capturing and integrating multimedia assets part 1

data/april_11_multimedia_data_collect/get_player_socials.py

@@ -0,0 +1,239 @@
+import requests
+import csv
+import os
+import time
+import sys
+from dotenv import load_dotenv
+from pathlib import Path
+
+# Load environment variables from .env file (for API key)
+load_dotenv()
+
+# Get the directory where this script is located
+SCRIPT_DIR = Path(os.path.dirname(os.path.abspath(__file__)))
+
+SERP_API_KEY = os.getenv("SERP_API_KEY")  # Or just hardcode for testing, not recommended
+
+def get_instagram_handle(query, timeout=10, retries=3, delay_between_retries=2):
+    """
+    Uses SerpAPI to search for query: e.g. 'Brock Purdy Instagram'
+    Returns the best guess at Instagram handle/page URL if found, else empty string.
+    
+    Args:
+        query: Search query string
+        timeout: Request timeout in seconds
+        retries: Number of retries if request fails
+        delay_between_retries: Seconds to wait between retries
+    """
+    if not SERP_API_KEY:
+        raise ValueError("SERP_API_KEY environment variable not set or provided!")
+    
+    url = "https://serpapi.com/search"
+    params = {
+        "engine": "google",
+        "q": query,
+        "api_key": SERP_API_KEY,
+    }
+    
+    for attempt in range(retries):
+        try:
+            print(f"[DEBUG] Sending API request for: {query}")
+            response = requests.get(url, params=params, timeout=timeout)
+            response.raise_for_status()
+            data = response.json()
+            
+            # Check if we have organic results
+            if "organic_results" not in data:
+                print(f"[WARNING] No organic_results found in API response for {query}")
+                print(f"[DEBUG] Response keys: {list(data.keys())}")
+                return ""
+            
+            # Typical structure: data['organic_results'] - parse each for relevant domain
+            results = data.get("organic_results", [])
+            print(f"[DEBUG] Found {len(results)} organic results")
+            
+            for r in results:
+                link = r.get("link", "")
+                # If it has 'instagram.com', let's assume it's correct
+                if "instagram.com" in link.lower():
+                    print(f"[DEBUG] Found Instagram link: {link}")
+                    return link
+            
+            print(f"[WARNING] No Instagram links found for {query}")
+            return ""
+            
+        except requests.exceptions.Timeout:
+            print(f"[ERROR] Request timed out for {query} (attempt {attempt+1}/{retries})")
+            if attempt < retries - 1:
+                print(f"[INFO] Retrying in {delay_between_retries} seconds...")
+                time.sleep(delay_between_retries)
+            else:
+                print(f"[ERROR] All retries failed for {query}")
+                return ""
+                
+        except requests.exceptions.RequestException as e:
+            print(f"[ERROR] Request failed for {query}: {str(e)} (attempt {attempt+1}/{retries})")
+            if attempt < retries - 1:
+                print(f"[INFO] Retrying in {delay_between_retries} seconds...")
+                time.sleep(delay_between_retries)
+            else:
+                print(f"[ERROR] All retries failed for {query}")
+                return ""
+                
+        except Exception as e:
+            print(f"[ERROR] Unexpected error for {query}: {str(e)}")
+            return ""
+
+def enrich_niners_socials(input_csv='niners_players_headshots.csv',
+                          output_csv='niners_players_headshots_with_socials.csv',
+                          delay_between_requests=1,
+                          start_player=None,
+                          max_players=None):
+    """
+    Reads the roster CSV, queries Instagram for each player's best match,
+    then writes the results to a new CSV.
+    
+    Args:
+        input_csv: Path to input CSV file
+        output_csv: Path to output CSV file
+        delay_between_requests: Seconds to wait between API requests to avoid rate limiting
+        start_player: Player number to start processing from (1-indexed)
+        max_players: Maximum number of players to process (None for all)
+    """
+    # Convert relative paths to absolute paths based on script directory
+    if not os.path.isabs(input_csv):
+        input_csv = os.path.join(SCRIPT_DIR, input_csv)
+    if not os.path.isabs(output_csv):
+        output_csv = os.path.join(SCRIPT_DIR, output_csv)
+        
+    print(f"[INFO] Input CSV path: {input_csv}")
+    print(f"[INFO] Output CSV path: {output_csv}")
+    if not SERP_API_KEY:
+        print("[ERROR] SERP_API_KEY not set. Please set your environment variable or update the script.")
+        return
+    
+    # Check if input file exists
+    if not os.path.exists(input_csv):
+        print(f"[ERROR] Input file not found: {input_csv}")
+        return
+    
+    try:
+        # Read existing output CSV if it exists to continue from where we left off
+        existing_data = []
+        if os.path.exists(output_csv):
+            with open(output_csv, 'r', encoding='utf-8') as f_existing:
+                existing_reader = csv.DictReader(f_existing)
+                existing_data = list(existing_reader)
+                print(f"[INFO] Loaded {len(existing_data)} existing entries")
+        
+        # Count total players for progress reporting
+        with open(input_csv, 'r', encoding='utf-8') as f:
+            total_players = sum(1 for _ in csv.DictReader(f))
+        
+        print(f"[INFO] Total players: {total_players}")
+        
+        # Determine start and end points
+        start_index = start_player - 1 if start_player is not None else len(existing_data)
+        end_index = min(total_players, start_index + (max_players or total_players))
+        
+        print(f"[INFO] Will process players from {start_index + 1} to {end_index}")
+        
+        # Reopen input CSV to start processing
+        with open(input_csv, 'r', encoding='utf-8') as f:
+            reader = csv.DictReader(f)
+            input_fieldnames = reader.fieldnames
+            
+            # Skip to the start player
+            for _ in range(start_index):
+                next(reader)
+            
+            # Process remaining players
+            for i, row in enumerate(reader, start_index + 1):
+                if i > end_index:
+                    print(f"[INFO] Reached maximum number of players. Stopping.")
+                    break
+                    
+                player_name = row['name']
+                print(f"[INFO] Processing player {i}/{end_index}: {player_name}")
+                
+                # Skip if already processed
+                if any(existing_row['name'] == player_name for existing_row in existing_data):
+                    print(f"[INFO] {player_name} already processed. Skipping.")
+                    continue
+                
+                # Construct a query like 'PLAYER NAME instagram'
+                query = f"{player_name} NFL 49ers instagram"
+                
+                try:
+                    insta_url = get_instagram_handle(query)
+                    row['instagram_url'] = insta_url
+                    
+                    # Print result
+                    if insta_url:
+                        print(f"[SUCCESS] Found Instagram for {player_name}: {insta_url}")
+                    else:
+                        print(f"[WARNING] No Instagram found for {player_name}")
+                    
+                    # Append new data
+                    existing_data.append(row)
+                    
+                    # Save progress after each player
+                    with open(output_csv, 'w', newline='', encoding='utf-8') as f_out:
+                        output_fieldnames = input_fieldnames + ['instagram_url']
+                        writer = csv.DictWriter(f_out, fieldnames=output_fieldnames)
+                        writer.writeheader()
+                        writer.writerows(existing_data)
+                    
+                    # Add delay between requests to avoid rate limiting
+                    if i < end_index:
+                        print(f"[INFO] Waiting {delay_between_requests} seconds before next request...")
+                        time.sleep(delay_between_requests)
+                        
+                except KeyboardInterrupt:
+                    print("\n[INFO] Process interrupted by user. Saving progress...")
+                    break
+        
+        print(f"[INFO] Social data saved to {output_csv}")
+        print(f"[INFO] Processed {len(existing_data)}/{total_players} players")
+        
+    except Exception as e:
+        print(f"[ERROR] An unexpected error occurred: {str(e)}")
+        # Try to save any data collected so far
+        if existing_data:
+            try:
+                with open(output_csv, 'w', newline='', encoding='utf-8') as f_out:
+                    output_fieldnames = input_fieldnames + ['instagram_url']
+                    writer = csv.DictWriter(f_out, fieldnames=output_fieldnames)
+                    writer.writeheader()
+                    writer.writerows(existing_data)
+                print(f"[INFO] Partial data saved to {output_csv}")
+            except Exception:
+                print("[ERROR] Failed to save partial data")
+
+if __name__ == "__main__":
+    print("[INFO] Starting player social media enrichment script")
+    
+    # Parse command line arguments
+    delay = 1  # Default delay
+    start_player = 51  # Default to start from 51st player
+    max_players = None  # Process all remaining players
+    
+    if len(sys.argv) > 1:
+        try:
+            delay = float(sys.argv[1])
+            print(f"[INFO] Using custom delay between requests: {delay} seconds")
+        except ValueError:
+            print(f"[WARNING] Invalid delay value: {sys.argv[1]}. Using default: 1 second")
+    
+    if len(sys.argv) > 2:
+        try:
+            start_player = int(sys.argv[2])
+            print(f"[INFO] Will start processing from player {start_player}")
+        except ValueError:
+            print(f"[WARNING] Invalid start_player value: {sys.argv[2]}. Using default: 51")
+    
+    enrich_niners_socials(
+        delay_between_requests=delay, 
+        start_player=start_player, 
+        max_players=max_players
+    )

