merge #4832: [tmohentai] add 'gallery' extractor (#4808)

docs/supportedsites.md

@@ -823,6 +823,12 @@ Consider all sites to be NSFW unless otherwise known.
     <td>Galleries</td>
     <td></td>
 </tr>
+<tr>
+    <td>TMOHentai</td>
+    <td>https://tmohentai.com/</td>
+    <td>Galleries</td>
+    <td></td>
+</tr>
 <tr>
     <td>Toyhouse</td>
     <td>https://toyhou.se/</td>

gallery_dl/extractor/__init__.py

@@ -147,6 +147,7 @@ modules = [
     "tapas",
     "tcbscans",
     "telegraph",
+    "tmohentai",
     "toyhouse",
     "tsumino",
     "tumblr",

gallery_dl/extractor/tmohentai.py

@@ -0,0 +1,48 @@
+# -*- 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.
+
+"""Extractors for https://tmohentai.com/"""
+
+from .common import GalleryExtractor
+from .. import text
+
+BASE_PATTERN = r"(?:https?://)?tmohentai\.com"
+
+
+class TmohentaiGalleryExtractor(GalleryExtractor):
+    category = "tmohentai"
+    root = "http://tmohentai.com"
+    directory_fmt = ("{category}", "{title} ({gallery_id})")
+    pattern = BASE_PATTERN + r"/(?:contents|reader)/(\w+)"
+    example = "https://tmohentai.com/contents/12345a67b89c0"
+
+    def __init__(self, match):
+        self.gallery_id = match.group(1)
+        url = "{}/contents/{}".format(self.root, self.gallery_id)
+        GalleryExtractor.__init__(self, match, url)
+
+    def images(self, page):
+        fmt = "https://imgrojo.tmohentai.com/contents/{}/{{:>03}}.webp".format(
+            self.gallery_id).format
+        cnt = page.count('class="lanzador')
+        return [(fmt(i), None) for i in range(0, cnt)]
+
+    def metadata(self, page):
+        extr = text.extract_from(page)
+
+        return {
+            "gallery_id": self.gallery_id,
+            "title"     : text.unescape(extr("<h3>", "<").strip()),
+            "artists"   : text.split_html(extr(
+                "<label>Artists and Artists Groups</label>", "</ul>")),
+            "categories": text.split_html(extr(
+                "<label>Genders</label>", "</ul>")),
+            "tags"      : text.split_html(extr(
+                "<label>Tags</label>", "</ul>")),
+            "uploader"  : text.remove_html(extr(
+                "<label>Uploaded By</label>", "</ul>")),
+            "language"  : extr("&nbsp;", "\n"),
+        }

scripts/supportedsites.py

@@ -122,6 +122,7 @@ CATEGORY_MAP = {
     "tbib"           : "The Big ImageBoard",
     "tcbscans"       : "TCB Scans",
     "tco"            : "Twitter t.co",
+    "tmohentai"      : "TMOHentai",
     "thatpervert"    : "ThatPervert",
     "thebarchive"    : "The /b/ Archive",
     "thecollection"  : "The /co/llection",

test/results/tmohentai.py

@@ -0,0 +1,54 @@
+# -*- 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 tmohentai
+
+
+__tests__ = (
+{
+    "#url"     : "https://tmohentai.com/contents/653c2aeaa693c",
+    "#category": ("", "tmohentai", "gallery"),
+    "#class"   : tmohentai.TmohentaiGalleryExtractor,
+    "#pattern" : r"https://imgrojo\.tmohentai\.com/contents/653c2aeaa693c/\d\d\d\.webp",
+    "#count"   : 46,
+
+    "artists"   : ["Andoryu"],
+    "categories": [
+        "Big Breasts",
+        "BlowJob",
+        "Cheating",
+        "Mature",
+        "Milf",
+        "Student",
+    ],
+    "count"     : 46,
+    "extension" : "webp",
+    "gallery_id": "653c2aeaa693c",
+    "language"  : "Español",
+    "num"       : int,
+    "tags"      : [
+        "milf",
+        "Madre",
+        "enormes pechos",
+        "Peluda",
+        "nakadashi",
+        "cheating",
+        "madura",
+        "sexo a escondidas",
+        "Ama de casa",
+        "mamada",
+    ],
+    "title"     : "La Mama de mi Novia es tan Pervertida que no Pude Soportarlo mas",
+    "uploader"  : "NekoCreme Fansub",
+},
+
+{
+    "#url"     : "https://tmohentai.com/reader/653c2aeaa693c/paginated/1",
+    "#category": ("", "tmohentai", "gallery"),
+    "#class"   : tmohentai.TmohentaiGalleryExtractor,
+},
+
+)