Update player_matcher.py

player_matcher.py

@@ -1,63 +1,43 @@
-"""
-Player Name Matching Engine
-Scans text for NBA player mentions and returns matched players.
-"""
+"""Fast Player Name Matching - Optimized for speed"""
 
 import re
 from nba_players import NBA_PLAYERS
 
-# Build search patterns
-PLAYER_PATTERNS = {}
+# Build lookup structures once at import time
+FULL_NAMES = set(p.lower() for p in NBA_PLAYERS)
 FIRST_NAMES = {}
 LAST_NAMES = {}
 
 for player in NBA_PLAYERS:
     parts = player.split()
     if len(parts) >= 2:
-        first = parts[0]
-        last = parts[-1]
-        
-        # Full name pattern
-        PLAYER_PATTERNS[player.lower()] = player
-        
-        # Track first and last names for partial matching
-        if first.lower() not in FIRST_NAMES:
-            FIRST_NAMES[first.lower()] = []
-        FIRST_NAMES[first.lower()].append(player)
-        
-        if last.lower() not in LAST_NAMES:
-            LAST_NAMES[last.lower()] = []
-        LAST_NAMES[last.lower()].append(player)
+        first, last = parts[0].lower(), parts[-1].lower()
+        if last not in LAST_NAMES:
+            LAST_NAMES[last] = []
+        LAST_NAMES[last].append(player)
 
-# Common names that need full name to match (first or last names)
-AMBIGUOUS_NAMES = {
-    # Last names
-    "james", "johnson", "williams", "brown", "jones", "davis", "smith",
-    "thompson", "jackson", "white", "harris", "martin", "moore", "taylor",
-    "thomas", "clark", "robinson", "walker", "green", "hall", "allen",
-    "young", "king", "wright", "hill", "scott", "adams", "baker", "nelson",
-    "mitchell", "carter", "roberts", "turner", "phillips", "campbell",
-    "parker", "evans", "edwards", "collins", "stewart", "morris", "murphy",
-    "gordon", "murray", "barnes", "curry", "ball", "wagner", "miller",
-    "porter", "grant", "reed", "powell", "butler", "jordan", "paul",
-    "hunter", "love", "smart", "holiday", "mann", "wade", "monk", "dunn",
-    "nance", "cole", "reed", "huff", "temple", "bates", "cook", "bell",
-    # First names that are common
-    "will", "aaron", "chris", "kevin", "gary", "nick", "cole", "josh",
-    "mike", "anthony", "cooper", "tre", "cam", "max", "pat", "drew",
-    "jordan", "isaiah", "marcus", "jalen", "tyler", "ryan", "kyle",
-    "joe", "ben", "sam", "tim", "dan", "tom", "john", "david", "alex",
-    "james", "robert", "michael", "william", "richard", "charles",
-    "daniel", "matthew", "andrew", "joshua", "christopher", "joseph",
-    # Common words that match player names
-    "post", "bridges", "rose", "wolf", "wells", "rice", "prince",
-    "little", "green", "brown", "white", "black", "gray", "best",
-    "freeman", "washington", "boston", "houston", "phoenix", "orlando",
-    "brooks", "rivers", "banks", "stone", "wood", "waters", "field",
-    "hart", "fox", "dean", "terry", "craig", "ross", "lane", "day",
+# Unique last names (only one player has this last name)
+UNIQUE_LAST_NAMES = {k: v[0] for k, v in LAST_NAMES.items() if len(v) == 1}
+
+# Common words that should NOT match even if they're unique last names
+EXCLUDED_WORDS = {
+    "post", "love", "smart", "holiday", "green", "brown", "white", "black",
+    "gray", "king", "young", "long", "little", "best", "hart", "wade",
+    "rose", "bridges", "wall", "hill", "rice", "prince", "mann", "day",
+    "wells", "reed", "waters", "brooks", "banks", "bell", "ford", "lane",
+    "grant", "stone", "wood", "wolf", "fox", "dean", "ross", "terry",
+    "craig", "cole", "huff", "cook", "monk", "dunn", "bates", "temple",
+    "joe", "ben", "sam", "tim", "dan", "tom", "john", "will", "max",
+    "cam", "tre", "pat", "drew", "alex", "nick", "gary", "josh", "mike",
+    "chris", "kyle", "ryan", "tyler", "jordan", "isaiah", "marcus", "jalen",
+    "boston", "houston", "phoenix", "orlando", "washington", "cleveland",
 }
 