data/april_11_multimedia_data_collect/get_youtube_playlist_videos.py

@@ -0,0 +1,65 @@
+import os
+import csv
+from googleapiclient.discovery import build
+from dotenv import load_dotenv
+from pathlib import Path
+
+# Load environment variables from .env file (for API key)
+load_dotenv()
+
+API_KEY = os.getenv("YOUTUBE_API_KEY")  # Or replace with your key in code
+# Example 49ers highlights playlist:
+PLAYLIST_ID = "PLBB205pkCsyvZ6tjCh_m5s21D0eeYJ8Ly"
+
+def get_youtube_videos(playlist_id=PLAYLIST_ID, output_csv='youtube_highlights.csv'):
+    """
+    Fetches videos from a YouTube playlist (title, video ID, published date, etc.)
+    Writes output to CSV.
+    """
+    if not API_KEY:
+        raise ValueError("YOUTUBE_API_KEY environment variable not set or provided!")
+    
+    youtube = build('youtube', 'v3', developerKey=API_KEY)
+    
+    video_data = []
+    page_token = None
+
+    while True:
+        playlist_req = youtube.playlistItems().list(
+            part="snippet",
+            playlistId=playlist_id,
+            maxResults=50,
+            pageToken=page_token
+        )
+        playlist_res = playlist_req.execute()
+        
+        for item in playlist_res['items']:
+            snippet = item['snippet']
+            title = snippet['title']
+            description = snippet['description']
+            video_id = snippet['resourceId']['videoId']
+            published_at = snippet['publishedAt']
+            
+            video_data.append({
+                "video_id": video_id,
+                "title": title,
+                "description": description,
+                "published_at": published_at,
+                "video_url": f"https://www.youtube.com/watch?v={video_id}"
+            })
+        
+        page_token = playlist_res.get('nextPageToken')
+        if not page_token:
+            break
+    
+    # Write to CSV
+    with open(output_csv, 'w', newline='', encoding='utf-8') as f:
+        fieldnames = ["video_id", "title", "description", "published_at", "video_url"]
+        writer = csv.DictWriter(f, fieldnames=fieldnames)
+        writer.writeheader()
+        writer.writerows(video_data)
+    
+    print(f"[INFO] YouTube playlist data saved to {output_csv}")
+
+if __name__ == "__main__":
+    get_youtube_videos()

