Update api/main.py

api/main.py

@@ -36,29 +36,31 @@ def load() -> dict:
 
 
 def sorted_matches(d: dict):
-    """Return all matches sorted by timestamp, fixing any year-rollover issues."""
+    """
+    Split by score: has score = past, no score = future.
+    Combines all_past + all_future, deduplicates, filters year-rollover ghosts.
+    """
     now_ts     = datetime.now(PT_TZ).timestamp()
+    max_future = now_ts + 60 * 60 * 24 * 180  # 6 months ahead
 
-    past   = []
+    all_matches = []
-    future = []
+    for m in d.get("all_past", []) + d.get("all_future", []):
+        ts = m.get("timestamp", 0)
+        if ts > 0 and ts < max_future:
+            all_matches.append(m)
 
-    for m in d.get("all_past", []):
+    # Deduplicate by (home, away, date)
-        if m.get("timestamp", 0) > 0:
+    seen, unique = set(), []
-            past.append(m)
+    for m in all_matches:
+        key = (m.get("home"), m.get("away"), m.get("date"))
+        if key not in seen:
+            seen.add(key)
+            unique.append(m)
 
-    for m in d.get("all_future", []):
+    unique.sort(key=lambda x: x["timestamp"])
-        if m.get("timestamp", 0) > 0:
-            future.append(m)
 
-    # Sort both
+    past   = [m for m in unique if m.get("score")]
-    past.sort(key=lambda x: x["timestamp"])
+    future = [m for m in unique if not m.get("score")]
-    future.sort(key=lambda x: x["timestamp"])
-
-    # Edge case: some "future" matches may have wrong year from scraper
-    # Keep only those with timestamp > now
-    future = [m for m in future if m["timestamp"] > now_ts]
-    # Keep only past with timestamp <= now
-    past   = [m for m in past if m["timestamp"] <= now_ts]
 
     return past, future