Paywall detection: swap to Wayback snapshot when article body is gated

Fetch each new article once and measure visible <p> text. If body is below
500 chars (or fetch fails), resolve a Wayback snapshot via the
archive.org/wayback/available API and use that URL for both the og:image
fallback and the posted link. Dedup keys stay derived from the original
URL so paywalled/non-paywalled hits collide as before.

- New: internal/ingestion/article.go (FetchArticleMeta via goquery)
- New: internal/ingestion/wayback.go (ResolveWayback)
- Removed: internal/ingestion/og.go (folded into article.go)
- poller.go: dedup first, then one article fetch, then snapshot fallback

internal/ingestion/poller.go

@@ -106,17 +106,10 @@ func (p *Poller) pollOnceWithErr(src config.SourceConfig) error {
 			continue
 		}
 
-		// Last-resort image fallback: feed gave us nothing, scrape og:image
-		// from the article page. Only runs for genuinely new items.
-		if items[i].ImageURL == "" {
-			if og := FetchOGImage(items[i].ArticleURL); og != "" {
-				items[i].ImageURL = og
-				slog.Debug("og:image fallback used",
-					"guid", items[i].GUID, "url", items[i].ArticleURL, "image", og)
-			}
-		}
-
-		canonical := dedup.CanonicalURL(items[i].ArticleURL)
+		// Dedup checks first — canonical and headline use the ORIGINAL URL so
+		// they stay stable whether we end up swapping to a Wayback snapshot.
+		originalURL := items[i].ArticleURL
+		canonical := dedup.CanonicalURL(originalURL)
 		headlineNorm := dedup.NormalizeHeadline(items[i].Headline)
 
 		if storage.IsCanonicalSeen(canonical) {
@@ -130,6 +123,38 @@ func (p *Poller) pollOnceWithErr(src config.SourceConfig) error {
 			continue
 		}
 
+		// Genuinely new story — fetch the article page once. We use it for:
+		//   1. og:image fallback when the feed didn't give us one.
+		//   2. Paywall detection: visible body text below the threshold means
+		//      the article is gated, so swap in a Wayback snapshot for both
+		//      Pete (image) and the reader (posted link).
+		meta := FetchArticleMeta(originalURL)
+		paywalled := !meta.Fetched || meta.BodyChars < PaywallBodyThreshold
+		if paywalled {
+			if snap := ResolveWayback(originalURL); snap != "" {
+				snapMeta := FetchArticleMeta(snap)
+				if snapMeta.Fetched {
+					items[i].ArticleURL = snap
+					if items[i].ImageURL == "" && snapMeta.ImageURL != "" {
+						items[i].ImageURL = snapMeta.ImageURL
+					}
+					slog.Info("paywall detected, using wayback snapshot",
+						"guid", items[i].GUID, "original", originalURL,
+						"snapshot", snap, "body_chars", meta.BodyChars)
+				} else {
+					slog.Debug("paywall detected but snapshot fetch failed",
+						"guid", items[i].GUID, "original", originalURL, "snapshot", snap)
+				}
+			} else {
+				slog.Debug("paywall detected but no wayback snapshot available",
+					"guid", items[i].GUID, "url", originalURL, "body_chars", meta.BodyChars)
+			}
+		} else if items[i].ImageURL == "" && meta.ImageURL != "" {
+			items[i].ImageURL = meta.ImageURL
+			slog.Debug("og:image fallback used",
+				"guid", items[i].GUID, "url", originalURL, "image", meta.ImageURL)
+		}
+
 		// Stamp source metadata onto the item
 		items[i].Source = src.Name
 		items[i].FeedHint = src.FeedHint