Spaces:
Sleeping
Sleeping
File size: 19,838 Bytes
36cc9c0 abcb943 944060d 36cc9c0 944060d abcb943 291e330 284ad13 36cc9c0 291e330 24265ea 291e330 24265ea 291e330 3224c57 291e330 78c8d84 291e330 78c8d84 291e330 944060d 291e330 78c8d84 291e330 97a08ff 291e330 11400bb abcb943 bea7945 97a08ff 36cc9c0 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 |
from Recent_match_scrapper import get_matches_stats
import os
import pandas as pd
import numpy as np
from Meta_scrapper import *
from helper import merge_stats, process_kda_perfect, ChampionConverter
from Player_scrapper import get_player_stats
from Weekly_meta_scrapper import *
import pandas as pd
import re
# ============================================ my functions =========================================================
def create_champion_features_and_return_df(merged_player_stats=None, meta_stats=None, weekly_meta=None, debug=None, consider_team_comp=True, test_mode=False):
"""
Create features for champion prediction using player data.
Champion names will be used as column headers.
Uses pd.concat to avoid DataFrame fragmentation.
"""
try:
if merged_player_stats is None:
print("Loading merged player stats...")
input_file = os.path.join("util", "data", "player_stats_merged.csv")
merged_player_stats = pd.read_csv(input_file, low_memory=False)
#processing kda value
merged_player_stats = process_kda_perfect(merged_player_stats)
if test_mode:
print("Test mode: Using only first 100 rows")
merged_player_stats = merged_player_stats.head(100)
if meta_stats is None:
print("Loading meta stats...")
meta_file = os.path.join("util", "data", "meta_stats.csv")
meta_stats = pd.read_csv(meta_file, low_memory=False)
if weekly_meta is None:
print("Loading weekly meta stats...")
weekly_file = os.path.join("util", "data", "weekly_meta_stats.csv")
weekly_meta = pd.read_csv(weekly_file, low_memory=False)
# Initialize variables
debug_data = []
original_columns = merged_player_stats.columns.tolist()
feature_dict = {}
# Copy original columns
for col in merged_player_stats.columns:
feature_dict[col] = merged_player_stats[col].values.copy()
# Initialize the champion converter
converter = ChampionConverter()
all_champions = converter.champions
#total_champions = len(converter.champions)
# Get low tier champions and counter information
tier_penalties = {3: 0.9, 4: 0.85, 5: 0.8}
# Create tier_map as a dictionary of lists
tier_map = {}
for _, row in meta_stats.iterrows():
champ = row['champion']
tier = row['tier']
if pd.notna(tier):
if champ in tier_map:
tier_map[champ].append(tier)
else:
tier_map[champ] = [tier]
counter_map = {}
for _, row in meta_stats.iterrows():
if pd.notna(row['counter1']):
champ = row['champion']
counters = [row['counter1'], row['counter2'], row['counter3']]
if champ in counter_map:
counter_map[champ].extend([c for c in counters if pd.notna(c)])
else:
counter_map[champ] = [c for c in counters if pd.notna(c)]
# Ensure unique counters and remove duplicates
for champ, counters in counter_map.items():
counter_map[champ] = list(set(counters))
# Move 'champion' column to the first position
cols = ['champion'] + [col for col in merged_player_stats if col != 'champion']
merged_player_stats = merged_player_stats[cols]
# Define importance weights
weights = {
'recent': 0.3, # Last 20 games
'weekly': 0.4, # Last 7 days
'meta': 0.2, # Only from weekly_stats
'season': 0.06, # Current season
'mastery': 0.04 # All-time mastery
}
# Process rows in batches
batch_size = 100
total_rows = len(merged_player_stats)
print(f"Total rows: {total_rows}")
for batch_start in range(0, total_rows, batch_size):
batch_end = min(batch_start + batch_size, total_rows)
batch_rows = merged_player_stats.iloc[batch_start:batch_end]
print(f"\nProcessing rows {batch_start} to {batch_end} ({batch_start/total_rows*100:.2f}% complete)")
# Initialize batch scores dictionary
batch_scores = {champion: np.zeros(len(batch_rows)) for champion in all_champions}
# Process each row in this batch
for batch_idx, (idx, row) in enumerate(batch_rows.iterrows()):
# Process each champion for this row
for champion in all_champions:
# Initialize scores for this champion and row
champion_scores = {
'recent_score': 0,
'weekly_score': 0,
'meta_score': 0,
'season_score': 0,
'mastery_score': 0
}
# Store debug info if needed
base_score_before_penalty = 0
counter_penalty = 0
counter_debug = []
# 1. Recent Performance
for i in range(1, 4):
if row.get(f'most_champ_{i}') == champion:
wr = float(row[f'WR_{i}']) if pd.notna(row[f'WR_{i}']) else 0
