diff --git a/.gitignore b/.gitignore index 73b1d2d..11a1fc8 100644 --- a/.gitignore +++ b/.gitignore @@ -10,4 +10,5 @@ __pycache__ # ignore any CSV files written to the data directory: +data/*.csv data/*/*.csv diff --git a/app/ai/README.md b/app/ai/README.md deleted file mode 100644 index ab5ba79..0000000 --- a/app/ai/README.md +++ /dev/null @@ -1,56 +0,0 @@ - -# Tic Tac Toe (AI) - -> NOTE: this is currently a work in progress! - -The [computer players](/app/player.py) in the Tic Tac Toe app use a predefined algorithm to select the best move. Let's see if we can train machine learning models to play the game instead. - -## Generating Moves Data - -Generates moves dataset for model training and evaluation. The test/train splits are out of scope for this step, and will be done later, during model training. - -The inputs for each move are the board state (e.g. "-X-O-X-OX" and the active player (i.e. "X" or "O"). The output is the player's selected move, represented by the selected square's index position in the board notation string (0-8). - -We simulate alternating moves until we reach an outcome (win, lose, tie). After reaching an outcome, we can assign the eventual outcome value to all moves that player made leading up until the outcome: - + all winning player's moves get assigned a positive value (+1) - + all losing player's moves get assigned a negative value (-1) - + all moves resulting in a tie get a neutral score (0) - -Generating the datasets: - -```sh -GAME_COUNT=100000 X_STRATEGY="RANDOM" O_STRATEGY="RANDOM" python -m app.jobs.play_moves -``` - -Exports a CSV file in the "data" directory (e.g "/data/moves/x_random_vs_o_random_10000.csv"). Example results: - -|game_id|move_id|board_state|player|square_idx|reward| -|-------|-------|-----------|------|---------|------| -|1 |1 | `---------` |`X` |5 | 0 | -|1 |2 | `-----X---` |`O` |7 | 0 | -|1 |3 | `-----X-O-` |`X` |8 | 0 | -|1 |4 | `-----X-OX` |`O` |3 | 0 | -|1 |5 | `---O-X-OX` |`X` |1 | 0 | -|1 |6 | `-X-O-X-OX` |`O` |2 | 0 | -|1 |7 | `-XOO-X-OX` |`X` |6 | 0 | -|1 |8 | `-XOO-XXOX` |`O` |4 | 0 | -|1 |9 | `-XOOOXXOX` |`X` |0 | 0 | -|2 |1 |`---------` |`X` |1 |1 | -|2 |2 |`-X-------` |`O` |8 |-1 | -|2 |3 |`-X------O` |`X` |0 |1 | -|2 |4 |`XX------O` |`O` |4 |-1 | -|2 |5 |`XX--O---O` |`X` |2 |1 | -|3 |1 |`---------` |`X` |4 |-1 | -|3 |2 |`----X----` |`O` |7 |1 | -|3 |3 |`----X--O-` |`X` |2 |-1 | -|3 |4 |`--X-X--O-` |`O` |8 |1 | -|3 |5 |`--X-X--OO` |`X` |5 |-1 | -|3 |6 |`--X-XX-OO` |`O` |6 |1 | - -## Model Selection and Evaluation - -TBD - -## Model Training - -TBD diff --git a/app/board.py b/app/board.py index 7fd3f75..697d7fb 100644 --- a/app/board.py +++ b/app/board.py @@ -39,14 +39,14 @@ def __repr__(self): """ - @property - def notation(self) -> str: - """ - Represents the board's current state in simple string format like "-X-O-X-OX". - - Position corresponds with square names ['A1','B1','C1','A2','B2','C2','A3','B3','C3'] and indices [0,1,2,3,4,5,6,7,8]. - """ - return "".join([square.notation for square in self.squares]) + #@property + #def notation(self) -> str: + # """ + # Represents the board's current state in simple string format like "-X-O-X-OX". +# + # Position corresponds with square names ['A1','B1','C1','A2','B2','C2','A3','B3','C3'] and indices [0,1,2,3,4,5,6,7,8]. + # """ + # return "".join([square.notation for square in self.squares]) def get_square(self, square_name): @@ -57,7 +57,13 @@ def get_square(self, square_name): def get_squares(self, square_names): return [square for square in self.squares if square.name in square_names] - def set_square(self, square_name, player_letter): + def set_square(self, square_name: str, player_letter: str): + """ + Params: + square_name + + player_letter + """ square = self.get_square(square_name) if not square.letter: square.letter = player_letter diff --git a/app/game.py b/app/game.py index b5b5c82..ccd6546 100644 --- a/app/game.py +++ b/app/game.py @@ -1,5 +1,6 @@ from itertools import cycle +from copy import deepcopy from app.board import Board from app.player import select_player @@ -46,9 +47,8 @@ def take_turn(self, turn: tuple): Pass the turn param as a tuple in the form of (player_letter, square_name). """ player_letter, square_name = turn - initial_board_state = self.board.notation # important to note this before changing the board - - move = Move(board_state=initial_board_state, active_player=player_letter, selected_square=square_name) + initial_board = deepcopy(self.board) # important to note this before changing the board + move = Move(board=initial_board, active_player=player_letter, selected_square=square_name) # make the move / change the board state: self.board.set_square(square_name, player_letter) diff --git a/app/jobs/generate_moves.py b/app/jobs/generate_moves.py new file mode 100644 index 0000000..13671cc --- /dev/null +++ b/app/jobs/generate_moves.py @@ -0,0 +1,113 @@ + + + +import os + +from pandas import DataFrame + +from app import OPPOSITE_LETTERS +#from app.board import SQUARE_NAMES +from app.game import Game +from app.player import select_player +from app.jobs.timer import Timer + + +# for the strategies, use "RANDOM" for random moves, or "MINIMAX-AB" for expert moves +X_STRATEGY = os.getenv("X_STRATEGY", default="RANDOM") +O_STRATEGY = os.getenv("O_STRATEGY", default="RANDOM") + +GAME_COUNT = int(os.getenv("GAME_COUNT", default="1_000")) + + + +class EvaluatedGame(Game): + @property + def player_rewards(self): + if self.winner: + # reward the winner and punish the loser: + winning_letter = self.winner["letter"] + losing_letter = OPPOSITE_LETTERS[winning_letter] + return {winning_letter: 1, losing_letter: 0} + else: + # give neutral scores to both players: + return {"X": 0.5, "O": 0.5} + + +class MoveEvaluator: + def __init__(self): + self.timer = Timer() + self.players = [ + select_player(letter="X", strategy=X_STRATEGY), + select_player(letter="O", strategy=O_STRATEGY), + ] + self.GAME_COUNT = GAME_COUNT + self.moves_df = None + + def perform(self, export=True): + self.timer.start() + records = [] + for game_counter in range(0, self.GAME_COUNT): + game = EvaluatedGame(players=self.players) + game.play() + + player_rewards = game.player_rewards + for move_counter, move in enumerate(game.move_history): + active_player = move.active_player + + # if the active player takes this move they will get the outcome + move.board.set_square(square_name=move.selected_square, player_letter=active_player) + + records.append({ + "game_id": game_counter + 1, # start ids at 1 instead of 0 + "move_id": move_counter + 1, # start ids at 1 instead of 0 + #"board_state": move.board.notation, + "a1": move.board.get_square("A1").notation, + "b1": move.board.get_square("B1").notation, + "c1": move.board.get_square("C1").notation, + "a2": move.board.get_square("A2").notation, + "b2": move.board.get_square("B2").notation, + "c2": move.board.get_square("C2").notation, + "a3": move.board.get_square("A3").notation, + "b3": move.board.get_square("B3").notation, + "c3": move.board.get_square("C3").notation, + "player": active_player, + #"square_name": move.selected_square, + #"square_idx": SQUARE_NAMES.index(move.selected_square), # translate squares to index 0-8 to match board notation (maybe) + "outcome": player_rewards[active_player], + }) + + self.timer.end() + print("------------------------") + print("PLAYED", self.GAME_COUNT, "GAMES", f"IN {self.timer.duration_seconds} SECONDS") + print("TOTAL MOVES:", len(records)) + self.moves_df = DataFrame(records) + print(self.moves_df.head()) + + if export: + print("------------------------") + print("SAVING DATA TO FILE...") + #csv_filename = f"{self.players[0].letter}_{self.players[0].player_type.replace('-','')}" + #csv_filename += "_vs_" + #csv_filename += f"_{self.players[1].letter}_{self.players[1].player_type.replace('-','')}" + + csv_filename = self.players[0].player_type.replace('-','') + "_" + csv_filename += self.players[1].player_type.replace('-','') + csv_filename += f"_{self.GAME_COUNT}.csv" + csv_filename = csv_filename.lower() + csv_filepath = os.path.join(os.path.dirname(__file__), "..", "..", "data", "moves", csv_filename) + + self.moves_df.to_csv(csv_filepath, index=False) + print(os.path.abspath(csv_filepath)) + + return self.moves_df + + + + +if __name__ == "__main__": + + + + job = MoveEvaluator() + + job.perform() diff --git a/app/jobs/play_moves.py b/app/jobs/play_moves.py deleted file mode 100644 index 2e6e4ed..0000000 --- a/app/jobs/play_moves.py +++ /dev/null @@ -1,92 +0,0 @@ - - - -import os - -from pandas import DataFrame - -from app import OPPOSITE_LETTERS -from app.board import SQUARE_NAMES -from app.game import Game -from app.player import select_player -from app.jobs.timer import Timer - - -# for the strategies, use "RANDOM" for random moves, or "MINIMAX-AB" for expert moves -X_STRATEGY = os.getenv("X_STRATEGY", default="RANDOM") -O_STRATEGY = os.getenv("O_STRATEGY", default="RANDOM") - -GAME_COUNT = int(os.getenv("GAME_COUNT", default="100_000")) - -if __name__ == "__main__": - - timer = Timer() - timer.start() - - records = [] - for game_counter in range(0, GAME_COUNT): - game = Game(players=[ - select_player(letter="X", strategy=X_STRATEGY), - select_player(letter="O", strategy=O_STRATEGY), - ]) - - # - # PLAY - # - - game.play() - - # - # OUTCOME EVAL - # - - # determine reward values for each player - if game.winner: - winning_letter = game.winner["letter"] - losing_letter = OPPOSITE_LETTERS[winning_letter] - # reward the winner and punish the loser: - rewards = {winning_letter: 1, losing_letter: -1} - else: - # give neutral scores to both players: - rewards = {"X": 0, "O": 0} - print("------------------------") - print("REWARDS:", rewards) - - # - # PLAYBACK - # - - for move_counter, move in enumerate(game.move_history): - active_player = move.active_player - records.append({ - "game_id": game_counter + 1, # start ids at 1 instead of 0 - "move_id": move_counter + 1, # start ids at 1 instead of 0 - "board_state": move.board_state, - "player": active_player, - "square_name": move.selected_square, - "square_idx": SQUARE_NAMES.index(move.selected_square), # translate squares to index 0-8 to match board notation (maybe) - "reward": rewards[active_player], - }) - - timer.end() - print("------------------------") - print("PLAYED", GAME_COUNT, "GAMES", f"IN {timer.duration_seconds} SECONDS") - print("TOTAL MOVES:", len(records)) - - df = DataFrame(records) - print(df.head()) - - print("------------------------") - print("SAVING DATA TO FILE...") - - csv_filename = f"{game.players[0].letter}_{game.players[0].player_type}" - csv_filename += "_vs_" - csv_filename += f"{game.players[1].letter}_{game.players[1].player_type}" - csv_filename += f"_{GAME_COUNT}.csv" - csv_filename = csv_filename.lower() - csv_filepath = os.path.join(os.path.dirname(__file__), "..", "..", "data", "moves", csv_filename) - - df.to_csv(csv_filepath, index=False) - print(os.path.abspath(csv_filepath)) - - #breakpoint() diff --git a/app/move.py b/app/move.py index e065407..6b544de 100644 --- a/app/move.py +++ b/app/move.py @@ -3,19 +3,19 @@ class Move: - def __init__(self, board_state, active_player, selected_square): + def __init__(self, board, active_player, selected_square): """ Params - board_state (str) the initial board state before the player made the move + board (Board) the initial board state before the player made the move active_player (str) the letter of the player who made the move ("X" or "O") selected_square (str) the name of the square the player selected (e.g "A1") """ - self.board_state = board_state #> "XX-OO----" - self.active_player = active_player #> "X" - self.selected_square = selected_square #> "C1" + self.board = board + self.active_player = active_player + self.selected_square = selected_square def __repr__(self): return f"" diff --git a/ml/README.md b/ml/README.md new file mode 100644 index 0000000..1fac79a --- /dev/null +++ b/ml/README.md @@ -0,0 +1,51 @@ + +# Tic Tac Toe (AI/ML) + +> NOTE: this is currently a work in progress! + +The [computer players](/app/player.py) in the Tic Tac Toe app use a predefined algorithm to select the best move. Let's see if we can train machine learning models to play the game instead. + +## Existing Datasets + +There is a [dataset of tic-tac-toe endgames](https://archive.ics.uci.edu/ml/datasets/Tic-Tac-Toe+Endgame) from UC Irvine. The inputs are the terminal board states, and the output is the game outcome ("positive" or "negative") for the "X" player. It is possible to [train a nice classifier](/endgames/Endgame_Model_Training.ipynb) on their data. But it represents terminal game states only, whereas an AI agent would need to assess non-terminal game states as well. And it assumes the perspective of the "X" player only, whereas an AI agent would need to be able to play as "O" as well. + +## Dataset Generation + +So let's generate a dataset of terminal and non-terminal game states, for both players (in case there are strategy differences between "X" going first and "O" going second). + +In this step, we generate moves datasets for model training and evaluation. The test/train splits are out of scope for this step, and will be done later, during model training. + +The inputs are the board state after the active player makes a move (e.g. "-X-O-X-OX"). The output is the eventual outcome of that move for the given player (win, loss, or tie). + +We simulate alternating moves until we reach an outcome (win, lose, tie). After reaching an outcome, we assign the eventual outcome value to all moves that player made leading up until the outcome: + + all winning player's moves get assigned a positive value (1.0) + + all losing player's moves get assigned a zero (0.0) + + all moves resulting in a tie get a neutral score (0.5) + +Generating the datasets: + +```sh +X_STRATEGY="RANDOM" O_STRATEGY="RANDOM" GAME_COUNT=1000 python -m app.jobs.generate_moves + +X_STRATEGY="RANDOM" O_STRATEGY="MINIMAX-AB" GAME_COUNT=1000 python -m app.jobs.generate_moves + +X_STRATEGY="MINIMAX-AB" O_STRATEGY="RANDOM" GAME_COUNT=1000 python -m app.jobs.generate_moves + +X_STRATEGY="MINIMAX-AB" O_STRATEGY="MINIMAX-AB" GAME_COUNT=1000 python -m app.jobs.generate_moves +``` + +After generating the datasets, we uploaded the CSV files to GitHub, and used a [notebook](/ml/data_prep/Training_Data_Prep.ipynb) to combine the datasets into a single CSV file ("move_values.csv"), which we also upload to GitHub as well: + +Datasets: + + + [random_random_1000.csv](https://github.com/s2t2/tic-tac-toe-py/files/7921041/random_random_1000.csv) + + [random_minimaxab_1000.csv](https://github.com/s2t2/tic-tac-toe-py/files/7921043/random_minimaxab_1000.csv) + + [minimaxab_random_1000.csv](https://github.com/s2t2/tic-tac-toe-py/files/7921050/minimaxab_random_1000.csv) + + [minimaxab_minimaxab_1000.csv](https://github.com/s2t2/tic-tac-toe-py/files/7921045/minimaxab_minimaxab_1000.csv) + +Combined Datasets: + + + [move_values.csv](https://github.com/s2t2/tic-tac-toe-py/files/7921159/move_values.csv) + + [move_values_x.csv](https://github.com/s2t2/tic-tac-toe-py/files/7921160/move_values_x.csv) + + [move_values_o.csv](https://github.com/s2t2/tic-tac-toe-py/files/7921161/move_values_o.csv) + + [move_values_normalized.csv](https://github.com/s2t2/tic-tac-toe-py/files/7921162/move_values_normalized.csv) diff --git a/ml/data_prep/Training_Data_Prep.ipynb b/ml/data_prep/Training_Data_Prep.ipynb new file mode 100644 index 0000000..6e6d1f6 --- /dev/null +++ b/ml/data_prep/Training_Data_Prep.ipynb @@ -0,0 +1,697 @@ +{ + "nbformat": 4, + "nbformat_minor": 0, + "metadata": { + "colab": { + "name": "Tic Tac Toe - Training Data Prep", + "provenance": [] + }, + "kernelspec": { + "name": "python3", + "display_name": "Python 3" + }, + "language_info": { + "name": "python" + } + }, + "cells": [ + { + "cell_type": "markdown", + "source": [ + "## Import datasets" + ], + "metadata": { + "id": "0pG_Kg24wWz6" + } + }, + { + "cell_type": "code", + "execution_count": 77, + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "Hj6PchBDokUF", + "outputId": "8353ecc3-0174-4805-a565-e2d6372af1ad" + }, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "--------------------\n", + "X RANDOM vs O RANDOM:\n", + "1.0 0.460388\n", + "0.0 0.386149\n", + "0.5 0.153463\n", + "Name: outcome, dtype: float64\n", + "--------------------\n", + "X RANDOM vs O EXPERT:\n", + "1.0 0.381239\n", + "0.0 0.381239\n", + "0.5 0.237522\n", + "Name: outcome, dtype: float64\n", + "--------------------\n", + "X EXPERT vs O RANDOM:\n", + "1.0 0.568935\n", + "0.0 0.398080\n", + "0.5 0.032984\n", + "Name: outcome, dtype: float64\n", + "--------------------\n", + "X EXPERT vs O EXPERT:\n", + "0.5 1.0\n", + "Name: outcome, dtype: float64\n" + ] + } + ], + "source": [ + "from pandas import read_csv\n", + "\n", + "print(\"--------------------\")\n", + "print(\"X RANDOM vs O RANDOM:\")\n", + "rr = read_csv(\"https://github.com/s2t2/tic-tac-toe-py/files/7921041/random_random_1000.csv\")\n", + "rr[\"dataset_id\"] = \"random_vs_random\"\n", + "print(rr[\"outcome\"].value_counts(normalize=True))\n", + "\n", + "print(\"--------------------\")\n", + "print(\"X RANDOM vs O EXPERT:\")\n", + "rm = read_csv(\"https://github.com/s2t2/tic-tac-toe-py/files/7921043/random_minimaxab_1000.csv\")\n", + "rm[\"dataset_id\"] = \"random_vs_expert\"\n", + "print(rm[\"outcome\"].value_counts(normalize=True))\n", + "\n", + "print(\"--------------------\")\n", + "print(\"X EXPERT vs O RANDOM:\")\n", + "mr = read_csv(\"https://github.com/s2t2/tic-tac-toe-py/files/7921050/minimaxab_random_1000.csv\")\n", + "mr[\"dataset_id\"] = \"expert_vs_random\"\n", + "print(mr[\"outcome\"].value_counts(normalize=True))\n", + "\n", + "print(\"--------------------\")\n", + "print(\"X EXPERT vs O EXPERT:\")\n", + "mm = read_csv(\"https://github.com/s2t2/tic-tac-toe-py/files/7921045/minimaxab_minimaxab_1000.csv\")\n", + "print(mm[\"outcome\"].value_counts(normalize=True))\n", + "mm[\"dataset_id\"] = \"expert_vs_expert\"\n" + ] + }, + { + "cell_type": "markdown", + "source": [ + "## Combine Datasets" + ], + "metadata": { + "id": "FVHOoqIYwf39" + } + }, + { + "cell_type": "code", + "source": [ + "from pandas import concat\n", + "\n", + "combined_df = concat([rr, rm, mr, mm])\n", + "\n", + "print(combined_df)\n", + "\n", + "print(\"-------------\")\n", + "print(\"OUTCOMES:\")\n", + "print(combined_df[\"outcome\"].value_counts(normalize=True))\n", + "\n", + "print(\"-------------\")\n", + "print(\"DATASETS:\")\n", + "print(combined_df[\"dataset_id\"].value_counts(normalize=True))" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "SZb0fWyQpshF", + "outputId": "fdbc27a6-cfb2-4ad0-ef58-33c1181bdd2c" + }, + "execution_count": 78, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + " game_id move_id a1 b1 c1 a2 ... a3 b3 c3 player outcome dataset_id\n", + "0 1 1 X - - - ... - - - X 1.0 random_vs_random\n", + "1 1 2 X - - - ... - O - O 0.0 random_vs_random\n", + "2 1 3 X X - - ... - O - X 1.0 random_vs_random\n", + "3 1 4 X X - - ... - O - O 0.0 random_vs_random\n", + "4 1 5 X X - X ... - O - X 1.0 random_vs_random\n", + "... ... ... .. .. .. .. ... .. .. .. ... ... ...\n", + "8995 1000 5 X X O - ... X - - X 0.5 expert_vs_expert\n", + "8996 1000 6 X X O O ... X - - O 0.5 expert_vs_expert\n", + "8997 1000 7 X X O O ... X - - X 0.5 expert_vs_expert\n", + "8998 1000 8 X X O O ... X O - O 0.5 expert_vs_expert\n", + "8999 1000 9 X X O O ... X O X X 0.5 expert_vs_expert\n", + "\n", + "[29326 rows x 14 columns]\n", + "-------------\n", + "OUTCOMES:\n", + "0.5 0.409705\n", + "1.0 0.321489\n", + "0.0 0.268806\n", + "Name: outcome, dtype: float64\n", + "-------------\n", + "DATASETS:\n", + "expert_vs_expert 0.306895\n", + "random_vs_random 0.259974\n", + "random_vs_expert 0.237741\n", + "expert_vs_random 0.195390\n", + "Name: dataset_id, dtype: float64\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "combined_df.to_csv(\"move_values.csv\", index=False)\n" + ], + "metadata": { + "id": "S9oYL8hFqALA" + }, + "execution_count": 79, + "outputs": [] + }, + { + "cell_type": "markdown", + "source": [ + "# Separate Player Perspectives" + ], + "metadata": { + "id": "oiKMMt0BMyKk" + } + }, + { + "cell_type": "code", + "source": [ + "df = combined_df.copy()\n", + "\n", + "x_rows = df[df[\"player\"] == \"X\"]\n", + "o_rows = df[df[\"player\"] == \"O\"]\n", + "print(\"X ROWS:\", len(x_rows))\n", + "print(\"O ROWS:\", len(o_rows))\n", + "print(\"TOTAL :\", len(df))" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "Yb5xL0mcM0T8", + "outputId": "0b270a88-ffa2-40a5-96fc-2fd3ef84770f" + }, + "execution_count": 80, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "X ROWS: 16103\n", + "O ROWS: 13223\n", + "TOTAL : 29326\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "o_rows.to_csv(\"move_values_o.csv\")\n", + "x_rows.to_csv(\"move_values_x.csv\")" + ], + "metadata": { + "id": "rz6yYomRM9aJ" + }, + "execution_count": 81, + "outputs": [] + }, + { + "cell_type": "markdown", + "source": [ + "## Normalize O Player Perspective" + ], + "metadata": { + "id": "XG1xWzn4Gvwk" + } + }, + { + "cell_type": "code", + "source": [ + "OPPOSITE_PERSPECTIVES = {\"X\":\"O\", \"O\":\"X\", \"-\":\"-\"}\n", + "OPPOSITE_OUTCOMES = {1:0, 0:1, 0.5:0.5}\n", + "\n", + "def normalize_square(val):\n", + " return OPPOSITE_PERSPECTIVES[val]\n", + "\n", + "#def normalize_board(board_notation):\n", + "# return \"\".join([normalize(char) for char in board_notation])\n", + "\n", + "def normalize_outcome(val):\n", + " return OPPOSITE_OUTCOMES[val]\n", + "\n", + "assert normalize(\"X\") == \"O\"\n", + "assert normalize(\"O\") == \"X\"\n", + "assert normalize(\"-\") == \"-\"\n", + "#assert normalize_board(\"XO-XO----\") == \"OX-OX----\"\n", + "\n", + "assert normalize_outcome(1) == 0\n", + "assert normalize_outcome(0.5) == 0.5\n", + "assert normalize_outcome(0) == 1\n" + ], + "metadata": { + "id": "530ZTDGEGz2u" + }, + "execution_count": 82, + "outputs": [] + }, + { + "cell_type": "code", + "source": [ + "print(\"\\n--------------------\")\n", + "print(\"ORIGINAL PERSPECTIVES:\")\n", + "print(o_rows[[\"a1\",\"b1\", \"c1\", \"a2\",\"b2\", \"c2\", \"a3\",\"b3\", \"c3\", \"player\", \"outcome\"]].head())\n", + "print(o_rows[\"outcome\"].value_counts(normalize=True))\n", + "\n", + "# NORMALIZING...\n", + "new_o_rows = o_rows.copy()\n", + "\n", + "#new_o_rows[\"board_state\"] = o_rows[\"board_state\"].apply(normalize_board)\n", + "for col in [\"a1\",\"b1\",\"c1\",\"a2\",\"b2\",\"c2\",\"a3\",\"b3\",\"c3\"]:\n", + " new_o_rows[col] = o_rows[col].apply(normalize)\n", + "new_o_rows[\"outcome\"] = o_rows[\"outcome\"].apply(normalize_outcome)\n", + "new_o_rows[\"player\"] = \"X\" # normalize player (or drop it)\n", + "\n", + "print(\"\\n--------------------\")\n", + "print(\"NORMALIZED PERSPECTIVES:\")\n", + "print(new_o_rows[[\"a1\",\"b1\", \"c1\", \"a2\",\"b2\", \"c2\", \"a3\",\"b3\", \"c3\", \"player\", \"outcome\"]].head())\n", + "print(new_o_rows[\"outcome\"].value_counts(normalize=True))" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "5KusQOKcHXq3", + "outputId": "bde4dea7-6a7b-4289-c472-007756bc231a" + }, + "execution_count": 83, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "\n", + "--------------------\n", + "ORIGINAL PERSPECTIVES:\n", + " a1 b1 c1 a2 b2 c2 a3 b3 c3 player outcome\n", + "1 X - - - - - - O - O 0.0\n", + "3 X X - - O - - O - O 0.0\n", + "5 X X - X O - - O O O 0.0\n", + "8 - - - O - - - X - O 0.0\n", + "10 X - - O - O - X - O 0.0\n", + "0.5 0.403842\n", + "0.0 0.310141\n", + "1.0 0.286017\n", + "Name: outcome, dtype: float64\n", + "\n", + "--------------------\n", + "NORMALIZED PERSPECTIVES:\n", + " a1 b1 c1 a2 b2 c2 a3 b3 c3 player outcome\n", + "1 O - - - - - - X - X 1.0\n", + "3 O O - - X - - X - X 1.0\n", + "5 O O - O X - - X X X 1.0\n", + "8 - - - X - - - O - X 1.0\n", + "10 O - - X - X - O - X 1.0\n", + "0.5 0.403842\n", + "1.0 0.310141\n", + "0.0 0.286017\n", + "Name: outcome, dtype: float64\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "# https://pandas.pydata.org/docs/reference/api/pandas.concat.html\n", + "from pandas import concat\n", + "\n", + "# single-player df\n", + "print(\"SINGLE PLAYER'S PERSPECTIVE:\")\n", + "df_normalized = concat([x_rows, new_o_rows])\n", + "df_normalized.drop(columns=[\"player\"], inplace=True)\n", + "df_normalized.sort_values(by=[\"dataset_id\", \"game_id\", \"move_id\"], inplace=True)\n", + "print(len(df_normalized))\n", + "\n", + "df_normalized" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 494 + }, + "id": "LlHj10zTHhcZ", + "outputId": "d6e555b4-4a48-4bd9-b00c-17fa85e6753e" + }, + "execution_count": 84, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "SINGLE PLAYER'S PERSPECTIVE:\n", + "29326\n" + ] + }, + { + "output_type": "execute_result", + "data": { + "text/html": [ + "\n", + "
\n", + "
\n", + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
game_idmove_ida1b1c1a2b2c2a3b3c3outcomedataset_id
011-----X---0.5expert_vs_expert
112--X--O---0.5expert_vs_expert
213X-O--X---0.5expert_vs_expert
314O-XX-O---0.5expert_vs_expert
415XXOO-X---0.5expert_vs_expert
..........................................
761910001-X-------1.0random_vs_random
762010002-O---X---1.0random_vs_random
762110003-XX--O---1.0random_vs_random
762210004-OO--X--X1.0random_vs_random
762310005XXX--O--O1.0random_vs_random
\n", + "

29326 rows × 13 columns

\n", + "
\n", + " \n", + " \n", + " \n", + "\n", + " \n", + "
\n", + "
\n", + " " + ], + "text/plain": [ + " game_id move_id a1 b1 c1 a2 b2 c2 a3 b3 c3 outcome dataset_id\n", + "0 1 1 - - - - - X - - - 0.5 expert_vs_expert\n", + "1 1 2 - - X - - O - - - 0.5 expert_vs_expert\n", + "2 1 3 X - O - - X - - - 0.5 expert_vs_expert\n", + "3 1 4 O - X X - O - - - 0.5 expert_vs_expert\n", + "4 1 5 X X O O - X - - - 0.5 expert_vs_expert\n", + "... ... ... .. .. .. .. .. .. .. .. .. ... ...\n", + "7619 1000 1 - X - - - - - - - 1.0 random_vs_random\n", + "7620 1000 2 - O - - - X - - - 1.0 random_vs_random\n", + "7621 1000 3 - X X - - O - - - 1.0 random_vs_random\n", + "7622 1000 4 - O O - - X - - X 1.0 random_vs_random\n", + "7623 1000 5 X X X - - O - - O 1.0 random_vs_random\n", + "\n", + "[29326 rows x 13 columns]" + ] + }, + "metadata": {}, + "execution_count": 84 + } + ] + }, + { + "cell_type": "code", + "source": [ + "df_normalized.to_csv(\"move_values_normalized.csv\")" + ], + "metadata": { + "id": "L2hwCS19P3F3" + }, + "execution_count": 85, + "outputs": [] + } + ] +} \ No newline at end of file diff --git a/ml/endgames/Endgame_Model_Training.ipynb b/ml/endgames/Endgame_Model_Training.ipynb new file mode 100644 index 0000000..1cbf1c1 --- /dev/null +++ b/ml/endgames/Endgame_Model_Training.ipynb @@ -0,0 +1,1831 @@ +{ + "nbformat": 4, + "nbformat_minor": 0, + "metadata": { + "colab": { + "name": "Tic Tac Toe Model Training (END STATES) - 20220102 ", + "provenance": [] + }, + "kernelspec": { + "name": "python3", + "display_name": "Python 3" + }, + "language_info": { + "name": "python" + } + }, + "cells": [ + { + "cell_type": "markdown", + "source": [ + "# Dataset Info\n" + ], + "metadata": { + "id": "u6mV0e_fHOCu" + } + }, + { + "cell_type": "markdown", + "source": [ + "\n", + "http://archive.ics.uci.edu/ml/datasets/Tic-Tac-Toe+Endgame\n", + "\n", + "\n", + "This database encodes the **complete set of possible board configurations at the end of tic-tac-toe games**, where \"x\" is assumed to have played first. \n", + "\n", + "The target concept is: \"win for x\" (i.e. positive when \"x\" has a way to win).\n", + "\n", + "\n", + "Attributes / Columns:\n", + "\n", + "1. top-left-square: {x,o,b}\n", + "2. top-middle-square: {x,o,b}\n", + "3. top-right-square: {x,o,b}\n", + "4. middle-left-square: {x,o,b}\n", + "5. middle-middle-square: {x,o,b}\n", + "6. middle-right-square: {x,o,b}\n", + "7. bottom-left-square: {x,o,b}\n", + "8. bottom-middle-square: {x,o,b}\n", + "9. bottom-right-square: {x,o,b}\n", + "10. Class: {positive,negative}\n", + "\n" + ], + "metadata": { + "id": "2a0xpH6f4kM7" + } + }, + { + "cell_type": "markdown", + "source": [ + "# Formatter Functions" + ], + "metadata": { + "id": "q3-WGcWYHGAP" + } + }, + { + "cell_type": "code", + "source": [ + "\n", + "\n", + "# HELPER FUNCTIONS\n", + "\n", + "def fmt_n(large_number):\n", + " \"\"\"\n", + " Formats a large number with thousands separator, for printing and logging.\n", + " \n", + " Param large_number (int) like 1_000_000_000\n", + " \n", + " Returns (str) like '1,000,000,000'\n", + " \"\"\"\n", + " return f\"{large_number:,.0f}\"\n", + "\n", + "\n", + "def fmt_pct(decimal_number):\n", + " \"\"\"\n", + " Formats a large number with thousands separator, for printing and logging.\n", + " \n", + " Param decimal_number (float) like 0.95555555555\n", + " \n", + " Returns (str) like '95.5%'\n", + " \"\"\"\n", + " return f\"{decimal_number:.2f}%\"\n", + "\n", + "\n", + "assert fmt_n(1_000_000_000) == '1,000,000,000'\n", + "assert fmt_pct(9.67890987) == '9.68%'" + ], + "metadata": { + "id": "b7416I9bELB5" + }, + "execution_count": 1, + "outputs": [] + }, + { + "cell_type": "markdown", + "source": [ + "# Importing the Data\n" + ], + "metadata": { + "id": "SY8YP2-p53Mh" + } + }, + { + "cell_type": "code", + "source": [ + "#\n", + "# IMPORT DATA\n", + "# https://pandas.pydata.org/docs/reference/api/pandas.read_csv.html\n", + "#\n", + "\n", + "from pandas import read_csv\n", + "\n", + "url = \"http://archive.ics.uci.edu/ml/machine-learning-databases/tic-tac-toe/tic-tac-toe.data\"\n", + "\n", + "columns = [\"a1\", \"a2\", \"a3\", \"b1\", \"b2\", \"b3\", \"c1\", \"c2\", \"c3\", \"outcome\"]\n", + "\n", + "raw_df = read_csv(url, names=columns)\n", + "raw_df.head()" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 204 + }, + "id": "IYxmq5QT564m", + "outputId": "54096f45-7388-4864-dad2-9188a4ae575d" + }, + "execution_count": 2, + "outputs": [ + { + "output_type": "execute_result", + "data": { + "text/html": [ + "\n", + "
