Spaces:
Sleeping
Sleeping
File size: 11,553 Bytes
b0a8dea 97c3bf5 b0a8dea 97c3bf5 b0a8dea 46c25c8 97c3bf5 46c25c8 97c3bf5 b0a8dea 97c3bf5 b0a8dea 97c3bf5 b0a8dea 97c3bf5 b0a8dea 46c25c8 b0a8dea 97c3bf5 b0a8dea 97c3bf5 b0a8dea 97c3bf5 b0a8dea 46c25c8 97c3bf5 b0a8dea 97c3bf5 b0a8dea 46c25c8 97c3bf5 46c25c8 b0a8dea 46c25c8 97c3bf5 46c25c8 b0a8dea 46c25c8 97c3bf5 46c25c8 b0a8dea 97c3bf5 b0a8dea 97c3bf5 46c25c8 b0a8dea 46c25c8 b0a8dea 97c3bf5 b0a8dea 46c25c8 b0a8dea 97c3bf5 b0a8dea 46c25c8 97c3bf5 46c25c8 97c3bf5 b0a8dea 97c3bf5 b0a8dea 46c25c8 b0a8dea 46c25c8 b0a8dea 46c25c8 b0a8dea 97c3bf5 b0a8dea |
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 |
from fastapi import FastAPI, File, UploadFile, HTTPException
import os
import tempfile
import chess.pgn
import chess.engine
from enum import Enum
from typing import List, Dict
from datetime import datetime
import csv
import json
from fastapi.responses import JSONResponse
import asyncio
import sys
from routes.tex_based_review import review_chess_game, validate_json
if sys.platform == "win32":
asyncio.set_event_loop_policy(asyncio.WindowsProactorEventLoopPolicy())
class GamePhase(Enum):
OPENING = "opening"
MIDDLEGAME = "middlegame"
ENDGAME = "endgame"
class Classification(Enum):
BRILLIANT = "brilliant"
GREAT = "great"
BEST = "best"
EXCELLENT = "excellent"
GOOD = "good"
INACCURACY = "inaccuracy"
MISTAKE = "mistake"
MISS = "miss"
BLUNDER = "blunder"
BOOK = "book"
FORCED = "forced"
classification_values = {
Classification.BLUNDER: 0,
Classification.MISTAKE: 0.2,
Classification.MISS: 0.3,
Classification.INACCURACY: 0.4,
Classification.GOOD: 0.65,
Classification.EXCELLENT: 0.9,
Classification.BEST: 1,
Classification.GREAT: 1,
Classification.BRILLIANT: 1,
Classification.BOOK: 1,
Classification.FORCED: 1,
}
centipawn_classifications = [
Classification.BEST,
Classification.EXCELLENT,
Classification.GOOD,
Classification.INACCURACY,
Classification.MISS,
Classification.MISTAKE,
Classification.BLUNDER,
]
# Analysis parameters
FORCED_WIN_THRESHOLD = 500
MISS_CENTIPAWN_LOSS = 300
MISS_MATE_THRESHOLD = 3
ENDGAME_MATERIAL_THRESHOLD = 24
QUEEN_VALUE = 9
def detect_game_phase(board: chess.Board, in_opening: bool) -> GamePhase:
if in_opening:
return GamePhase.OPENING
total_material = 0
queens = 0
for color in [chess.WHITE, chess.BLACK]:
for piece_type in chess.PIECE_TYPES:
if piece_type == chess.KING:
continue
count = len(board.pieces(piece_type, color))
value = {
chess.PAWN: 1,
chess.KNIGHT: 3,
chess.BISHOP: 3,
chess.ROOK: 5,
chess.QUEEN: QUEEN_VALUE
}[piece_type]
total_material += count * value
if piece_type == chess.QUEEN:
queens += count
endgame_conditions = [
total_material <= ENDGAME_MATERIAL_THRESHOLD,
queens == 0 and total_material <= ENDGAME_MATERIAL_THRESHOLD * 2,
]
return GamePhase.ENDGAME if any(endgame_conditions) else GamePhase.MIDDLEGAME
def get_evaluation_loss_threshold(classif: Classification, prev_eval: float) -> float:
prev_eval = abs(prev_eval)
if classif == Classification.BEST:
return max(0.0001 * prev_eval**2 + 0.0236 * prev_eval - 3.7143, 0)
elif classif == Classification.EXCELLENT:
return max(0.0002 * prev_eval**2 + 0.1231 * prev_eval + 27.5455, 0)
elif classif == Classification.GOOD:
return max(0.0002 * prev_eval**2 + 0.2643 * prev_eval + 60.5455, 0)
