fix: prefer first-post F95 covers and banners over metadata thumbnails

backend/app/f95.py

@@ -4,7 +4,7 @@ import urllib.error
 import urllib.request
 from html import unescape
 from html.parser import HTMLParser
-from urllib.parse import urljoin, urlparse
+from urllib.parse import unquote, urljoin, urlparse
 
 THREAD_RE = re.compile(r"^/threads/[^/]+\.\d+/?|^/threads/\d+/?")
 UA = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124 Safari/537.36"
@@ -132,6 +132,7 @@ class _Parser(HTMLParser):
         self.og_title = ""
         self.og_image = None
         self.post_image = None
+        self.cover_image = None
         self.tags: list[str] = []
         self.post: list[str] = []
         self.template = ""
@@ -164,10 +165,18 @@ class _Parser(HTMLParser):
             self._depth += 1
             if self._skip or {"bbCodeSpoiler", "messageHide"} & set(cls):
                 self._skip += 1
-        elif tag == "img" and "bbImage" in cls and not self._skip and not self.post_image:
+        elif tag == "img" and "bbImage" in cls and not self._skip:
             for key in ("data-src", "src"):
                 if candidate := image_url(a.get(key) or ""):
-                    self.post_image = candidate
+                    if not self.post_image:
+                        self.post_image = candidate
+                    # Prefer a cover/banner explicitly identified in the opening post.
+                    # Metadata thumbnails and earlier inline images may be unrelated.
+                    name = " ".join((a.get("alt") or "", a.get("title") or "",
+                                     unquote(urlparse(candidate).path.rsplit("/", 1)[-1])))
+                    name = re.sub(r"([a-z])([A-Z])", r"\1_\2", name)
+                    if not self.cover_image and re.search(r"(?:^|[^a-z])(?:f95[ _-]*)?(?:cover|banner)(?:[^a-z]|$)", name, re.I):
+                        self.cover_image = candidate
                     break
         elif tag == "br" and not self._skip:
             self.post.append("\n")
@@ -207,7 +216,7 @@ def parse(html: str) -> dict:
     title = re.sub(r"\s*\[[^\]]*\]", "", title)
     title = re.sub(r"^(?:[A-Za-z0-9'’ ]{1,20} - )+", "", title).strip()
     return {"title": title, "description": _description("".join(p.post)),
-            "tags": sorted(set(p.tags)), "image_url": p.og_image or p.post_image,
+            "tags": sorted(set(p.tags)), "image_url": p.cover_image or p.post_image or p.og_image,
             "logged_in": p.logged_in}
 
 

backend/tests/test_f95.py

@@ -29,11 +29,41 @@ def page(meta="", post="", reply=""):
 
 
 def test_og_preview_and_existing_metadata():
-    result = f95.parse(page(f'<meta property="og:image" content="{IMAGE}">',
-                            '<img class="bbImage" src="/attachments/other.jpg">'))
+    result = f95.parse(page(f'<meta property="og:image" content="{IMAGE}">'))
     assert result == {"title": "Example", "description": "A story", "tags": [], "image_url": IMAGE, "logged_in": None}
 
 
+def test_mist_banner_beats_metadata_thumbnail_and_earlier_inline_image():
+    banner = ("https://attachments.f95zone.to/2022/03/1736427_F95Banner_Sideversion.png"
+              "?h=test-signature&e=1793113200&u=42")
+    escaped = banner.replace("&", "&amp;")
+    html = page('<meta property="og:image" content="/attachments/thumbnail.jpg">',
+                '<img class="bbImage" src="/attachments/screenshot.jpg">'
+                f'<img src="{escaped}" data-src="{escaped}" data-url="" '
+                'class="bbImage lazyloaded" data-zoom-target="1" alt="F95 Cover.png" '
+                'title="F95 Cover.png" width="1200" height="500" loading="lazy">',
+                '<img class="bbImage" src="/attachments/reply-cover.png">')
+    assert f95.parse(html)["image_url"] == banner
+
+
+@pytest.mark.parametrize("attributes", [
+    'alt="F95 Cover.png"', 'title="Game Cover"',
+    'data-src="/attachments/123_F95Banner_Sideversion.png"',
+])
+def test_cover_identified_by_alt_title_or_filename(attributes):
+    html = page(post='<img class="bbImage" src="/first.jpg">'
+                f'<img class="bbImage" src="/chosen.jpg" {attributes}>')
+    expected = "/attachments/123_F95Banner_Sideversion.png" if "data-src" in attributes else "/chosen.jpg"
+    assert f95.parse(html)["image_url"] == "https://f95zone.to" + expected
+
+
+def test_first_post_image_preferred_over_og_thumbnail_without_cover_marker():
+    html = page('<meta property="og:image" content="/thumbnail.jpg">',
+                '<div class="bbCodeSpoiler"><img class="bbImage" src="/hidden-cover.png"></div>'
+                '<img class="bbImage" src="/first.jpg"><img class="bbImage" src="/second.jpg">')
+    assert f95.parse(html)["image_url"] == "https://f95zone.to/first.jpg"
+
+
 def test_first_post_lazy_image_fallback_ignores_logo_spoilers_and_replies():
     html = page('<meta property="og:image" content="/styles/logo.png">',
                 '<img class="avatar" src="/avatar.jpg">'