\n", + "
\n", + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
a1a2a3b1b2b3c1c2c3outcome
0xxxxooxoopositive
1xxxxoooxopositive
2xxxxooooxpositive
3xxxxooobbpositive
4xxxxoobobpositive
\n", + "
\n", + " \n", + " \n", + " \n", + "\n", + " \n", + "
\n", + "
\n", + " " + ], + "text/plain": [ + " a1 a2 a3 b1 b2 b3 c1 c2 c3 outcome\n", + "0 x x x x o o x o o positive\n", + "1 x x x x o o o x o positive\n", + "2 x x x x o o o o x positive\n", + "3 x x x x o o o b b positive\n", + "4 x x x x o o b o b positive" + ] + }, + "metadata": {}, + "execution_count": 2 + } + ] + }, + { + "cell_type": "markdown", + "source": [ + "The positive outcome rows are first and the negative outcome rows are later, so we'll have to shuffle the row order / randomize when splitting the data." + ], + "metadata": { + "id": "7kEqVWLm56T1" + } + }, + { + "cell_type": "code", + "source": [ + "print(raw_df[\"outcome\"].value_counts(normalize=True))" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "FQ_XwcoX-5vW", + "outputId": "2a698c41-e09a-4ed7-86db-eb76d77220b7" + }, + "execution_count": 3, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "positive 0.653445\n", + "negative 0.346555\n", + "Name: outcome, dtype: float64\n" + ] + } + ] + }, + { + "cell_type": "markdown", + "source": [ + "there is a 2:1 ratio between positive and negative outcomes. this is generally imbalanced. \n", + "\n", + "https://machinelearningmastery.com/types-of-classification-in-machine-learning/" + ], + "metadata": { + "id": "-jSzc1cm-8rq" + } + }, + { + "cell_type": "markdown", + "source": [ + "# Splitting the Data" + ], + "metadata": { + "id": "ABHaHgU9_h2o" + } + }, + { + "cell_type": "code", + "source": [ + "\n", + "\n", + "# https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html\n", + "\n", + "from sklearn.model_selection import train_test_split\n", + "\n", + "\n", + "train_df, test_df = train_test_split(raw_df, train_size=0.8, shuffle=True, random_state=99, stratify= raw_df[\"outcome\"])\n", + "\n", + "print(\"--------------------\")\n", + "print(\"TRAIN:\", len(train_df))\n", + "print(train_df[\"outcome\"].value_counts(normalize=True))\n", + "\n", + "print(\"--------------------\")\n", + "print(\"TEST:\", len(test_df))\n", + "print(test_df[\"outcome\"].value_counts(normalize=True))\n", + "\n", + "print(\"--------------------\")\n", + "train_df" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 606 + }, + "id": "oFWpPzJi7I1M", + "outputId": "9a31a5cd-cda5-48ec-de87-b6b737926f23" + }, + "execution_count": 4, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "--------------------\n", + "TRAIN: 766\n", + "positive 0.654047\n", + "negative 0.345953\n", + "Name: outcome, dtype: float64\n", + "--------------------\n", + "TEST: 192\n", + "positive 0.651042\n", + "negative 0.348958\n", + "Name: outcome, dtype: float64\n", + "--------------------\n" + ] + }, + { + "output_type": "execute_result", + "data": { + "text/html": [ + "\n", + "
\n", + "
\n", + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
a1a2a3b1b2b3c1c2c3outcome
783oxoobxoxxnegative
468oboobxxxxpositive
405ooxbbxoxxpositive
820oooxoxxxbnegative
874obboxxoxbnegative
.................................
232xbxoxbooxpositive
449obxbxoxbbpositive
381ooxoxxxxopositive
838ooobxxoxxnegative
403ooxbobxxxpositive
\n", + "

766 rows × 10 columns

\n", + "
\n", + " \n", + " \n", + " \n", + "\n", + " \n", + "
\n", + "
\n", + " " + ], + "text/plain": [ + " a1 a2 a3 b1 b2 b3 c1 c2 c3 outcome\n", + "783 o x o o b x o x x negative\n", + "468 o b o o b x x x x positive\n", + "405 o o x b b x o x x positive\n", + "820 o o o x o x x x b negative\n", + "874 o b b o x x o x b negative\n", + ".. .. .. .. .. .. .. .. .. .. ...\n", + "232 x b x o x b o o x positive\n", + "449 o b x b x o x b b positive\n", + "381 o o x o x x x x o positive\n", + "838 o o o b x x o x x negative\n", + "403 o o x b o b x x x positive\n", + "\n", + "[766 rows x 10 columns]" + ] + }, + "metadata": {}, + "execution_count": 4 + } + ] + }, + { + "cell_type": "markdown", + "source": [ + "# Feature Selection" + ], + "metadata": { + "id": "z84Eh1FOFUA0" + } + }, + { + "cell_type": "code", + "source": [ + "target_col = \"outcome\"\n", + "\n", + "feature_cols = columns\n", + "if target_col in feature_cols:\n", + " feature_cols.remove(target_col)\n", + "print(\"FEATURE COLS:\", feature_cols)\n", + "\n", + "train_x = train_df[feature_cols]\n", + "train_y = train_df[[target_col]]\n", + "\n", + "test_x = test_df[feature_cols]\n", + "test_y = test_df[[target_col]]" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "4Z6IYvoS_7rr", + "outputId": "28e1b4db-2003-4d74-b57f-fa9aa89c4be1" + }, + "execution_count": 5, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "FEATURE COLS: ['a1', 'a2', 'a3', 'b1', 'b2', 'b3', 'c1', 'c2', 'c3']\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "train_x" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 419 + }, + "id": "5Ty17vtsCl89", + "outputId": "df1ed55f-45ca-4253-9a50-22710c4abd48" + }, + "execution_count": 6, + "outputs": [ + { + "output_type": "execute_result", + "data": { + "text/html": [ + "\n", + "
\n", + "
\n", + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
a1a2a3b1b2b3c1c2c3
783oxoobxoxx
468oboobxxxx
405ooxbbxoxx
820oooxoxxxb
874obboxxoxb
..............................