elif classif == Classification.INACCURACY:
return max(0.0002 * prev_eval**2 + 0.3624 * prev_eval + 108.0909, 0)
elif classif == Classification.MISS:
return max(0.00025 * prev_eval**2 + 0.38255 * prev_eval + 166.9541, 0)
elif classif == Classification.MISTAKE:
return max(0.0003 * prev_eval**2 + 0.4027 * prev_eval + 225.8182, 0)
else:
return float("inf")
def load_opening_book(csv_path):
opening_book = {}
try:
with open(csv_path, newline='', encoding='utf-8') as csvfile:
reader = csv.reader(csvfile)
next(reader)
for row in reader:
if len(row) < 3:
continue
pgn_moves = row[2]
game = chess.pgn.Game()
board = game.board()
for move in pgn_moves.split():
if "." in move:
continue
try:
chess_move = board.push_san(move)
fen = " ".join(board.fen().split()[:4])
opening_book[fen] = chess_move.uci()
except ValueError:
break
except Exception as e:
print(f"Error loading opening book: {e}")
return opening_book
def is_book_move(board, opening_book, max_depth=8):
if board.fullmove_number > max_depth:
return None
fen = " ".join(board.fen().split()[:4])
return opening_book.get(fen)
engine_path = os.path.join(os.getcwd(), "models", "stockfish-windows-x86-64-avx2.exe")
book_csv_path = os.path.join(os.getcwd(), "assets", "openings_master.csv")
def analyze_pgn(pgn_file: str) -> Dict:
opening_book = load_opening_book(book_csv_path)
text_based_result = review_chess_game(pgn_file)
with open(pgn_file) as pgn:
game = chess.pgn.read_game(pgn)
if not game:
return {"error": "No game found in the PGN file."}
result = {
"move_analysis": [],
"phase_analysis": {},
"player_summaries": {},
"test_based_review": text_based_result
}
with chess.engine.SimpleEngine.popen_uci(engine_path) as engine:
board = game.board()
classifications = {
"white": {phase: [] for phase in GamePhase},
"black": {phase: [] for phase in GamePhase}
}
phase_data = {phase: [] for phase in GamePhase}
in_opening = True
for move_number, node in enumerate(game.mainline(), start=1):
# Analyze position before the move
pre_info = engine.analyse(board, chess.engine.Limit(time=0.3), multipv=3)[0]
pre_eval = pre_info["score"].white().score(mate_score=10000) or 0
pre_pv_moves = pre_info.get("pv", [])
# Get best move and follow-up moves in UCI notation
best_move_pre = pre_pv_moves[0].uci() if pre_pv_moves else None
follow_up_pre = [m.uci() for m in pre_pv_moves[:min(len(pre_pv_moves), 5)]]
# Make the user move
move = node.move
board.push(move) # Update the board state
# Analyze position after the move
post_info = engine.analyse(board, chess.engine.Limit(time=0.3), multipv=3)[0]
# Get best move and follow-up moves AFTER move is played (in UCI notation)
post_pv_moves = post_info.get("pv", [])
best_move_post = post_pv_moves[0].uci() if post_pv_moves else None
follow_up_post = [m.uci() for m in post_pv_moves[:min(len(post_pv_moves), 5)]]
post_eval = post_info["score"].white().score(mate_score=10000) or 0
# Determine game phase
book_move = is_book_move(board, opening_book)
current_phase = detect_game_phase(board, in_opening)
if not book_move and in_opening:
in_opening = False
# Calculate evaluation loss
eval_loss = abs(pre_eval - post_eval)
# Initial classification
classification = Classification.BOOK if book_move else None
if not classification:
for classif in centipawn_classifications:
