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}