data/april_11_multimedia_data_collect/match_highlights.py

@@ -0,0 +1,193 @@
+import csv
+import re
+from collections import defaultdict
+
+# Define file paths
+YOUTUBE_HIGHLIGHTS_PATH = "ifx-sandbox/data/april_11_multimedia_data_collect/youtube_highlights.csv"
+PLAYERS_ROSTER_PATH = "ifx-sandbox/data/april_11_multimedia_data_collect/niners_players_headshots_with_socials_merged.csv"
+GAMES_SCHEDULE_PATH = "ifx-sandbox/data/nfl-2024-san-francisco-49ers-with-results.csv"
+OUTPUT_PLAYERS_PATH = "ifx-sandbox/data/niners_output/players_with_highlights.csv"
+OUTPUT_GAMES_PATH = "ifx-sandbox/data/niners_output/games_with_highlights.csv"
+OUTPUT_TEAM_VIDEOS_PATH = "ifx-sandbox/data/niners_output/team_highlights.csv"
+
+def load_youtube_highlights():
+    """Load YouTube highlights data from CSV file."""
+    highlights = []
+    with open(YOUTUBE_HIGHLIGHTS_PATH, 'r', encoding='utf-8') as file:
+        reader = csv.DictReader(file)
+        for row in reader:
+            highlights.append({
+                'video_id': row['video_id'],
+                'title': row['title'],
+                'description': row['description'],
+                'published_at': row['published_at'],
+                'video_url': row['video_url']
+            })
+    return highlights
+
+def load_players():
+    """Load player roster data from CSV file."""
+    players = []
+    with open(PLAYERS_ROSTER_PATH, 'r', encoding='utf-8') as file:
+        reader = csv.DictReader(file)
+        for row in reader:
+            players.append({
+                'name': row['name'],
+                'headshot_url': row['headshot_url'],
+                'instagram_url': row['instagram_url'],
+                'highlight_video_url': ''  # Initialize with empty string
+            })
+    return players
+
+def load_games():
+    """Load game schedule data from CSV file."""
+    games = []
+    with open(GAMES_SCHEDULE_PATH, 'r', encoding='utf-8') as file:
+        reader = csv.DictReader(file)
+        for row in reader:
+            opponent = row['Away Team'] if row['Home Team'] == 'San Francisco 49ers' else row['Home Team']
+            opponent = opponent.replace('San Francisco 49ers', '').strip()
+            
+            games.append({
+                'match_number': row['Match Number'],
+                'round_number': row['Round Number'],
+                'date': row['Date'],
+                'location': row['Location'],
+                'home_team': row['Home Team'],
+                'away_team': row['Away Team'],
+                'result': row['Result'],
+                'game_result': row['game_result'],
+                'opponent': opponent,
+                'highlight_video_url': ''  # Initialize with empty string
+            })
+    return games
+
+def match_highlights_to_players_and_games(highlights, players, games):
+    """Match YouTube highlights to players and games."""
+    # Create a copy of highlights to track which ones are assigned
+    unassigned_highlights = highlights.copy()
+    
+    # Track assigned videos
+    assigned_video_ids = set()
+    
+    # Match players first
+    for player in players:
+        player_name = player['name']
+        first_name = player_name.split()[0]
+        last_name = player_name.split()[-1]
+        
+        # Create patterns to match player names
+        full_name_pattern = re.compile(r'\b' + re.escape(player_name) + r'\b', re.IGNORECASE)
+        last_name_pattern = re.compile(r'\b' + re.escape(last_name) + r'\b', re.IGNORECASE)
+        
+        # Try to find a match in the unassigned highlights
+        for highlight in unassigned_highlights:
+            if highlight['video_id'] in assigned_video_ids:
+                continue
+                
+            title = highlight['title']
+            description = highlight['description']
+            
+            # Check for full name match in title first (most specific)
+            if full_name_pattern.search(title):
+                player['highlight_video_url'] = highlight['video_url']
+                assigned_video_ids.add(highlight['video_id'])
+                break
+            
+            # Then check for last name match in title
+            elif last_name_pattern.search(title):
+                player['highlight_video_url'] = highlight['video_url']
+                assigned_video_ids.add(highlight['video_id'])
+                break
+    
+    # Match games next
+    for game in games:
+        opponent = game['opponent']
+        week_pattern = re.compile(r'\bWeek\s+' + re.escape(game['round_number']) + r'\b', re.IGNORECASE)
+        opponent_pattern = re.compile(r'\b' + re.escape(opponent) + r'\b', re.IGNORECASE)
+        
+        # Try to find a match in the unassigned highlights
+        for highlight in unassigned_highlights:
+            if highlight['video_id'] in assigned_video_ids:
+                continue
+                
+            title = highlight['title']
+            description = highlight['description']
+            
+            # Check for both week and opponent match in title (most specific)
+            if week_pattern.search(title) and opponent_pattern.search(title):
+                game['highlight_video_url'] = highlight['video_url']
+                assigned_video_ids.add(highlight['video_id'])
+                break
+            
+            # Then check for opponent match in title
+            elif opponent_pattern.search(title):
+                game['highlight_video_url'] = highlight['video_url']
+                assigned_video_ids.add(highlight['video_id'])
+                break
+    
+    # Collect team videos (unassigned highlights)
+    team_videos = []
+    for highlight in highlights:
+        if highlight['video_id'] not in assigned_video_ids:
+            team_videos.append(highlight)
+    
+    return team_videos
+
+def save_players_with_highlights(players):
+    """Save players with highlight videos to CSV file."""
+    with open(OUTPUT_PLAYERS_PATH, 'w', newline='', encoding='utf-8') as file:
+        fieldnames = ['name', 'headshot_url', 'instagram_url', 'highlight_video_url']
+        writer = csv.DictWriter(file, fieldnames=fieldnames)
+        writer.writeheader()
+        for player in players:
+            writer.writerow(player)
+
+def save_games_with_highlights(games):
+    """Save games with highlight videos to CSV file."""
+    with open(OUTPUT_GAMES_PATH, 'w', newline='', encoding='utf-8') as file:
+        fieldnames = ['match_number', 'round_number', 'date', 'location', 'home_team', 'away_team', 
+                     'result', 'game_result', 'opponent', 'highlight_video_url']
+        writer = csv.DictWriter(file, fieldnames=fieldnames)
+        writer.writeheader()
+        for game in games:
+            writer.writerow(game)
+
+def save_team_videos(team_videos):
+    """Save team videos to CSV file."""
+    with open(OUTPUT_TEAM_VIDEOS_PATH, 'w', newline='', encoding='utf-8') as file:
+        fieldnames = ['video_id', 'title', 'description', 'published_at', 'video_url']
+        writer = csv.DictWriter(file, fieldnames=fieldnames)
+        writer.writeheader()
+        for video in team_videos:
+            writer.writerow(video)
+
+def main():
+    # Load data
+    highlights = load_youtube_highlights()
+    players = load_players()
+    games = load_games()
+    
+    # Match highlights to players and games
+    team_videos = match_highlights_to_players_and_games(highlights, players, games)
+    
+    # Save results
+    save_players_with_highlights(players)
+    save_games_with_highlights(games)
+    save_team_videos(team_videos)
+    
+    # Print summary
+    player_matches = sum(1 for player in players if player['highlight_video_url'])
+    game_matches = sum(1 for game in games if game['highlight_video_url'])
+    
+    print(f"Total YouTube highlights: {len(highlights)}")
+    print(f"Players with highlight videos: {player_matches}/{len(players)}")
+    print(f"Games with highlight videos: {game_matches}/{len(games)}")
+    print(f"Team videos (unassigned): {len(team_videos)}")
+    print(f"\nOutput files created:")
+    print(f"- {OUTPUT_PLAYERS_PATH}")
+    print(f"- {OUTPUT_GAMES_PATH}")
+    print(f"- {OUTPUT_TEAM_VIDEOS_PATH}")
+
+if __name__ == "__main__":
+    main()