-# Famous nicknames -> canonical names
+# Remove excluded words from unique last names
+for word in EXCLUDED_WORDS:
+    UNIQUE_LAST_NAMES.pop(word, None)
+
+# Nicknames mapping
 NICKNAMES = {
     "lebron": "LeBron James",
     "steph": "Stephen Curry",
@@ -67,7 +47,6 @@ NICKNAMES = {
     "embiid": "Joel Embiid",
     "kd": "Kevin Durant",
     "ad": "Anthony Davis",
-    "pg": "Paul George",
     "cp3": "Chris Paul",
     "ja": "Ja Morant",
     "trae": "Trae Young",
@@ -85,7 +64,6 @@ NICKNAMES = {
     "harden": "James Harden",
     "westbrook": "Russell Westbrook",
     "sabonis": "Domantas Sabonis",
-    "fox": "De'Aaron Fox",
     "mobley": "Evan Mobley",
     "cade": "Cade Cunningham",
     "paolo": "Paolo Banchero",
@@ -93,57 +71,55 @@ NICKNAMES = {
     "scottie": "Scottie Barnes",
     "brunson": "Jalen Brunson",
     "flagg": "Cooper Flagg",
+    "maxey": "Tyrese Maxey",
+    "halliburton": "Tyrese Haliburton",
+    "haliburton": "Tyrese Haliburton",
+    "garland": "Darius Garland",
+    "murray": "Jamal Murray",
+    "dejounte": "Dejounte Murray",
+    "bam": "Bam Adebayo",
+    "herro": "Tyler Herro",
+    "jimmy": "Jimmy Butler",
+    "kawhi": "Kawhi Leonard",
+    "pg13": "Paul George",
 }
 
+# Precompile word pattern
+WORD_PATTERN = re.compile(r"[a-zA-Z']+")
+
 
 def find_player_mentions(text):
-    """
-    Find all NBA player mentions in text.
-    
-    Returns:
-        dict: {canonical_name: [list of matched strings]}
-    """
+    """Fast player mention detection."""
     if not text:
         return {}
     
-    text_lower = text.lower()
     found = {}
+    text_lower = text.lower()
+    words = WORD_PATTERN.findall(text_lower)
+    word_set = set(words)
     
-    # Check nicknames first
+    # Check nicknames (fast set lookup)
     for nick, player in NICKNAMES.items():
-        pattern = r'\b' + re.escape(nick) + r'\b'
-        if re.search(pattern, text_lower):
-            if player not in found:
-                found[player] = []
-            found[player].append(nick)
-    
-    # Check full names
-    for player in NBA_PLAYERS:
-        pattern = r'\b' + re.escape(player.lower()) + r'\b'
-        if re.search(pattern, text_lower):
-            if player not in found:
-                found[player] = []
-            found[player].append(player)
+        if nick in word_set:
+            found[player] = [nick]
     
-    # Check unique last names (not ambiguous)
-    words = set(re.findall(r'\b[a-zA-Z]+\b', text_lower))
+    # Check full names by looking at word pairs
+    for i in range(len(words) - 1):
+        full_name = words[i] + " " + words[i + 1]
+        if full_name in FULL_NAMES:
+            # Find proper case version
+            for player in NBA_PLAYERS:
+                if player.lower() == full_name:
+                    if player not in found:
+                        found[player] = []
+                    found[player].append(full_name)
+                    break
     
-    for word in words:
-        if word in AMBIGUOUS_NAMES:
-            continue
-        
-        # Check if it's a unique last name
-        if word in LAST_NAMES:
-            matches = LAST_NAMES[word]
-            if len(matches) == 1:
-                player = matches[0]
-                if player not in found:
-                    found[player] = []
-                    found[player].append(word)
+    # Check unique last names (but not excluded common words)
+    for word in word_set:
+        if word in UNIQUE_LAST_NAMES:
+            player = UNIQUE_LAST_NAMES[word]
+            if player not in found:
+                found[player] = [word]
     
-    return found
-
-
-def get_mentioned_players(text):
-    """Get just the list of mentioned players."""
-    return list(find_player_mentions(text).keys())
+    return found