kda = float(row[f'KDA_{i}']) if pd.notna(row[f'KDA_{i}']) else 0
wins = float(row[f'W_{i}']) if pd.notna(row[f'W_{i}']) else 0
losses = float(row[f'L_{i}']) if pd.notna(row[f'L_{i}']) else 0
games = wins + losses
total_games = float(row['total_games']) if pd.notna(row['total_games']) else 20
performance_quality = (
(wr * 0.7) +
(min(kda, 10) / 10 * 0.3)
)
games_factor = min(games / 5, 1.0)
games_ratio = games / total_games
if games >= 5:
if performance_quality < 0.4:
performance_quality *= 0.8
elif performance_quality > 0.7:
performance_quality *= 1.2
champion_scores['recent_score'] = (
performance_quality * (0.7 + (0.3 * games_factor))
) * (1 + games_ratio * 0.2)
break # Exit loop once found
# 2. Weekly Performance
for i in range(1, 4):
if row.get(f'7d_champ_{i}') == champion:
weekly_wins = float(row[f'7d_W_{i}']) if pd.notna(row[f'7d_W_{i}']) else 0
weekly_losses = float(row[f'7d_L_{i}']) if pd.notna(row[f'7d_L_{i}']) else 0
weekly_games = float(row[f'7d_total_{i}']) if pd.notna(row[f'7d_total_{i}']) else 0
weekly_wr = float(row[f'7d_WR_{i}']) if pd.notna(row[f'7d_WR_{i}']) else 0
profile_wr = float(row['win_rate']) if pd.notna(row['win_rate']) else 0.5
if weekly_games > 0:
wr_trend = (weekly_wr - profile_wr) / profile_wr if profile_wr > 0 else 0
weekly_intensity = min(weekly_games / 10, 1.0)
win_ratio = weekly_wins / weekly_games if weekly_games > 0 else 0
weekly_performance = (
(weekly_wr * 0.4) +
(max(min(wr_trend, 1), -1) * 0.2) +
(weekly_intensity * 0.2) +
(win_ratio * 0.2)
)
if weekly_games >= 5:
if weekly_performance < 0.4:
weekly_performance *= 0.8
elif weekly_performance > 0.7:
weekly_performance *= 1.2
champion_scores['weekly_score'] = weekly_performance * (
0.7 + (0.3 * min(weekly_games / 5, 1.0))
)
break # Exit loop once found
# 3. Meta Score
if champion in weekly_meta['champion'].values:
weekly_row = weekly_meta[weekly_meta['champion'] == champion].iloc[0]
rank = weekly_row['rank']
games = weekly_row['games']
pick_rate = weekly_row['pick']
ban_rate = weekly_row['ban']
weight = (
1 / rank * 0.5 +
games / 100 * 0.3 +
pick_rate * 0.1 -
ban_rate * 0.1
)
champion_scores['meta_score'] = weight
# 4. Season Performance
for i in range(1, 8):
if row.get(f'season_champ_{i}') == champion:
wr = float(row[f'wr_ssn_{i}']) if pd.notna(row[f'wr_ssn_{i}']) else 0
games = float(row[f'games_ssn_{i}']) if pd.notna(row[f'games_ssn_{i}']) else 0
kda = float(row[f'kda_ssn_{i}']) if pd.notna(row[f'kda_ssn_{i}']) else 0
champion_scores['season_score'] = (
wr * 0.7 +
(kda / 10) * 0.3
) * (games / 100)
break # Exit loop once found
# 5. Mastery Score
for i in range(1, 17):
if row.get(f'mastery_champ_{i}') == champion:
mastery = float(row[f'm_lv_{i}']) if pd.notna(row[f'm_lv_{i}']) else 0
champion_scores['mastery_score'] = mastery / 7
break # Exit loop once found
# Calculate base score for this champion and row
base_score = (
champion_scores['recent_score'] * weights['recent'] +
champion_scores['weekly_score'] * weights['weekly'] +
champion_scores['meta_score'] * weights['meta'] +
champion_scores['season_score'] * weights['season'] +
champion_scores['mastery_score'] * weights['mastery']
)
# Store the pre-penalty score for debugging
base_score_before_penalty = base_score
# Apply tier penalties
if champion in tier_map:
highest_tier = min(tier_map[champion])
if highest_tier in tier_penalties:
base_score *= tier_penalties[highest_tier]
# Process team composition and counter penalties
if consider_team_comp:
# Check team champions
for i in range(1, 5):
team_col = f'team_champ{i}'
if team_col in row and pd.notna(row[team_col]):
if row[team_col] == champion:
base_score = 0
break
# Only check opponents if base_score isn't already 0
if base_score != 0:
counter_penalty = 0
counter_debug = [] # For debug information
for i in range(1, 6):
opp_col = f'opp_champ{i}'
if opp_col in row and pd.notna(row[opp_col]):
opp_champ = row[opp_col]
if opp_champ == champion:
base_score = 0
break
if champion in counter_map and opp_champ in counter_map[champion]:
counter_penalty += 0.1
counter_debug.append(opp_champ)
if counter_penalty > 0:
base_score = base_score * (1 - counter_penalty)
# Store the final score for this champion and row
batch_scores[champion][batch_idx] = max(base_score, 0)
# Collect debug data if this is the debug champion
if debug == champion:
counter_list = []
for i in range(1, 6):
opp_col = f'opp_champ{i}'
if opp_col in row and pd.notna(row[opp_col]):
if champion in counter_map and row[opp_col] in counter_map[champion]:
