Wiki Jokes: daily LLM jokes from Wikipedia featured article

- FastAPI + APScheduler + SQLite, single container - Daily generation at 06:00 Europe/Copenhagen + cold-start generation - Retry with backoff (3 attempts/15 min), stale fallback with banner - OpenAI-compatible endpoint via env (OPENAI_BASE_URL/MODEL/API_KEY) - docker-compose with named volume for joke persistence
2026-09-29 19:38:01 +00:00
commit 075b72ec6c
14 changed files with 601 additions and 0 deletions
@@ -0,0 +1,79 @@
+"""Fetch Wikipedia's 'Today's Featured Article' via the REST API feed."""
+import re
+from dataclasses import dataclass
+from datetime import date
+
+import httpx
+
+from .config import settings
+
+FEED_URL = (
+    "https://{lang}.wikipedia.org/api/rest_v1/feed/featured/{year:04d}/{month:02d}/{day:02d}"
+)
+
+
+@dataclass
+class FeaturedArticle:
+    title: str
+    url: str
+    extract: str  # plain-text extract, HTML stripped
+
+
+def _strip_html(html: str) -> str:
+    text = re.sub(r"<[^>]+>", " ", html)
+    text = (
+        text.replace("&amp;", "&")
+        .replace("&quot;", '"')
+        .replace("&#39;", "'")
+        .replace("&lt;", "<")
+        .replace("&gt;", ">")
+        .replace("&nbsp;", " ")
+    )
+    return re.sub(r"\s+", " ", text).strip()
+
+
+def get_featured_article(day: date) -> FeaturedArticle:
+    url = FEED_URL.format(
+        lang=settings.wikipedia_lang, year=day.year, month=day.month, day=day.day
+    )
+    resp = httpx.get(
+        url,
+        headers={
+            "User-Agent": "wiki-jokes/1.0 (daily joke generator; contact: admin@example.com)",
+            "Accept": "application/json",
+        },
+        timeout=30,
+        follow_redirects=True,
+    )
+    resp.raise_for_status()
+    data = resp.json()
+
+    tfa = data.get("tfa")
+    if not tfa:
+        raise ValueError(f"No featured article (tfa) in feed for {day.isoformat()}")
+
+    # Prefer the REST "extracts.plain.text" if present, else the summary HTML.
+    extract = ""
+    extracts = tfa.get("extracts") or {}
+    plain = extracts.get("plain") or {}
+    if plain.get("text"):
+        extract = plain["text"]
+    elif tfa.get("extract"):
+        extract = tfa["extract"]
+    elif tfa.get("summary"):
+        extract = _strip_html(tfa["summary"])
+
+    extract = _strip_html(extract) if "<" in extract else extract
+    if not extract:
+        raise ValueError(f"Featured article for {day.isoformat()} has no extract text")
+
+    # Cap the extract so the prompt stays bounded (~4000 chars).
+    extract = extract[:4000]
+
+    display = (tfa.get("displaytitle") or tfa.get("title", "Unknown article"))
+    display = _strip_html(display).replace("_", " ")
+    return FeaturedArticle(
+        title=display,
+        url=tfa.get("content_urls", {}).get("desktop", {}).get("page", url),
+        extract=extract,
+    )