From 71bd40a2e8174853b80dddbc35ad38052e4477a9 Mon Sep 17 00:00:00 2001
From: arabcoders <admin@arabcoders.org>
Date: Wed, 27 Aug 2025 16:28:19 +0300
Subject: [PATCH] added new endpoint to get urls archive ids

---
 API.md                   | 55 ++++++++++++++++++++++++++++++++++++++++
 app/library/Utils.py     | 24 +++++++++++-------
 app/routes/api/yt_dlp.py | 37 ++++++++++++++++++++++++---
 3 files changed, 104 insertions(+), 12 deletions(-)

diff --git a/API.md b/API.md
index 84588a36..8d647c6c 100644
--- a/API.md
+++ b/API.md
@@ -44,7 +44,9 @@ This document describes the available endpoints and their usage. All endpoints r
     - [GET /api/logs](#get-apilogs)
     - [GET /api/notifications](#get-apinotifications)
     - [PUT /api/notifications](#put-apinotifications)
+    - [POST /api/yt-dlp/archive\_id/](#post-apiyt-dlparchive_id)
     - [POST /api/notifications/test](#post-apinotificationstest)
+    - [GET /api/yt-dlp/options](#get-apiyt-dlpoptions)
   - [Error Responses](#error-responses)
 
 ---
@@ -824,6 +826,41 @@ Binary image data with appropriate `Content-Type` header.
   "allowedTypes": ["added", "completed", "error", "cancelled", "cleared", "log_info", "log_success", ...]
 }
 ```
+---
+
+### POST /api/yt-dlp/archive_id/
+**Purpose**: Get the archive ID for a given URLs.
+**Body**: Array of URLs.
+```json
+[
+    "https://youtube.com/...",
+    "https://..."
+}
+```
+
+**Response**:
+```json
+[
+  {
+    "index": "index_of_the_url_in_request_array",
+    "url": "the_url",
+    "id": "the_video_id_or_null_if_not_found",
+    "ie_key": "the_extractor_key_or_null_if_not_found",
+    "archive_id": "the_archive_id_or_null_if_not_found",
+    "error": "error_message_if_any_or_null"
+  },
+  ...
+]
+```
+
+or an error:
+```json
+{
+  "error": "text"
+}
+```
+
+- If the body is not a valid JSON array, returns `400 Bad Request`.
 
 ---
 
@@ -840,6 +877,24 @@ Binary image data with appropriate `Content-Type` header.
 
 ---
 
+### GET /api/yt-dlp/options
+**Purpose**: Get the current yt-dlp CLI options as a JSON object.
+
+**Response**:
+```json
+[
+  {
+    "description": "Description of the option",
+    "flags":[ "--option", "-o" ],
+    "group": "Option Group",
+    "ignored": false, // true if this option is ignored by ytptube.
+  },
+  ...
+]
+```
+
+---
+
 ## Error Responses
 
 Most endpoints return standard error codes (`400`, `403`, `404`, `500`, etc.) and a JSON body on failure. For example:
diff --git a/app/library/Utils.py b/app/library/Utils.py
index b1f74ad8..4a54145f 100644
--- a/app/library/Utils.py
+++ b/app/library/Utils.py
@@ -405,13 +405,9 @@ def check_id(file: Path) -> bool | str:
 
 @lru_cache(maxsize=512)
 def is_private_address(hostname: str) -> bool:
-    try:
-        ip = socket.gethostbyname(hostname)
-        ip_obj = ipaddress.ip_address(ip)
-        return ip_obj.is_private or ip_obj.is_loopback or ip_obj.is_reserved or ip_obj.is_link_local
-    except socket.gaierror:
-        # Could not resolve - treat as invalid or restricted
-        return True
+    ip = socket.gethostbyname(hostname)
+    ip_obj = ipaddress.ip_address(ip)
+    return ip_obj.is_private or ip_obj.is_loopback or ip_obj.is_reserved or ip_obj.is_link_local
 
 
 def validate_url(url: str, allow_internal: bool = False) -> bool:
@@ -447,10 +443,20 @@ def validate_url(url: str, allow_internal: bool = False) -> bool:
         raise ValueError(msg)
 
     hostname: str | None = parsed_url.host
-    if allow_internal is False and (not hostname or is_private_address(hostname)):
-        msg = "Access to internal urls or private networks is not allowed."
+    if not hostname:
+        msg = "Invalid hostname."
         raise ValueError(msg)
 
+    if allow_internal is False:
+        try:
+            if is_private_address(hostname):
+                msg = "Access to internal urls or private networks is not allowed."
+                raise ValueError(msg)
+        except socket.gaierror as e:
+            LOG.error(f"Error resolving hostname '{hostname}': {e!s}")
+            msg = "Invalid hostname."
+            raise ValueError(msg) from e
+
     return True
 
 
diff --git a/app/routes/api/yt_dlp.py b/app/routes/api/yt_dlp.py
index 49291c20..5ca2e459 100644
--- a/app/routes/api/yt_dlp.py
+++ b/app/routes/api/yt_dlp.py
@@ -309,6 +309,37 @@ async def get_options() -> Response:
     """
     from app.library.ytdlp import ytdlp_options
 
-    return web.json_response(
-        body=json.dumps(ytdlp_options(), indent=4, default=str), status=web.HTTPOk.status_code
-    )
+    return web.json_response(body=json.dumps(ytdlp_options(), indent=4, default=str), status=web.HTTPOk.status_code)
+
+
+@route("POST", "api/yt-dlp/archive_id/", "get_archive_ids")
+async def get_archive_ids(request: Request) -> Response:
+    """
+    Get the yt-dlp CLI options.
+
+    Returns:
+        Response: The response object with the yt-dlp CLI options.
+
+    """
+    from app.library.Utils import get_archive_id
+
+    data = (await request.json()) if request.body_exists else None
+    if not data or not isinstance(data, list):
+        return web.json_response(
+            data={"error": "Invalid request. expecting list with URLs."},
+            status=web.HTTPBadRequest.status_code,
+        )
+
+    response = []
+
+    for i, url in enumerate(data):
+        dct = {"index": i, "url": url}
+        try:
+            validate_url(url)
+            dct.update(get_archive_id(url))
+        except ValueError as e:
+            dct.update({"id": None, "ie_key": None, "archive_id": None, "error": str(e)})
+
+        response.append(dct)
+
+    return web.json_response(data=response, status=web.HTTPOk.status_code)