No /list suffix)
    const endpoint = `/v3/crud/${obj_type}/`;

    // 2. Build Query Params
    const params: any = {
        enabled,
        hidden,
        limit,
        offset
    };

    if (for_obj_type) params.for_obj_type = for_obj_type;
    if (for_obj_id) params.for_obj_id = for_obj_id;
    if (order_by_li) params.order_by_li = JSON.stringify(order_by_li);
    if (delay_ms) params.delay_ms = delay_ms;

    return await get_object({ api_cfg, endpoint, params });
}
```

### B. Access Nested Objects (GET)
V3 allows you to enforce parent-child relationships directly in the URL.

```ts
// Example: Get all entries for a specific journal
// GET /v3/crud/journal/{journal_id}/journal_entry/
export async function get_nested_obj_li_v3({
    api_cfg,
    parent_type,
    parent_id,
    child_type,
    limit = 100
}) {
    const endpoint = `/v3/crud/${parent_type}/${parent_id}/${child_type}/`;
    return await get_object({ api_cfg, endpoint, params: { limit } });
}
```

### C. Advanced Search (POST)
Use the new `/search` endpoint for complex logic (AND/OR, LIKE, IN) that would exceed URL length limits.

```ts
export async function search_ae_obj_v3({
    api_cfg,
    obj_type,
    search_query, // Complex SearchQuery object
    order_by_li = null
}) {
    const endpoint = `/v3/crud/${obj_type}/search`;
    
    const params: any = {};
    if (order_by_li) params.order_by_li = JSON.stringify(order_by_li);

    // Note: search_query is sent in the BODY via POST
    return await post_object({ 
        api_cfg, 
        endpoint, 
        params, 
        data: search_query 
    });
}

/**
 * Example search_query usage:
 * {
 *   "and": [
 *     { "field": "enable", "op": "eq", "value": true },
 *     { 
 *       "or": [
 *         { "field": "name", "op": "like", "value": "%Meeting%" },
 *         { "field": "priority", "op": "gt", "value": 5 }
 *       ]
 *     }
 *   ]
 * }
 */
```

---

## 3. Best Practices for V3

1.  **Stop using Headers for Sorting**: In V2, we used `headers['order_by_li']`. In V3, always use `params['order_by_li']`.
2.  **Prefer `/search` for Filters**: If your query involves more than a simple `for_obj_id`, use the `POST .../search` endpoint. It is safer, more readable, and bypasses the 2083-character URL limit.
3.  **Non-Blocking UI**: Use the `delay_ms` parameter during development to test how your Svelte components handle loading states and skeleton screens.
4.  **Singular Nouns**: Always use singular names for `obj_type` (e.g., `journal`, not `journals`), following the backend's `obj_type_kv_li` registry.

---

## 4. Special Case: Account Context
V3 strictly enforces account context. Ensure your `api_cfg` logic correctly sets the `X-Account-ID` header. 

*   If a request doesn't need an account (e.g., public site data), set the `X-No-Account-ID: bypass` header.
*   If using a temporary token, use the `x_no_account_id_token` query parameter.