feat: add --max-age-days filter to prevent posting stale RSS entries Introduce an age-based filter

rss2bsky.py

@@ -401,6 +401,7 @@ def process_title(title: str) -> str:
     title_text = clean_whitespace(title_text)
     return title_text
 
+
 def build_post_text_variants(title_text: str, link: str, max_length: int = 300):
     title_text = clean_whitespace(title_text)
     link = canonicalize_url(link) or link or ""
@@ -414,13 +415,13 @@ def build_post_text_variants(title_text: str, link: str, max_length: int = 300):
             seen.add(cleaned)
             variants.append(cleaned)
 
-    # Variant 1: títol + link (si cap sencer)
+    # Variant 1: title + link (if fits whole)
     if title_text and link:
         full = f"{title_text}\n\n{link}"
         if len(full) <= max_length:
             add_variant(full)
 
-    # Variant 2: només títol complet (sense truncar)
+    # Variant 2: full title only (no truncation)
     if title_text:
         if len(title_text) <= max_length:
             add_variant(title_text)
@@ -428,7 +429,7 @@ def build_post_text_variants(title_text: str, link: str, max_length: int = 300):
             truncated = title_text[:max_length].rstrip(" .")
             add_variant(truncated)
 
-    # Variant 3: títol truncat + link (si el títol complet+link no hi cap)
+    # Variant 3: truncated title + link (when full title+link doesn't fit)
     if title_text and link:
         full = f"{title_text}\n\n{link}"
         if len(full) > max_length:
@@ -438,12 +439,13 @@ def build_post_text_variants(title_text: str, link: str, max_length: int = 300):
                 truncated_title = title_text[:available].rstrip(" .")
                 add_variant(f"{truncated_title}\n\n{link}")
 
-    # Variant 4: només link (si no hi ha títol)
+    # Variant 4: link only (when no title)
     if link and not title_text:
         add_variant(link)
 
     return variants
 
+
 def is_x_or_twitter_domain(url: str) -> bool:
     try:
         hostname = (urlparse(url).hostname or "").lower()
@@ -1076,7 +1078,7 @@ def build_candidates_from_feed(feed, max_length: int = 300) -> List[EntryCandida
 
 
 # ============================================================
-# Orchestration
+# Login
 # ============================================================
 def login_with_backoff(
     client: Client,
@@ -1110,7 +1112,6 @@ def login_with_backoff(
         except Exception as e:
             logging.exception("❌ Login exception")
 
-            # Rate-limited login: retry first, cooldown only if exhausted
             if is_rate_limited_error(e):
                 if attempt < max_attempts:
                     wait_seconds = get_rate_limit_wait_seconds(e, base_delay, cfg)
@@ -1125,19 +1126,16 @@ def login_with_backoff(
                 activate_post_creation_cooldown_from_error(e, cooldown_path, cfg)
                 return False
 
-            # Bad credentials: fail fast
             if is_auth_error(e):
                 logging.error("❌ Authentication failed (bad handle/password/app-password).")
                 return False
 
-            # Network/transient: bounded retry
             if attempt < max_attempts and (is_network_error(e) or is_timeout_error(e)):
                 delay = min(base_delay * attempt, max_delay) + random.uniform(0, jitter_max)
                 logging.warning(f"⏳ Transient login failure. Retrying in {delay:.1f}s...")
                 time.sleep(delay)
                 continue
 
-            # Other errors: bounded retry
             if attempt < max_attempts:
                 delay = min(base_delay * attempt, max_delay) + random.uniform(0, jitter_max)
                 logging.warning(f"⏳ Login retry in {delay:.1f}s...")
@@ -1148,6 +1146,10 @@ def login_with_backoff(
 
     return False
 
+
+# ============================================================
+# Orchestration
+# ============================================================
 def run_once(
     rss_feed: str,
     bsky_handle: str,
@@ -1158,13 +1160,26 @@ def run_once(
     state_path: str,
     cooldown_path: str,
     cfg: AppConfig,
-    max_posts: int = 5              # ← NEW PARAMETER
+    max_posts: int = 5,
+    max_age_days: int = 7,          # ← NEW: 0 = disabled
 ) -> RunResult:
     if not PIL_AVAILABLE:
         logging.warning("🟡 Pillow is not installed. External card thumbnail compression is disabled.")
 
     logging.info(f"🌍 Posting language(s): {post_langs}")
 
