Merge pull request #34 from cuappdev/lax-fix

Fix missing team field and bad formatting for lacrosse

Author: skyeslattery

src/scrapers/game_details_scrape.py

@@ -1,7 +1,26 @@
+import re
 import requests
 from bs4 import BeautifulSoup
 from src.utils.constants import *
 
+def clean_name(name):
+    """Strip extra information from player names, keeping only first and last name."""
+    # try to match firstname, lastname format
+    if ',' in name:
+        match = re.match(r'^([^,]+),\s*(\w+)', name)
+        if match:
+            return f"{match.group(1)}, {match.group(2)}"
+    else:
+        match = re.match(r'^(\w+)\s+(\w+)', name)
+        if match:
+            return f"{match.group(1)} {match.group(2)}"
+    
+    # fallback for removing common extra characters
+    cleaned = re.sub(r'\s*\([^)]*\).*$', '', name)
+    cleaned = re.sub(r'\s*\d+.*$', '', cleaned)
+    cleaned = cleaned.strip()
+    return cleaned
+
 def fetch_page(url):
     response = requests.get(url)
     return BeautifulSoup(response.text, 'html.parser')
@@ -150,16 +169,25 @@ def lacrosse_summary(box_score_section):
         scoring_rows = scoring_table.find(TAG_TBODY)
         if scoring_rows:
             for row in scoring_rows.find_all(TAG_TR):
+                team = row.find_all(TAG_TD)[1].find(TAG_IMG)[ATTR_ALT]
                 period = row.find_all(TAG_TD)[2].text.strip()
                 time = row.find_all(TAG_TD)[3].text.strip()
-                scorer = row.find_all(TAG_TD)[4].text.strip()
-                assist = row.find_all(TAG_TD)[5].text.strip()
+                scorer = clean_name(row.find_all(TAG_TD)[4].text.strip())
+                assist = clean_name(row.find_all(TAG_TD)[5].text.strip())
                 opp_score = row.find_all(TAG_TD)[7].text.strip()
                 cor_score = row.find_all(TAG_TD)[6].text.strip()
+                
+                if assist and assist != "Unassisted":
+                    desc = f"Scored by {scorer}, assisted by {assist}"
+                else:
+                    desc = f"Scored by {scorer}"
+
                 summary.append({
+                    'team': team,
                     'period': period,
                     'time': time,
                     'scorer': scorer,
+                    'description': desc,
                     'assist': assist,
                     'cor_score': cor_score,
                     'opp_score': opp_score,

src/scrapers/games_scraper.py

@@ -170,7 +170,9 @@ def process_game_data(game_data):
         city = geo_location
         state = geo_location
     else:
-        city, state = map(str.strip, geo_location.split(","))
+        parts = [part.strip() for part in geo_location.split(",")]
+        city = parts[0]
+        state = parts[-1]
     location = location_data[1] if len(location_data) > 1 else None
 
     team = TeamService.get_team_by_name(game_data["opponent_name"])
@@ -228,7 +230,6 @@ def process_game_data(game_data):
 
             # If they don't match, flip the arrays
             if str(final_box_cor_score) != str(cor_final) or str(final_box_opp_score) != str(opp_final):
-                print("flipping")
                 game_data["score_breakdown"] = game_data["score_breakdown"][::-1]
 
     # finds any existing game with the same key fields regardless of time