[hdoujin] add support (#6810)

docs/configuration.rst

@@ -3092,6 +3092,70 @@ Description
     Recursively download files from subfolders.
 
 
+extractor.hdoujin.crt
+---------------------
+Type
+    ``string``
+Example
+    * ``"0542daa9-352c-4fd5-a497-6c6d5cf07423"``
+    * ``"/12345/a1b2c3d4e5f6?crt=0542daa9-352c-4fd5-a497-6c6d5cf07423"``
+Description
+    The ``crt`` query parameter value
+    sent when fetching gallery data.
+
+    To get this value:
+
+    * Open your browser's Developer Tools (F12)
+    * Select `Network` -> `XHR`
+    * Open a gallery page
+    * Select the last `Network` entry and copy its ``crt`` value
+
+    Note: You will also need your browser's
+    `user-agent <extractor.*.user-agent_>`__
+
+
+extractor.hdoujin.format
+------------------------
+Type
+    * ``string``
+    * ``list`` of ``strings``
+Default
+    ``["0", "1600", "1280", "980", "780"]``
+Description
+    Name(s) of the image format to download.
+
+    When more than one format is given, the first available one is selected.
+
+    | Possible formats are
+    | ``"780"``, ``"980"``, ``"1280"``, ``"1600"``, ``"0"`` (original)
+
+
+extractor.hdoujin.tags
+----------------------
+Type
+    ``bool``
+Default
+    ``false``
+Description
+    Group ``tags`` by type and
+    provide them as ``tags_<type>`` metadata fields,
+    for example ``tags_artist`` or ``tags_character``.
+
+
+extractor.hdoujin.token
+-----------------------
+Type
+    ``string``
+Example
+    * ``"3f1a9b72-4e4d-4f4e-9e5d-4a2b99f7c893"``
+    * ``"Bearer 3f1a9b72-4e4d-4f4e-9e5d-4a2b99f7c893"``
+    * ``"Authorization: Bearer 3f1a9b72-4e4d-4f4e-9e5d-4a2b99f7c893"``
+Description
+    ``Authorization`` header value
+    used for requests to ``https://api.hdoujin.org``
+    to access ``favorite`` galleries.
+
+
 extractor.hentaifoundry.descriptions
 ------------------------------------
 Type
@@ -3586,70 +3650,6 @@ Description
     the first in the list gets chosen (usually `mp3`).
 
 
-extractor.schalenetwork.crt
----------------------------
-Type
-    ``string``
-Example
-    * ``"0542daa9-352c-4fd5-a497-6c6d5cf07423"``
-    * ``"/12345/a1b2c3d4e5f6?crt=0542daa9-352c-4fd5-a497-6c6d5cf07423"``
-Description
-    The ``crt`` query parameter value
-    sent when fetching gallery data.
-
-    To get this value:
-
-    * Open your browser's Developer Tools (F12)
-    * Select `Network` -> `XHR`
-    * Open a gallery page
-    * Select the last `Network` entry and copy its ``crt`` value
-
-    Note: You will also need your browser's
-    `user-agent <extractor.*.user-agent_>`__
-
-
-extractor.schalenetwork.format
-------------------------------
-Type
-    * ``string``
-    * ``list`` of ``strings``
-Default
-    ``["0", "1600", "1280", "980", "780"]``
-Description
-    Name(s) of the image format to download.
-
-    When more than one format is given, the first available one is selected.
-
-    | Possible formats are
-    | ``"780"``, ``"980"``, ``"1280"``, ``"1600"``, ``"0"`` (original)
-
-
-extractor.schalenetwork.tags
-----------------------------
-Type
-    ``bool``
-Default
-    ``false``
-Description
-    Group ``tags`` by type and
-    provide them as ``tags_<type>`` metadata fields,
-    for example ``tags_artist`` or ``tags_character``.
-
-
-extractor.schalenetwork.token
------------------------------
-Type
-    ``string``
-Example
-    * ``"3f1a9b72-4e4d-4f4e-9e5d-4a2b99f7c893"``
-    * ``"Bearer 3f1a9b72-4e4d-4f4e-9e5d-4a2b99f7c893"``
-    * ``"Authorization: Bearer 3f1a9b72-4e4d-4f4e-9e5d-4a2b99f7c893"``
-Description
-    ``Authorization`` header value
-    used for requests to ``https://api.schale.network``
-    to access ``favorite`` galleries.
-
-
 extractor.lolisafe.domain
 -------------------------
 Type
@@ -4880,6 +4880,70 @@ Description
     Download videos.
 