data/april_11_multimedia_data_collect/niners_players_headshots.csv

@@ -0,0 +1,74 @@
+name,headshot_url
+Israel Abanikanda,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/wo7d9oli06eki4mnh3i8.png
+Brandon Aiyuk,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/khwofxjjwx0hcaigzxhw.png
+Isaac Alarcon,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/mlhuuxukyusodzlfsmnv.jpg
+Evan Anderson,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/ng7oamywxvqgkx6l6kqc.png
+Tre Avery,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/a7kfv7xjftqlaqghk6sg
+Alex Barrett,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/bm0ay22de39d1enrxwiq.jpg
+Ben Bartch,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/aqaslodzr7y0yvh5zzxa.jpg
+Robert Beal Jr.,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/jwwhmt5d8mi0vdb8nfic.jpg
+Tatum Bethune,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/vl08pinqpmoubdf0zy5s.png
+Nick Bosa,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/utiwswqvpkiwtocijwhz.jpg
+Jake Brendel,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/svsb41aekpzt3m9snilw.jpg
+Ji'Ayir Brown,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/urillpic02z774n09xvf.jpg
+Tre Brown,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/dpemqrrweakt8dci3qfb
+Spencer Burford,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/lje3ae25dntkdudp6eex.jpg
+Jacob Cowing,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/lg7aao0umc21oioufqdx.png
+Kalia Davis,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/rmnxj3sh7pyldmcxqe32.jpg
+Jordan Elliott,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/xbyky8r2yuzusd2tmrw8.jpg
+Luke Farrell,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/f2z7wpmx7ngtxcqqedla
+Russell Gage Jr.,https://static.www.nfl.com/image/private/t_thumb_squared_2x/f_auto/league/lkqhshv0dss1b9c6mdnj
+Jonathan Garvin,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/rapfcxut6vu50vcevswe.png
+Luke Gifford,https://static.www.nfl.com/image/private/t_thumb_squared_2x/f_auto/league/mhdbbzj8amttnpd1nbpn
+Kevin Givens,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/mstmgft0e0ancdzspboy.jpg
+Jalen Graham,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/pbl2a1ujopvwqrfct0jp.jpg
+Richie Grant,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/szeswtvt6jmbu3so3phd
+Renardo Green,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/v79obx9v7tgcjjlo6hiy.png
+Yetur Gross-Matos,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/etuaajmvhbc5qkebgoow.jpg
+Isaac Guerendo,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/b66rpzr9iauo5rdprvka.png
+Sebastian Gutierrez,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/am9sywgkga6jq65hvboe.jpg
+Matt Hennessy,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/zk8b21o8ncxnyu0gyf23
+Isaiah Hodgins,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/ax1oft9kqida0eokvtes.jpg
+Drake Jackson,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/y2luyplzpvbzokyfbmla.jpg
+Tarron Jackson,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/pnqrjp76bgpkmacxma3r
+Jauan Jennings,https://static.clubs.nfl.com/image/private/t_thumb_squared_2x/f_auto/49ers/wxsq7f4ajmhfs6tn4dg2.jpg
+Quindell Johnson,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/uga90lawcfxjcqna7opb
+Zack Johnson,https://static.www.nfl.com/image/private/t_thumb_squared_2x/f_auto/league/n4hy8uzhcl5cl0ricwoa
+Mac Jones,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/pedpdxybeus7mrovsoko
+Kyle Juszczyk,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/ywdz6y2pfzndqgmxxfbj.jpg
+George Kittle,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/elheepobwn1ahqwtfwat.jpg
+Deommodore Lenoir,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/f9fnuvbpcxku9ibt9qs8.jpg
+Chase Lucas,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/gjeejt5pbagnipodhdz4.jpg
+Darrell Luter Jr.,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/g5rohvooet9g5w7rlhrh.jpg
+Jaylen Mahoney,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/yv9inbia05nyxppuajv0.png
+Christian McCaffrey,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/a8fka6shomakkbllljgt.jpg
+Jalen McKenzie,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/gffxpns1ayxyjccymr6d.jpg
+Colton McKivitz,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/jugvoxjabgsbcfbuqfew.jpg
+Jake Moody,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/ygputwsbutemszr8xxkw.jpg
+Tanner Mordecai,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/y8gipodnkeapgmegnxs1.png
+Malik Mustapha,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/eyrgxgpbrycd9x8glk0j.png
+Siran Neal,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/muhthfs6owkkpsyop1e6
+Drake Nugent,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/qyb4kurtbv9uflmupfnc.png
+George Odum,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/sqpxhoycdpegkyjn6ooc.jpg
+Sam Okuayinonu,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/fyolr2zk2nplfbdze75l.jpg
+Terique Owens,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/okhin0uwdon2nimvbtwd.png
+Ricky Pearsall,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/to7q7w4kjiajseb4ljcx.png
+Jason Pinnock,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/on29awacb9frijyggtgt
+Austen Pleasants,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/wsbs5emdyzuc1sudbcls.png
+Mason Pline,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/mvjlaxpu8bu33ohspqot.png
+Dominick Puni,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/tq1snozjpjrgrjoflrfg.png
+Brock Purdy,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/wt42ykvuxpngm4m1axxn.png
+Curtis Robinson,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/x3xyzgeapcafr0gicl5y.jpg
+Demarcus Robinson,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/lakf0xue1qqb7ed4p6ge
+Patrick Taylor Jr.,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/hochjncae0hqcoveuexq.jpg
+Trent Taylor,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/j8lom4fnsveujt8hykef.jpg
+Tre Tomlinson,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/n5pfv126xw0psc0d1ydz
+Jake Tonges,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/la3z5y6u7tix6rnq2m5l.jpg
+Fred Warner,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/zo4ftfar4bshrbipceuk.jpg
+Jon Weeks,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/d9fvm74pu4vyinveopbf
+DaShaun White,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/mjnpmkw3ar6zcj2hxxzd
+Trent Williams,https://static.clubs.nfl.com/image/private/t_thumb_squared_2x/f_auto/49ers/bnq8i5urjualxre5caqz.jpg
+Brayden Willis,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/xmo7hsuho3ehmsjwvthc.jpg
+Dee Winters,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/ggf13riajo0kn0y6kbu0.jpg
+Mitch Wishnowsky,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/mkf1xr1x8nr9l55oq72a.jpg
+Nick Zakelj,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/o92tva22zezdz4aksadl.jpg

data/april_11_multimedia_data_collect/niners_players_headshots_with_socials_merged.csv