232xbxoxboox
449obxbxoxbb
381ooxoxxxxo
838ooobxxoxx
403ooxbobxxx
\n", + "

766 rows × 9 columns

\n", + "
\n", + " \n", + " \n", + " \n", + "\n", + " \n", + "
\n", + "
\n", + " " + ], + "text/plain": [ + " a1 a2 a3 b1 b2 b3 c1 c2 c3\n", + "783 o x o o b x o x x\n", + "468 o b o o b x x x x\n", + "405 o o x b b x o x x\n", + "820 o o o x o x x x b\n", + "874 o b b o x x o x b\n", + ".. .. .. .. .. .. .. .. .. ..\n", + "232 x b x o x b o o x\n", + "449 o b x b x o x b b\n", + "381 o o x o x x x x o\n", + "838 o o o b x x o x x\n", + "403 o o x b o b x x x\n", + "\n", + "[766 rows x 9 columns]" + ] + }, + "metadata": {}, + "execution_count": 6 + } + ] + }, + { + "cell_type": "code", + "source": [ + "train_y" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 419 + }, + "id": "mXbpOPMpCnZb", + "outputId": "69f32a6d-6521-435f-b6ce-9aece60eece9" + }, + "execution_count": 7, + "outputs": [ + { + "output_type": "execute_result", + "data": { + "text/html": [ + "\n", + "
\n", + "
\n", + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
outcome
783negative
468positive
405positive
820negative
874negative
......
232positive
449positive
381positive
838negative
403positive
\n", + "

766 rows × 1 columns

\n", + "
\n", + " \n", + " \n", + " \n", + "\n", + " \n", + "
\n", + "
\n", + " " + ], + "text/plain": [ + " outcome\n", + "783 negative\n", + "468 positive\n", + "405 positive\n", + "820 negative\n", + "874 negative\n", + ".. ...\n", + "232 positive\n", + "449 positive\n", + "381 positive\n", + "838 negative\n", + "403 positive\n", + "\n", + "[766 rows x 1 columns]" + ] + }, + "metadata": {}, + "execution_count": 7 + } + ] + }, + { + "cell_type": "markdown", + "source": [ + "# Feature Matrix Encoding" + ], + "metadata": { + "id": "9aridLI6Hf6b" + } + }, + { + "cell_type": "code", + "source": [ + "# https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.OneHotEncoder.html\n", + "\n", + "from sklearn.preprocessing import OneHotEncoder\n", + "\n", + "encoder = OneHotEncoder()\n", + "\n", + "encoder.fit(train_x) \n", + "\n", + "features = encoder.get_feature_names_out()\n", + "print(\"FEATURES :\", fmt_n(len(features)))\n", + "print(features)\n", + "\n", + "train_m = encoder.transform(train_x)\n", + "test_m = encoder.transform(test_x)" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "AYKvZl3GHeWm", + "outputId": "d0f9475f-9924-45fb-fb9d-02177bff3f45" + }, + "execution_count": 8, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "FEATURES : 27\n", + "['a1_b' 'a1_o' 'a1_x' 'a2_b' 'a2_o' 'a2_x' 'a3_b' 'a3_o' 'a3_x' 'b1_b'\n", + " 'b1_o' 'b1_x' 'b2_b' 'b2_o' 'b2_x' 'b3_b' 'b3_o' 'b3_x' 'c1_b' 'c1_o'\n", + " 'c1_x' 'c2_b' 'c2_o' 'c2_x' 'c3_b' 'c3_o' 'c3_x']\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "print(train_m.shape)\n", + "print(test_m.shape)" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "4mInxKSoKCse", + "outputId": "3f7c71ac-814c-43ca-df84-a8154608eab5" + }, + "execution_count": 9, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "(766, 27)\n", + "(192, 27)\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "print(train_m[0])" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "J9_I06q3KXEq", + "outputId": "b1e18e15-9d6c-4ba7-cc20-9eadbff6fb72" + }, + "execution_count": 10, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + " (0, 1)\t1.0\n", + " (0, 5)\t1.0\n", + " (0, 7)\t1.0\n", + " (0, 10)\t1.0\n", + " (0, 12)\t1.0\n", + " (0, 17)\t1.0\n", + " (0, 19)\t1.0\n", + " (0, 23)\t1.0\n", + " (0, 26)\t1.0\n" + ] + } + ] + }, + { + "cell_type": "markdown", + "source": [ + "# Model Selection and Training" + ], + "metadata": { + "id": "r4gEro9jBL6r" + } + }, + { + "cell_type": "code", + "source": [ + "from pprint import pprint\n", + "from sklearn.metrics import classification_report #, accuracy_score\n", + "\n", + "#def train_and_score(model, train_m, train_y, test_m, test_y):\n", + "def train_and_score(model):\n", + "\n", + " print(\"-------------------\")\n", + " print(\"TRAINING...\")\n", + " model.fit(train_m, train_y)\n", + "\n", + " train_y_pred = model.predict(train_m)\n", + " train_scores = classification_report(train_y, train_y_pred, output_dict=True)\n", + " #print(\"ACCY:\", fmt_pct(train_scores[\"accuracy\"]), \"GOAL:\", fmt_pct(train_scores[recall_class][\"recall\"]))\n", + " pprint(train_scores)\n", + "\n", + " print(\"-------------------\")\n", + " print(\"TESTING...\")\n", + " test_y_pred = model.predict(test_m)\n", + " test_scores = classification_report(test_y, test_y_pred, output_dict=True)\n", + " #print(\"ACCY:\", fmt_pct(test_scores[\"accuracy\"]), \"GOAL:\", fmt_pct(test_scores[recall_class][\"recall\"]))\n", + " pprint(test_scores)\n", + "\n" + ], + "metadata": { + "id": "KdlN86_CBN6r" + }, + "execution_count": 11, + "outputs": [] + }, + { + "cell_type": "code", + "source": [ + "# https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html\n", + "\n", + "from sklearn.linear_model import LogisticRegression\n", + "\n", + "lr = LogisticRegression(random_state=99, max_iter=2000)\n", + "\n", + "train_and_score(model=lr)" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "jDg96coCDA2k", + "outputId": "5f29fd82-f1d9-46f6-c71e-d4bdb063a356" + }, + "execution_count": 12, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "-------------------\n", + "TRAINING...