counter_list.append(row[opp_col])
debug_row = {
'champion': row['champion'],
'recent_score': champion_scores['recent_score'],
'weekly_score': champion_scores['weekly_score'],
'meta_score': champion_scores['meta_score'],
'base_score': base_score_before_penalty,
'final_score': base_score,
'counter_penalty': counter_penalty if consider_team_comp else 0,
'final_score_actual': feature_dict[row['champion']][idx] if row['champion'] in feature_dict else base_score,
'counter_list_debug': counter_list
}
debug_data.append(debug_row)
# Update feature_dict with batch results
for champion in batch_scores:
if champion not in feature_dict:
feature_dict[champion] = np.zeros(total_rows)
feature_dict[champion][batch_start:batch_end] = batch_scores[champion]
# Save after each batch with timestamp
temp_df = pd.DataFrame({
**{col: feature_dict[col] for col in original_columns}, # Original columns first
**{champion: feature_dict[champion] for champion in all_champions} # Then champion columns
})
batch_save_file = os.path.join("util", "data", f"feature_eng_stats.csv")
temp_df.to_csv(batch_save_file, index=False)
print(f"Saved batch progress to {batch_save_file}")
if debug:
print(f"{debug} is countered by: {counter_map[debug]}")
# Process debug data if any
if debug:
debug_df = pd.DataFrame(debug_data)
print("\nDebug Data:")
print(debug_df)
# Create final DataFrame
champion_features = pd.DataFrame(feature_dict)
# Create the final DataFrame by combining original data with new features
features = pd.concat([
merged_player_stats[original_columns], # Keep all original columns
champion_features[[col for col in champion_features.columns if col not in original_columns]] # Only new champion columns
], axis=1)
# Move the champion column to be the first column
if 'champion' in features.columns:
columns = ['champion'] + [col for col in features.columns if col != 'champion']
features = features[columns]
# Print confirmation message
print(f"Saved features in data frame.")
return features
except Exception as e:
print(f"\nError occurred: {str(e)}")
return None
def create_app_user_training_df(url):
try:
#meta_stats = get_meta_stats()
#weekly_meta_stats = get_weekly_meta()
# Input validation
if not url or not isinstance(url, str):
raise ValueError("Invalid URL provided")
# Extract region and username
match = re.search(r"/summoners/(\w+)/([\w\-]+)", url)
if not match:
raise ValueError(f"Could not parse region and username from URL: {url}\n Type(url): {type(url)}")
region = match.group(1)
username = match.group(2)
print(f"Extracted - Region: {region}, Username: {username}")
# Get recent stats
print("Fetching recent matches...")
recent_stats = get_matches_stats(region, username)
# Validate recent_stats
if recent_stats is None or recent_stats.empty:
raise ValueError("recent_stats is empty. type(recent_stats): ", type(recent_stats) , " recent_stats: \n", recent_stats)
#print("Recent matches columns:", recent_stats.columns.tolist())
# Process player_id
recent_stats['player_id'] = recent_stats['player_id'].str.replace(" #", "-", regex=False)
#print("Processed player_ids:", recent_stats['player_id'].head())
# Get player stats
print("Fetching player stats...")
player_stats = get_player_stats(region, username)
# Merge stats
print("Merging stats...")
merged_stats = merge_stats(recent_stats, player_stats)
# Validate merged stats
if merged_stats is None or merged_stats.empty:
raise ValueError("Failed to merge stats")
#print("Merged stats columns:", merged_stats.columns.tolist())
# Create features
print("Creating champion features...")
training_features = create_champion_features_and_return_df(
merged_player_stats=merged_stats,
debug=None,
consider_team_comp=True,
test_mode=False
)
# Final validation
if training_features is None or training_features.empty:
raise ValueError("Failed to create training features")
print("Training features created successfully")
return training_features
except Exception as e:
import traceback
error_trace = traceback.format_exc()
print(f"Error in create_app_user_training_df:\n{error_trace}")
raise Exception(f"Failed to create training dataframe: {str(e)}")
# ========================================= end of my functions =====================================================
#url = "https://www.op.gg/summoners/euw/Agurin-EUW"
#url = "https://www.op.gg/summoners/euw/Agurin-EUW?queue_type=TOTAL"
#return_value = create_app_user_training_df(url)
#print("type(Return_value):", type(return_value), "\n return value: \n", return_value)
|