threshold = get_evaluation_loss_threshold(classif, pre_eval)
if eval_loss <= threshold:
classification = classif
break
classification = classification or Classification.BLUNDER
# Check for missed opportunities
is_winning = abs(pre_eval) >= FORCED_WIN_THRESHOLD
is_forced_win = pre_info["score"].is_mate() and pre_info["score"].relative.mate() <= MISS_MATE_THRESHOLD
if is_winning and move != best_move_pre and (eval_loss >= MISS_CENTIPAWN_LOSS or is_forced_win):
classification = Classification.MISS
# Check for brilliant moves
if classification == Classification.BEST:
if pre_eval < -150 and post_eval >= 150:
classification = Classification.GREAT
elif pre_eval < -300 and post_eval >= 300:
classification = Classification.BRILLIANT
# Track classifications
player = "white" if board.turn == chess.BLACK else "black"
classifications[player][current_phase].append(classification)
phase_data[current_phase].append(classification)
# Add move analysis to result (using UCI notation)
result["move_analysis"].append({
"move_number": move_number,
"player": "White" if board.turn == chess.BLACK else "Black",
"user_move": move.uci(),
"evaluation": post_eval / 100,
"evaluation_loss": eval_loss / 100,
"classification": classification.value,
"best_move_pre": best_move_pre, # Best move BEFORE move is played (UCI)
"follow_up_pre": follow_up_pre, # Follow-up moves BEFORE move is played (UCI)
"best_move_post": best_move_post, # Best move AFTER move is played (UCI)
"follow_up_post": follow_up_post # Follow-up moves AFTER move is played (UCI)
})
# Phase analysis
for phase in GamePhase:
moves = phase_data[phase]
if moves:
rating = get_phase_rating(moves)
result["phase_analysis"][phase.value] = {
"rating": rating.value,
"move_count": len(moves)
}
# Player summaries
for color in ["white", "black"]:
player = game.headers["White" if color == "white" else "Black"]
counts = {c.value: 0 for c in Classification}
for phase in GamePhase:
phase_moves = classifications[color][phase]
for m in phase_moves:
m_enum = Classification(m) if isinstance(m, str) else m # Convert if needed
counts[m_enum.value] += 1
result["player_summaries"][player] = counts
def convert_enums(obj):
if isinstance(obj, Enum): # Convert Enum to its value
return obj.value
if isinstance(obj, dict): # Recursively handle dicts
return {k: convert_enums(v) for k, v in obj.items()}
if isinstance(obj, list): # Recursively handle lists
return [convert_enums(i) for i in obj]
return obj # Return other types as they are
json_result = convert_enums(result)
return JSONResponse(content=json_result)
def get_phase_rating(classified_moves: List[Classification]) -> Classification:
if not classified_moves:
return Classification.GOOD
classified_moves = [Classification(m) if isinstance(m, str) else m for m in classified_moves]
total = sum(classification_values[m] for m in classified_moves)
average = total / len(classified_moves)
rating_order = [
(Classification.BRILLIANT, 0.95),
(Classification.GREAT, 0.85),
(Classification.BEST, 0.75),
(Classification.EXCELLENT, 0.65),
(Classification.GOOD, 0.5),
(Classification.INACCURACY, 0.35),
(Classification.MISS, 0.25),
(Classification.MISTAKE, 0.15)
]
return next((c for c, t in rating_order if average >= t), Classification.BLUNDER) |