 
+extractor.schalenetwork.crt
+---------------------------
+Type
+    ``string``
+Example
+    * ``"0542daa9-352c-4fd5-a497-6c6d5cf07423"``
+    * ``"/12345/a1b2c3d4e5f6?crt=0542daa9-352c-4fd5-a497-6c6d5cf07423"``
+Description
+    The ``crt`` query parameter value
+    sent when fetching gallery data.
+
+    To get this value:
+
+    * Open your browser's Developer Tools (F12)
+    * Select `Network` -> `XHR`
+    * Open a gallery page
+    * Select the last `Network` entry and copy its ``crt`` value
+
+    Note: You will also need your browser's
+    `user-agent <extractor.*.user-agent_>`__
+
+
+extractor.schalenetwork.format
+------------------------------
+Type
+    * ``string``
+    * ``list`` of ``strings``
+Default
+    ``["0", "1600", "1280", "980", "780"]``
+Description
+    Name(s) of the image format to download.
+
+    When more than one format is given, the first available one is selected.
+
+    | Possible formats are
+    | ``"780"``, ``"980"``, ``"1280"``, ``"1600"``, ``"0"`` (original)
+
+
+extractor.schalenetwork.tags
+----------------------------
+Type
+    ``bool``
+Default
+    ``false``
+Description
+    Group ``tags`` by type and
+    provide them as ``tags_<type>`` metadata fields,
+    for example ``tags_artist`` or ``tags_character``.
+
+
+extractor.schalenetwork.token
+-----------------------------
+Type
+    ``string``
+Example
+    * ``"3f1a9b72-4e4d-4f4e-9e5d-4a2b99f7c893"``
+    * ``"Bearer 3f1a9b72-4e4d-4f4e-9e5d-4a2b99f7c893"``
+    * ``"Authorization: Bearer 3f1a9b72-4e4d-4f4e-9e5d-4a2b99f7c893"``
+Description
+    ``Authorization`` header value
+    used for requests to ``https://api.schale.network``
+    to access ``favorite`` galleries.
+
+
 extractor.sexcom.gifs
 ---------------------
 Type

docs/gallery-dl.conf

@@ -386,6 +386,15 @@
             "website-token": null,
             "recursive": false
         },
+        "hdoujin":
+        {
+            "crt"  : "",
+            "token": "",
+            "sleep-request": "0.5-1.5",
+
+            "format": ["0", "1600", "1280", "980", "780"],
+            "tags"  : false
+        },
         "hentaifoundry":
         {
             "descriptions": "text",

docs/supportedsites.md

@@ -373,6 +373,12 @@ Consider all listed sites to potentially be NSFW.
     <td>Archive, Individual Posts, Home Feed, Search Results</td>
     <td></td>
 </tr>
+<tr id="hdoujin" title="hdoujin">
+    <td>HDoujin Galleries</td>
+    <td>https://hdoujin.org/</td>
+    <td>Favorites, Galleries, Search Results</td>
+    <td></td>
+</tr>
 <tr id="hentaifoundry" title="hentaifoundry">
     <td>Hentai Foundry</td>
     <td>https://www.hentai-foundry.com/</td>

gallery_dl/extractor/__init__.py

@@ -73,6 +73,7 @@ modules = [
     "girlswithmuscle",
     "gofile",
     "hatenablog",
+    "hdoujin",
     "hentai2read",
     "hentaicosplays",
     "hentaifoundry",