@@ -0,0 +1,74 @@
+name,headshot_url,instagram_url,
+Israel Abanikanda,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/wo7d9oli06eki4mnh3i8.png,https://www.instagram.com/izzygetsbusy__/?hl=en,
+Brandon Aiyuk,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/khwofxjjwx0hcaigzxhw.png,https://www.instagram.com/brandonaiyuk/?hl=en,
+Isaac Alarcon,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/mlhuuxukyusodzlfsmnv.jpg,https://www.instagram.com/isaac_algar/?hl=en,
+Evan Anderson,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/ng7oamywxvqgkx6l6kqc.png,https://www.instagram.com/klamps8/?hl=en,
+Tre Avery,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/a7kfv7xjftqlaqghk6sg,https://www.instagram.com/t.avery21/?hl=en,
+Alex Barrett,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/bm0ay22de39d1enrxwiq.jpg,https://www.instagram.com/alex.barrett/?hl=en,
+Ben Bartch,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/aqaslodzr7y0yvh5zzxa.jpg,https://www.instagram.com/bartchben/,
+Robert Beal Jr.,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/jwwhmt5d8mi0vdb8nfic.jpg,https://www.instagram.com/oursf49ers/reel/C_CVQxxp2ti/,
+Tatum Bethune,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/vl08pinqpmoubdf0zy5s.png,https://www.instagram.com/tatumx15/?hl=en,
+Nick Bosa,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/utiwswqvpkiwtocijwhz.jpg,https://www.instagram.com/nbsmallerbear/?hl=en,
+Jake Brendel,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/svsb41aekpzt3m9snilw.jpg,https://www.instagram.com/jake.brendel/?hl=en,
+Ji'Ayir Brown,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/urillpic02z774n09xvf.jpg,https://www.instagram.com/_tiig/?hl=en,
+Tre Brown,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/dpemqrrweakt8dci3qfb,https://www.instagram.com/tre_brown25/?hl=en,
+Spencer Burford,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/lje3ae25dntkdudp6eex.jpg,https://www.instagram.com/spence__74/?hl=en,
+Jacob Cowing,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/lg7aao0umc21oioufqdx.png,https://www.instagram.com/jaycowing_/?hl=en,
+Kalia Davis,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/rmnxj3sh7pyldmcxqe32.jpg,https://www.instagram.com/ucf.football/p/C3No6rTugDe/,
+Jordan Elliott,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/xbyky8r2yuzusd2tmrw8.jpg,https://www.instagram.com/jordanelliott_nbcs/,
+Luke Farrell,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/f2z7wpmx7ngtxcqqedla,https://www.instagram.com/lukefarrell89/?hl=en,
+Russell Gage Jr.,https://static.www.nfl.com/image/private/t_thumb_squared_2x/f_auto/league/lkqhshv0dss1b9c6mdnj,https://www.instagram.com/w8k3mupruss/?hl=en,
+Jonathan Garvin,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/rapfcxut6vu50vcevswe.png,https://www.instagram.com/thesfniners/p/DCmgF8KSw2A/?hl=en,
+Luke Gifford,https://static.www.nfl.com/image/private/t_thumb_squared_2x/f_auto/league/mhdbbzj8amttnpd1nbpn,https://www.instagram.com/luke_gifford/?hl=en,
+Kevin Givens,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/mstmgft0e0ancdzspboy.jpg,https://www.instagram.com/49ers/p/DAg_Pvpz1vV/,
+Jalen Graham,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/pbl2a1ujopvwqrfct0jp.jpg,https://www.instagram.com/thexniners/p/CruR8IPrSV7/,
+Richie Grant,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/szeswtvt6jmbu3so3phd,https://www.instagram.com/richiegrant_/?hl=en,
+Renardo Green,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/v79obx9v7tgcjjlo6hiy.png,https://www.instagram.com/dondada.8/?hl=en,
+Yetur Gross-Matos,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/etuaajmvhbc5qkebgoow.jpg,https://www.instagram.com/__lobo99/?hl=en,
+Isaac Guerendo,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/b66rpzr9iauo5rdprvka.png,https://www.instagram.com/isaac_guerendo/?hl=en,
+Sebastian Gutierrez,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/am9sywgkga6jq65hvboe.jpg,https://www.instagram.com/sebastiandev1/?hl=en,
+Matt Hennessy,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/zk8b21o8ncxnyu0gyf23,https://www.instagram.com/matt___hennessy/?hl=en,
+Isaiah Hodgins,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/ax1oft9kqida0eokvtes.jpg,https://www.instagram.com/isaiahhodgins/?hl=en,
+Drake Jackson,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/y2luyplzpvbzokyfbmla.jpg,https://www.instagram.com/thefreak/?hl=en,
+Tarron Jackson,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/pnqrjp76bgpkmacxma3r,https://www.instagram.com/tarron_jackson/?hl=en,
+Jauan Jennings,https://static.clubs.nfl.com/image/private/t_thumb_squared_2x/f_auto/49ers/wxsq7f4ajmhfs6tn4dg2.jpg,https://www.instagram.com/u_aintjj/?hl=en,
+Quindell Johnson,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/uga90lawcfxjcqna7opb,https://www.instagram.com/p/DFGnwNlymc9/,
+Zack Johnson,https://static.www.nfl.com/image/private/t_thumb_squared_2x/f_auto/league/n4hy8uzhcl5cl0ricwoa,https://www.instagram.com/zack.johnson.68/,
+Mac Jones,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/pedpdxybeus7mrovsoko,https://www.instagram.com/macjones_10/?hl=en,
