Update app.py

app.py

@@ -63,7 +63,7 @@ SOCCER_LEAGUES = {
 
 def get_team_news(team: str, sport: str = "football"):
     url = f"https://news.google.com/search?q={team}+{sport}&hl=en"
-    r = requests.get(url, headers={"User-Agent": "Mozilla/5.0"})
+    r = requests.get(url, headers={"User-Agent": "Mozilla/5.0"}, timeout=10)
     soup = BeautifulSoup(r.text, "html.parser")
     headlines = [h.text for h in soup.select("h3")[:5]]
     return " ".join(headlines)
@@ -142,7 +142,7 @@ def scrape_nba_preseason_pairs_for_dates(dates):
             q = qtpl.format(d=d)
             url = f"https://www.google.com/search?q={q}"
             try:
-                r = requests.get(url, headers=headers, timeout=20)
+                r = requests.get(url, headers=headers, timeout=10)
                 soup = BeautifulSoup(r.text, "html.parser")
                 h3s = [h.text for h in soup.select("h3")[:30]]
                 pairs = _extract_pairs_from_texts(h3s)
@@ -168,6 +168,48 @@ def build_soccer_verdict(home, away, agg):
     conf_ou = max(ou_25.get("over", 0.0), ou_25.get("under", 0.0))
     return f"{home} vs {away}: {pick_1x2} ({conf_1x2:.2f}), {pick_ou} ({conf_ou:.2f})"
 
+def _score_zero_shot(sent_result):
+    # Expecting HF zero-shot output: {labels: [...], scores: [...]} possibly wrapped
+    try:
+        labels = sent_result.get('labels') or []
+        scores = sent_result.get('scores') or []
+        table = {lbl: scores[i] for i, lbl in enumerate(labels)}
+        positive = float(table.get('positive', 0.0))
+        negative = float(table.get('negative', 0.0))
+        injury = float(table.get('injury', 0.0))
+        motivation = float(table.get('motivation', 0.0))
+        transfer = float(table.get('transfer', 0.0))
+        return {
+            'positive': positive,
+            'negative': negative,
+            'injury': injury,
+            'motivation': motivation,
+            'transfer': transfer,
+            'net': positive - negative - 0.5 * injury + 0.25 * motivation
+        }
+    except Exception:
+        return {'positive': 0.0, 'negative': 0.0, 'injury': 0.0, 'motivation': 0.0, 'transfer': 0.0, 'net': 0.0}
+
+def build_nba_verdict(home, away, home_sig, away_sig):
+    # Convert net signals into a simple probability with softmax-like scaling
+    h = home_sig.get('net', 0.0)
+    a = away_sig.get('net', 0.0)
+    margin = h - a
+    # logistic transform to [0,1]
+    import math
+    p_home = 1.0 / (1.0 + math.exp(-3.0 * margin))
+    p_away = 1.0 - p_home
+    # Over/Under lean from total sentiment energy
+    energy = max(0.0, home_sig.get('positive', 0.0) + away_sig.get('positive', 0.0) + home_sig.get('motivation', 0.0) + away_sig.get('motivation', 0.0))
+    injury_load = home_sig.get('injury', 0.0) + away_sig.get('injury', 0.0)
+    ou_pick = 'Over' if energy >= (0.6 + 0.3 * injury_load) else 'Under'
+    return {
+        'winner_probs': { home: p_home, away: p_away },
+        'winner_pick': home if p_home >= p_away else away,
+        'winner_confidence': max(p_home, p_away),
+        'ou_pick': ou_pick
+    }
+
 # --- Helpers: Football-Data last-5 and simple Poisson model ---
 def football_headers():
     return {"X-Auth-Token": FOOTBALL_API_KEY}
@@ -323,6 +365,10 @@ def soccer_predictions():
             continue
 
         for match in matches:
+            # Skip matches that are already played or not upcoming
+            m_status = (match.get("status") or "").upper()
+            if m_status and m_status not in ("SCHEDULED", "TIMED", "POSTPONED"):  # e.g., FINISHED, IN_PLAY, PAUSED
+                continue
             home, away = match["homeTeam"]["name"], match["awayTeam"]["name"]
             home_id = match["homeTeam"].get("id")
             away_id = match["awayTeam"].get("id")
@@ -567,20 +613,26 @@ def nba_predictions():
             home_sent = sentiment_model(default_text, labels)
             away_sent = sentiment_model(default_text, labels)
 
+        sig_home = _score_zero_shot(home_sent if isinstance(home_sent, dict) else (home_sent[0] if isinstance(home_sent, list) else {}))
+        sig_away = _score_zero_shot(away_sent if isinstance(away_sent, dict) else (away_sent[0] if isinstance(away_sent, list) else {}))
+
         context = f"""
         Match: {home} vs {away}.
         Home news: {home_news}.
         Away news: {away_news}.
+        Signals: home={sig_home}, away={sig_away}.
         """
         reasoning = reasoning_model(
             f"Provide a concise NBA verdict: likely winner and Over/Under guidance. Context: {context}",
             max_new_tokens=256
         )[0]['generated_text']
 
+        verdict = build_nba_verdict(home, away, sig_home, sig_away)
+
         predictions.append({
             "match": f"{home} vs {away}",
             "reason": reasoning,
-            "verdict": f"{home} vs {away}: {reasoning[:140]}...",
+            "verdict": verdict,
             "news_summary": {
                 home: home_sent,
                 away: away_sent