gallery_dl/extractor/hdoujin.py

@@ -0,0 +1,42 @@
+# -*- coding: utf-8 -*-
+
+# Copyright 2025 Mike Fährmann
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License version 2 as
+# published by the Free Software Foundation.
+
+"""Extractors for https://hdoujin.org/"""
+
+from . import schalenetwork
+
+BASE_PATTERN = r"(?:https?://)?(?:www\.)?(hdoujin\.(?:org|net))"
+
+
+class HdoujinBase():
+    """Base class for hdoujin extractors"""
+    category = "hdoujin"
+    root = "https://hdoujin.org"
+    root_api = "https://api.hdoujin.org"
+    root_auth = "https://auth.hdoujin.org"
+
+
+class HdoujinGalleryExtractor(
+        HdoujinBase, schalenetwork.SchalenetworkGalleryExtractor):
+    pattern = rf"{BASE_PATTERN}/(?:g|reader)/(\d+)/(\w+)"
+    example = "https://hdoujin.org/g/12345/67890abcdef/"
+
+
+class HdoujinSearchExtractor(
+        HdoujinBase, schalenetwork.SchalenetworkSearchExtractor):
+    pattern = rf"{BASE_PATTERN}/(?:tag/([^/?#]+)|browse)?(?:/?\?([^#]*))?$"
+    example = "https://hdoujin.org/browse?s=QUERY"
+
+
+class HdoujinFavoriteExtractor(
+        HdoujinBase, schalenetwork.SchalenetworkFavoriteExtractor):
+    pattern = rf"{BASE_PATTERN}/favorites(?:\?([^#]*))?"
+    example = "https://hdoujin.org/favorites"
+
+
+HdoujinBase.extr_class = HdoujinGalleryExtractor

gallery_dl/extractor/schalenetwork.py

@@ -27,6 +27,7 @@ class SchalenetworkExtractor(Extractor):
     root = "https://niyaniya.moe"
     root_api = "https://api.schale.network"
     root_auth = "https://auth.schale.network"
+    extr_class = None
     request_interval = (0.5, 1.5)
 
     def _init(self):
@@ -38,6 +39,7 @@ class SchalenetworkExtractor(Extractor):
 
     def _pagination(self, endpoint, params):
         url_api = self.root_api + endpoint
+        cls = self.extr_class
 
         while True:
             data = self.request_json(
@@ -50,7 +52,7 @@ class SchalenetworkExtractor(Extractor):
 
             for entry in entries:
                 url = f"{self.root}/g/{entry['id']}/{entry['key']}"
-                entry["_extractor"] = SchalenetworkGalleryExtractor
+                entry["_extractor"] = cls
                 yield Message.Queue, url, entry
 
             try:
@@ -236,3 +238,6 @@ class SchalenetworkFavoriteExtractor(SchalenetworkExtractor):
         params["page"] = text.parse_int(params.get("page"), 1)
         self.headers["Authorization"] = self._token()
         return self._pagination(f"/books/favorites?crt={self._crt()}", params)
+
+
+SchalenetworkExtractor.extr_class = SchalenetworkGalleryExtractor

scripts/supportedsites.py

@@ -63,6 +63,7 @@ CATEGORY_MAP = {
     "girlswithmuscle": "Girls with Muscle",
     "hatenablog"     : "HatenaBlog",
     "hbrowse"        : "HBrowse",
+    "hdoujin"        : "HDoujin Galleries",
     "hentai2read"    : "Hentai2Read",
     "hentaicosplay"  : "Hentai Cosplay",
     "hentaienvy"     : "HentaiEnvy",

test/results/hdoujin.py

@@ -0,0 +1,40 @@
+# -*- coding: utf-8 -*-
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License version 2 as
+# published by the Free Software Foundation.
+
+from gallery_dl.extractor import hdoujin
+
+
+__tests__ = (
+{
+    "#url"     : "https://hdoujin.org/g/119874/bd0a5217dfc6",
+    "#class"   : hdoujin.HdoujinGalleryExtractor,
+},
+
+{
+    "#url"     : "https://hdoujin.net/g/119874/bd0a5217dfc6",
+    "#class"   : hdoujin.HdoujinGalleryExtractor,
+},
+
+{
+    "#url"     : "https://hdoujin.org/browse?s=beach",
+    "#class"   : hdoujin.HdoujinSearchExtractor,
+},
+
+{
+    "#url"     : "https://hdoujin.org/tag/female:maid",
+    "#class"   : hdoujin.HdoujinSearchExtractor,
+    "#pattern" : hdoujin.HdoujinGalleryExtractor.pattern,
+    "#range"   : "1-80",
+    "#count"   : 80,
+},
+
+{
+    "#url"     : "https://hdoujin.org/favorites",
+    "#class"   : hdoujin.HdoujinFavoriteExtractor,
+    "#auth"    : True,
+},
+
+)