+Kyle Juszczyk,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/ywdz6y2pfzndqgmxxfbj.jpg,https://www.instagram.com/juicecheck44/?hl=en,
+George Kittle,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/elheepobwn1ahqwtfwat.jpg,https://www.instagram.com/gkittle/?hl=en,
+Deommodore Lenoir,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/f9fnuvbpcxku9ibt9qs8.jpg,https://www.instagram.com/deommo.lenoir/?hl=en,
+Chase Lucas,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/gjeejt5pbagnipodhdz4.jpg,https://www.instagram.com/chase_lucas24/?hl=en,
+Darrell Luter Jr.,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/g5rohvooet9g5w7rlhrh.jpg,https://www.instagram.com/_d.ray4k/?hl=en,
+Jaylen Mahoney,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/yv9inbia05nyxppuajv0.png,https://www.instagram.com/jaylenmahoney_/,
+Christian McCaffrey,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/a8fka6shomakkbllljgt.jpg,https://www.instagram.com/christianmccaffrey/?hl=en,
+Jalen McKenzie,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/gffxpns1ayxyjccymr6d.jpg,https://www.instagram.com/jay_peez70/?hl=en,
+Colton McKivitz,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/jugvoxjabgsbcfbuqfew.jpg,https://www.instagram.com/cmckivitz53/?hl=en,
+Jake Moody,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/ygputwsbutemszr8xxkw.jpg,https://www.instagram.com/jmoods_/?hl=en,
+Tanner Mordecai,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/y8gipodnkeapgmegnxs1.png,https://www.instagram.com/t_mordecai/?hl=en,
+Malik Mustapha,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/eyrgxgpbrycd9x8glk0j.png,https://www.instagram.com/stapha/,
+Siran Neal,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/muhthfs6owkkpsyop1e6,https://www.instagram.com/siranneal/?hl=en,
+Drake Nugent,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/qyb4kurtbv9uflmupfnc.png,https://www.instagram.com/drakenugent9/?hl=en,
+George Odum,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/sqpxhoycdpegkyjn6ooc.jpg,https://www.instagram.com/george.w.odum/?hl=en,
+Sam Okuayinonu,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/fyolr2zk2nplfbdze75l.jpg,https://www.instagram.com/sam.ok97/?hl=en,
+Terique Owens,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/okhin0uwdon2nimvbtwd.png,https://www.instagram.com/terique_owens/?hl=en,
+Ricky Pearsall,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/to7q7w4kjiajseb4ljcx.png,https://www.instagram.com/ricky.pearsall/?hl=en,
+Jason Pinnock,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/on29awacb9frijyggtgt,https://www.instagram.com/jpinny15/?hl=en,
+Austen Pleasants,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/wsbs5emdyzuc1sudbcls.png,https://www.instagram.com/oursf49ers/p/DDr48a4PdcO/?hl=en,
+Mason Pline,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/mvjlaxpu8bu33ohspqot.png,https://www.instagram.com/mpline12/?hl=en,
+Dominick Puni,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/tq1snozjpjrgrjoflrfg.png,https://www.instagram.com/dompuni/?hl=en,
+Brock Purdy,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/wt42ykvuxpngm4m1axxn.png,https://www.instagram.com/brock.purdy13/?hl=en,
+Curtis Robinson,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/x3xyzgeapcafr0gicl5y.jpg,https://www.instagram.com/curtis_robinsonn/?hl=en,
+Demarcus Robinson,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/lakf0xue1qqb7ed4p6ge,https://www.instagram.com/demarcusrobinson/?hl=en,
+Patrick Taylor Jr.,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/hochjncae0hqcoveuexq.jpg,https://www.instagram.com/patricktaylor/?hl=en,
+Trent Taylor,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/j8lom4fnsveujt8hykef.jpg,https://www.instagram.com/trent5taylor/?hl=en,
+Tre Tomlinson,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/n5pfv126xw0psc0d1ydz,https://www.instagram.com/trevius/?hl=en,
+Jake Tonges,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/la3z5y6u7tix6rnq2m5l.jpg,https://www.instagram.com/jaketonges/?hl=en,
+Fred Warner,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/zo4ftfar4bshrbipceuk.jpg,https://www.instagram.com/fred_warner/?hl=en,
+Jon Weeks,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/d9fvm74pu4vyinveopbf,https://www.instagram.com/jonweeks46/?hl=en,
+DaShaun White,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/mjnpmkw3ar6zcj2hxxzd,https://www.instagram.com/demoeto/?hl=en,
+Trent Williams,https://static.clubs.nfl.com/image/private/t_thumb_squared_2x/f_auto/49ers/bnq8i5urjualxre5caqz.jpg,https://www.instagram.com/trentwilliams71/?hl=en,
+Brayden Willis,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/xmo7hsuho3ehmsjwvthc.jpg,https://www.instagram.com/brayden_willis/?hl=en,
+Dee Winters,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/ggf13riajo0kn0y6kbu0.jpg,https://www.instagram.com/dwints_/?hl=en,
+Mitch Wishnowsky,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/mkf1xr1x8nr9l55oq72a.jpg,https://www.instagram.com/mitchwish3/?hl=en,
+Nick Zakelj,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/o92tva22zezdz4aksadl.jpg,https://www.instagram.com/nickzakelj/?hl=en,