+    # ── Age-filter cutoff ────────────────────────────────────────
+    if max_age_days > 0:
+        cutoff = arrow.utcnow().shift(days=-max_age_days)
+        logging.info(
+            f"📅 Age filter active: skipping entries published before "
+            f"{cutoff.isoformat()} (max_age_days={max_age_days})"
+        )
+    else:
+        cutoff = None
+        logging.info("📅 Age filter disabled (max_age_days=0).")
+    # ────────────────────────────────────────────────────────────
+
     if check_post_cooldown_or_log(cooldown_path):
         return RunResult(published_count=0, stopped_reason="global_post_cooldown_active")
 
@@ -1197,12 +1212,35 @@ def run_once(
         logging.info(f"📰 Prepared {len(candidates)} feed entry candidates for duplicate comparison.")
 
         entries_to_post: List[EntryCandidate] = []
+
         for candidate in candidates:
+
+            # ── Age filter ───────────────────────────────────────────
+            if cutoff is not None:
+                pub = candidate.published_arrow
+                if pub is None:
+                    logging.info(
+                        f"⏭️ Skipping entry with no publication date "
+                        f"(age filter active, max_age_days={max_age_days}): "
+                        f"{candidate.canonical_link or candidate.title_text}"
+                    )
+                    continue
+                if pub < cutoff:
+                    logging.info(
+                        f"⏭️ Skipping old entry published {pub.isoformat()} "
+                        f"(cutoff: {cutoff.isoformat()}): "
+                        f"{candidate.canonical_link or candidate.title_text}"
+                    )
+                    continue
+            # ────────────────────────────────────────────────────────
+
+            # ── Deduplication: local state ───────────────────────────
             is_dup_state, reason_state = candidate_matches_state(candidate, state)
             if is_dup_state:
                 logging.info(f"⏭️ Skipping candidate due to local state duplicate match on: {reason_state}")
                 continue
 
+            # ── Deduplication: recent Bluesky posts ──────────────────
             is_dup_bsky, reason_bsky = candidate_matches_existing_bsky(candidate, recent_bsky_posts)
             if is_dup_bsky:
                 logging.info(f"⏭️ Skipping candidate due to recent Bluesky duplicate match on: {reason_bsky}")
@@ -1212,9 +1250,11 @@ def run_once(
 
         logging.info(f"📬 {len(entries_to_post)} entries remain after duplicate filtering.")
 
-        # ← NEW: log the effective cap before starting the loop
         if len(entries_to_post) > max_posts:
-            logging.info(f"🔢 max-posts cap is {max_posts}: will publish at most {max_posts} of {len(entries_to_post)} entries this run.")
+            logging.info(
+                f"🔢 max-posts cap is {max_posts}: will publish at most {max_posts} "
+                f"of {len(entries_to_post)} entries this run."
+            )
 
         if not entries_to_post:
             logging.info("ℹ️ Execution finished: no new entries to publish.")
@@ -1227,7 +1267,6 @@ def run_once(
 
         for candidate in entries_to_post:
 
-            # ← NEW: hard cap check at the top of every iteration
             if published >= max_posts:
                 logging.info(f"🔢 === MAX POSTS REACHED === Stopping after {published} posts (limit: {max_posts}).")
                 break
@@ -1311,6 +1350,7 @@ def run_once(
 
     return RunResult(published_count=published)
 
+
 # ============================================================
 # CLI
 # ============================================================
@@ -1330,10 +1370,15 @@ def main():
     )
     parser.add_argument("--state-path", default=DEFAULT_STATE_PATH, help="Path to local JSON state file")
     parser.add_argument("--cooldown-path", default=DEFAULT_COOLDOWN_STATE_PATH, help="Path to shared cooldown JSON state file")
-    parser.add_argument('--max-posts', type=int, default=5, help='Max new posts to publish per run')
+    parser.add_argument("--max-posts", type=int, default=5, help="Max new posts to publish per run (default: 5)")
+    parser.add_argument(
+        "--max-age-days",
+        type=int,
+        default=7,
+        help="Skip entries older than this many days (default: 7). Use 0 to disable the age filter.",
+    )
     args = parser.parse_args()
 
-    # Parse comma-separated langs: "ca,es" → ["ca", "es"]
     post_langs = [lang.strip() for lang in args.lang.split(",") if lang.strip()]
     if not post_langs:
         post_langs = ["ca"]
@@ -1348,13 +1393,14 @@ def main():
         bsky_username=args.bsky_username,
         bsky_password=args.bsky_app_password,
         service_url=args.service,
-        post_langs=args.lang.split(","),
+        post_langs=post_langs,
         state_path=args.state_path,
         cooldown_path=args.cooldown_path,
-        cfg=AppConfig(),
-        max_posts=args.max_posts 
+        cfg=cfg,
+        max_posts=args.max_posts,
+        max_age_days=args.max_age_days,     # ← NEW
     )
 
 
 if __name__ == "__main__":
-    main()
+    main()