data/april_11_multimedia_data_collect/player_headshots.py

@@ -0,0 +1,73 @@
+import requests
+from bs4 import BeautifulSoup
+import csv
+
+ROSTER_URL = "https://www.49ers.com/team/players-roster/"
+
+def scrape_49ers_roster(output_csv='niners_players_headshots.csv'):
+    """
+    Scrapes the 49ers roster page for player data and saves to CSV.
+    Extracts:
+        - Name
+        - Headshot Image URL
+    """
+    response = requests.get(ROSTER_URL)
+    response.raise_for_status()
+    soup = BeautifulSoup(response.text, 'html.parser')
+
+    player_rows = soup.select('div.d3-o-table--horizontal-scroll tbody tr')
+    if not player_rows:
+        raise ValueError("No player rows found. The page structure may have changed.")
+
+    roster_data = []
+    for row in player_rows:
+        try:
+            # Extract player name and headshot
+            player_cell = row.find('td')
+            name_tag = player_cell.select_one('.nfl-o-roster__player-name')
+            name = name_tag.get_text(strip=True) if name_tag else ""
+
+            img_tag = player_cell.find('img')
+            headshot_url = img_tag['src'] if img_tag and img_tag.get('src') else ""
+            
+            # Fix the URL by replacing t_lazy with t_thumb_squared_2x
+            if headshot_url:
+                headshot_url = headshot_url.replace('/t_thumb_squared/t_lazy/', '/t_thumb_squared_2x/')
+
+            # Other stats (in order of table columns)
+            # cells = row.find_all('td')
+            # jersey_number = cells[1].get_text(strip=True) if len(cells) > 1 else ""
+            # position = cells[2].get_text(strip=True) if len(cells) > 2 else ""
+            # height = cells[3].get_text(strip=True) if len(cells) > 3 else ""
+            # weight = cells[4].get_text(strip=True) if len(cells) > 4 else ""
+            # age = cells[5].get_text(strip=True) if len(cells) > 5 else ""
+            # experience = cells[6].get_text(strip=True) if len(cells) > 6 else ""
+            # college = cells[7].get_text(strip=True) if len(cells) > 7 else ""
+
+            roster_data.append({
+                'name': name,
+                # 'jersey_number': jersey_number,
+                # 'position': position,
+                # 'height': height,
+                # 'weight': weight,
+                # 'age': age,
+                # 'experience': experience,
+                # 'college': college,
+                'headshot_url': headshot_url
+            })
+
+        except Exception as e:
+            print(f"[WARNING] Skipping row due to error: {e}")
+            continue
+
+    # Save to CSV
+    with open(output_csv, 'w', newline='', encoding='utf-8') as f:
+        fieldnames = ['name', 'headshot_url']
+        writer = csv.DictWriter(f, fieldnames=fieldnames)
+        writer.writeheader()
+        writer.writerows(roster_data)
+
+    print(f"[INFO] Successfully saved {len(roster_data)} players to '{output_csv}'.")
+
+if __name__ == "__main__":
+    scrape_49ers_roster()

data/april_11_multimedia_data_collect/z_old/niners_players_headshots_with_socials.csv

@@ -0,0 +1,24 @@
+name,headshot_url,instagram_url
+George Odum,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/sqpxhoycdpegkyjn6ooc.jpg,https://www.instagram.com/george.w.odum/?hl=en
+Sam Okuayinonu,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/fyolr2zk2nplfbdze75l.jpg,https://www.instagram.com/sam.ok97/?hl=en
+Terique Owens,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/okhin0uwdon2nimvbtwd.png,https://www.instagram.com/terique_owens/?hl=en
+Ricky Pearsall,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/to7q7w4kjiajseb4ljcx.png,https://www.instagram.com/ricky.pearsall/?hl=en
+Jason Pinnock,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/on29awacb9frijyggtgt,https://www.instagram.com/jpinny15/?hl=en
+Austen Pleasants,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/wsbs5emdyzuc1sudbcls.png,https://www.instagram.com/oursf49ers/p/DDr48a4PdcO/?hl=en
+Mason Pline,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/mvjlaxpu8bu33ohspqot.png,https://www.instagram.com/mpline12/?hl=en
+Dominick Puni,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/tq1snozjpjrgrjoflrfg.png,https://www.instagram.com/dompuni/?hl=en
+Brock Purdy,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/wt42ykvuxpngm4m1axxn.png,https://www.instagram.com/brock.purdy13/?hl=en
+Curtis Robinson,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/x3xyzgeapcafr0gicl5y.jpg,https://www.instagram.com/curtis_robinsonn/?hl=en
+Demarcus Robinson,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/lakf0xue1qqb7ed4p6ge,https://www.instagram.com/demarcusrobinson/?hl=en
+Patrick Taylor Jr.,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/hochjncae0hqcoveuexq.jpg,https://www.instagram.com/patricktaylor/?hl=en
+Trent Taylor,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/j8lom4fnsveujt8hykef.jpg,https://www.instagram.com/trent5taylor/?hl=en
+Tre Tomlinson,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/n5pfv126xw0psc0d1ydz,https://www.instagram.com/trevius/?hl=en
+Jake Tonges,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/la3z5y6u7tix6rnq2m5l.jpg,https://www.instagram.com/jaketonges/?hl=en
+Fred Warner,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/zo4ftfar4bshrbipceuk.jpg,https://www.instagram.com/fred_warner/?hl=en
+Jon Weeks,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/d9fvm74pu4vyinveopbf,https://www.instagram.com/jonweeks46/?hl=en
+DaShaun White,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/mjnpmkw3ar6zcj2hxxzd,https://www.instagram.com/demoeto/?hl=en
+Trent Williams,https://static.clubs.nfl.com/image/private/t_thumb_squared_2x/f_auto/49ers/bnq8i5urjualxre5caqz.jpg,https://www.instagram.com/trentwilliams71/?hl=en
+Brayden Willis,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/xmo7hsuho3ehmsjwvthc.jpg,https://www.instagram.com/brayden_willis/?hl=en
+Dee Winters,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/ggf13riajo0kn0y6kbu0.jpg,https://www.instagram.com/dwints_/?hl=en
+Mitch Wishnowsky,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/mkf1xr1x8nr9l55oq72a.jpg,https://www.instagram.com/mitchwish3/?hl=en
+Nick Zakelj,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/o92tva22zezdz4aksadl.jpg,https://www.instagram.com/nickzakelj/?hl=en

data/april_11_multimedia_data_collect/z_old/niners_players_headshots_with_socials_v1_safe.csv

@@ -0,0 +1,51 @@
+name,headshot_url,instagram_url
+Israel Abanikanda,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/wo7d9oli06eki4mnh3i8.png,https://www.instagram.com/izzygetsbusy__/?hl=en
+Brandon Aiyuk,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/khwofxjjwx0hcaigzxhw.png,https://www.instagram.com/brandonaiyuk/?hl=en
+Isaac Alarcon,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/mlhuuxukyusodzlfsmnv.jpg,https://www.instagram.com/isaac_algar/?hl=en
+Evan Anderson,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/ng7oamywxvqgkx6l6kqc.png,https://www.instagram.com/klamps8/?hl=en
+Tre Avery,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/a7kfv7xjftqlaqghk6sg,https://www.instagram.com/t.avery21/?hl=en
+Alex Barrett,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/bm0ay22de39d1enrxwiq.jpg,https://www.instagram.com/alex.barrett/?hl=en
+Ben Bartch,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/aqaslodzr7y0yvh5zzxa.jpg,https://www.instagram.com/bartchben/
+Robert Beal Jr.,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/jwwhmt5d8mi0vdb8nfic.jpg,https://www.instagram.com/oursf49ers/reel/C_CVQxxp2ti/
+Tatum Bethune,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/vl08pinqpmoubdf0zy5s.png,https://www.instagram.com/tatumx15/?hl=en
+Nick Bosa,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/utiwswqvpkiwtocijwhz.jpg,https://www.instagram.com/nbsmallerbear/?hl=en
+Jake Brendel,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/svsb41aekpzt3m9snilw.jpg,https://www.instagram.com/jake.brendel/?hl=en
+Ji'Ayir Brown,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/urillpic02z774n09xvf.jpg,https://www.instagram.com/_tiig/?hl=en
+Tre Brown,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/dpemqrrweakt8dci3qfb,https://www.instagram.com/tre_brown25/?hl=en
+Spencer Burford,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/lje3ae25dntkdudp6eex.jpg,https://www.instagram.com/spence__74/?hl=en
+Jacob Cowing,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/lg7aao0umc21oioufqdx.png,https://www.instagram.com/jaycowing_/?hl=en
+Kalia Davis,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/rmnxj3sh7pyldmcxqe32.jpg,https://www.instagram.com/ucf.football/p/C3No6rTugDe/
+Jordan Elliott,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/xbyky8r2yuzusd2tmrw8.jpg,https://www.instagram.com/jordanelliott_nbcs/
+Luke Farrell,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/f2z7wpmx7ngtxcqqedla,https://www.instagram.com/lukefarrell89/?hl=en
+Russell Gage Jr.,https://static.www.nfl.com/image/private/t_thumb_squared_2x/f_auto/league/lkqhshv0dss1b9c6mdnj,https://www.instagram.com/w8k3mupruss/?hl=en
+Jonathan Garvin,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/rapfcxut6vu50vcevswe.png,https://www.instagram.com/thesfniners/p/DCmgF8KSw2A/?hl=en
+Luke Gifford,https://static.www.nfl.com/image/private/t_thumb_squared_2x/f_auto/league/mhdbbzj8amttnpd1nbpn,https://www.instagram.com/luke_gifford/?hl=en
+Kevin Givens,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/mstmgft0e0ancdzspboy.jpg,https://www.instagram.com/49ers/p/DAg_Pvpz1vV/
+Jalen Graham,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/pbl2a1ujopvwqrfct0jp.jpg,https://www.instagram.com/thexniners/p/CruR8IPrSV7/
+Richie Grant,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/szeswtvt6jmbu3so3phd,https://www.instagram.com/richiegrant_/?hl=en
+Renardo Green,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/v79obx9v7tgcjjlo6hiy.png,https://www.instagram.com/dondada.8/?hl=en
+Yetur Gross-Matos,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/etuaajmvhbc5qkebgoow.jpg,https://www.instagram.com/__lobo99/?hl=en
+Isaac Guerendo,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/b66rpzr9iauo5rdprvka.png,https://www.instagram.com/isaac_guerendo/?hl=en
+Sebastian Gutierrez,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/am9sywgkga6jq65hvboe.jpg,https://www.instagram.com/sebastiandev1/?hl=en
+Matt Hennessy,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/zk8b21o8ncxnyu0gyf23,https://www.instagram.com/matt___hennessy/?hl=en
+Isaiah Hodgins,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/ax1oft9kqida0eokvtes.jpg,https://www.instagram.com/isaiahhodgins/?hl=en
+Drake Jackson,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/y2luyplzpvbzokyfbmla.jpg,https://www.instagram.com/thefreak/?hl=en
+Tarron Jackson,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/pnqrjp76bgpkmacxma3r,https://www.instagram.com/tarron_jackson/?hl=en
+Jauan Jennings,https://static.clubs.nfl.com/image/private/t_thumb_squared_2x/f_auto/49ers/wxsq7f4ajmhfs6tn4dg2.jpg,https://www.instagram.com/u_aintjj/?hl=en
+Quindell Johnson,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/uga90lawcfxjcqna7opb,https://www.instagram.com/p/DFGnwNlymc9/
+Zack Johnson,https://static.www.nfl.com/image/private/t_thumb_squared_2x/f_auto/league/n4hy8uzhcl5cl0ricwoa,https://www.instagram.com/zack.johnson.68/
+Mac Jones,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/pedpdxybeus7mrovsoko,https://www.instagram.com/macjones_10/?hl=en
+Kyle Juszczyk,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/ywdz6y2pfzndqgmxxfbj.jpg,https://www.instagram.com/juicecheck44/?hl=en
+George Kittle,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/elheepobwn1ahqwtfwat.jpg,https://www.instagram.com/gkittle/?hl=en
+Deommodore Lenoir,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/f9fnuvbpcxku9ibt9qs8.jpg,https://www.instagram.com/deommo.lenoir/?hl=en
+Chase Lucas,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/gjeejt5pbagnipodhdz4.jpg,https://www.instagram.com/chase_lucas24/?hl=en
+Darrell Luter Jr.,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/g5rohvooet9g5w7rlhrh.jpg,https://www.instagram.com/_d.ray4k/?hl=en
+Jaylen Mahoney,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/yv9inbia05nyxppuajv0.png,https://www.instagram.com/jaylenmahoney_/
+Christian McCaffrey,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/a8fka6shomakkbllljgt.jpg,https://www.instagram.com/christianmccaffrey/?hl=en
+Jalen McKenzie,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/gffxpns1ayxyjccymr6d.jpg,https://www.instagram.com/jay_peez70/?hl=en
+Colton McKivitz,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/jugvoxjabgsbcfbuqfew.jpg,https://www.instagram.com/cmckivitz53/?hl=en
+Jake Moody,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_auto/49ers/ygputwsbutemszr8xxkw.jpg,https://www.instagram.com/jmoods_/?hl=en
+Tanner Mordecai,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/y8gipodnkeapgmegnxs1.png,https://www.instagram.com/t_mordecai/?hl=en
+Malik Mustapha,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/eyrgxgpbrycd9x8glk0j.png,https://www.instagram.com/stapha/
+Siran Neal,https://static.www.nfl.com/image/upload/t_thumb_squared_2x/f_auto/league/muhthfs6owkkpsyop1e6,https://www.instagram.com/siranneal/?hl=en
+Drake Nugent,https://static.clubs.nfl.com/image/upload/t_thumb_squared_2x/f_png/49ers/qyb4kurtbv9uflmupfnc.png,https://www.instagram.com/drakenugent9/?hl=en