\n", + "{'accuracy': 0.9817232375979112,\n", + " 'macro avg': {'f1-score': 0.9795443447476042,\n", + " 'precision': 0.9864077669902913,\n", + " 'recall': 0.9735849056603774,\n", + " 'support': 766},\n", + " 'negative': {'f1-score': 0.9728682170542635,\n", + " 'precision': 1.0,\n", + " 'recall': 0.9471698113207547,\n", + " 'support': 265},\n", + " 'positive': {'f1-score': 0.9862204724409448,\n", + " 'precision': 0.9728155339805825,\n", + " 'recall': 1.0,\n", + " 'support': 501},\n", + " 'weighted avg': {'f1-score': 0.9816012195982939,\n", + " 'precision': 0.9822200816243758,\n", + " 'recall': 0.9817232375979112,\n", + " 'support': 766}}\n", + "-------------------\n", + "TESTING...\n", + "{'accuracy': 0.9895833333333334,\n", + " 'macro avg': {'f1-score': 0.9884559884559885,\n", + " 'precision': 0.9921259842519685,\n", + " 'recall': 0.9850746268656716,\n", + " 'support': 192},\n", + " 'negative': {'f1-score': 0.9848484848484849,\n", + " 'precision': 1.0,\n", + " 'recall': 0.9701492537313433,\n", + " 'support': 67},\n", + " 'positive': {'f1-score': 0.9920634920634921,\n", + " 'precision': 0.984251968503937,\n", + " 'recall': 1.0,\n", + " 'support': 125},\n", + " 'weighted avg': {'f1-score': 0.9895457551707553,\n", + " 'precision': 0.989747375328084,\n", + " 'recall': 0.9895833333333334,\n", + " 'support': 192}}\n" + ] + }, + { + "output_type": "stream", + "name": "stderr", + "text": [ + "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:993: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n", + " y = column_or_1d(y, warn=True)\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "# https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegressionCV.html\n", + "# https://scikit-learn.org/stable/glossary.html#term-cross-validation-estimator\n", + "\n", + "from sklearn.linear_model import LogisticRegressionCV\n", + "\n", + "lr_cv = LogisticRegressionCV(random_state=99)\n", + "\n", + "train_and_score(lr_cv)" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "XuBbSRgQDXEt", + "outputId": "dae601dc-8cb0-46b0-c5db-d351f5ee4bfd" + }, + "execution_count": 13, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "-------------------\n", + "TRAINING...\n" + ] + }, + { + "output_type": "stream", + "name": "stderr", + "text": [ + "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:993: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n", + " y = column_or_1d(y, warn=True)\n" + ] + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "{'accuracy': 0.9817232375979112,\n", + " 'macro avg': {'f1-score': 0.9795443447476042,\n", + " 'precision': 0.9864077669902913,\n", + " 'recall': 0.9735849056603774,\n", + " 'support': 766},\n", + " 'negative': {'f1-score': 0.9728682170542635,\n", + " 'precision': 1.0,\n", + " 'recall': 0.9471698113207547,\n", + " 'support': 265},\n", + " 'positive': {'f1-score': 0.9862204724409448,\n", + " 'precision': 0.9728155339805825,\n", + " 'recall': 1.0,\n", + " 'support': 501},\n", + " 'weighted avg': {'f1-score': 0.9816012195982939,\n", + " 'precision': 0.9822200816243758,\n", + " 'recall': 0.9817232375979112,\n", + " 'support': 766}}\n", + "-------------------\n", + "TESTING...\n", + "{'accuracy': 0.9895833333333334,\n", + " 'macro avg': {'f1-score': 0.9884559884559885,\n", + " 'precision': 0.9921259842519685,\n", + " 'recall': 0.9850746268656716,\n", + " 'support': 192},\n", + " 'negative': {'f1-score': 0.9848484848484849,\n", + " 'precision': 1.0,\n", + " 'recall': 0.9701492537313433,\n", + " 'support': 67},\n", + " 'positive': {'f1-score': 0.9920634920634921,\n", + " 'precision': 0.984251968503937,\n", + " 'recall': 1.0,\n", + " 'support': 125},\n", + " 'weighted avg': {'f1-score': 0.9895457551707553,\n", + " 'precision': 0.989747375328084,\n", + " 'recall': 0.9895833333333334,\n", + " 'support': 192}}\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "# https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html\n", + "\n", + "from sklearn.tree import DecisionTreeClassifier\n", + "\n", + "dt = DecisionTreeClassifier(random_state=99)\n", + "\n", + "train_and_score(dt)" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "CVnedxS0JFdl", + "outputId": "63f67bda-21c4-4fbf-b808-05243c29aae3" + }, + "execution_count": 14, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "-------------------\n", + "TRAINING...\n", + "{'accuracy': 1.0,\n", + " 'macro avg': {'f1-score': 1.0,\n", + " 'precision': 1.0,\n", + " 'recall': 1.0,\n", + " 'support': 766},\n", + " 'negative': {'f1-score': 1.0, 'precision': 1.0, 'recall': 1.0, 'support': 265},\n", + " 'positive': {'f1-score': 1.0, 'precision': 1.0, 'recall': 1.0, 'support': 501},\n", + " 'weighted avg': {'f1-score': 1.0,\n", + " 'precision': 1.0,\n", + " 'recall': 1.0,\n", + " 'support': 766}}\n", + "-------------------\n", + "TESTING...\n", + "{'accuracy': 0.9583333333333334,\n", + " 'macro avg': {'f1-score': 0.9547543301519972,\n", + " 'precision': 0.9494820160633222,\n", + " 'recall': 0.9610746268656716,\n", + " 'support': 192},\n", + " 'negative': {'f1-score': 0.9420289855072463,\n", + " 'precision': 0.9154929577464789,\n", + " 'recall': 0.9701492537313433,\n", + " 'support': 67},\n", + " 'positive': {'f1-score': 0.967479674796748,\n", + " 'precision': 0.9834710743801653,\n", + " 'recall': 0.952,\n", + " 'support': 125},\n", + " 'weighted avg': {'f1-score': 0.9585984446800989,\n", + " 'precision': 0.9597495440965352,\n", + " 'recall': 0.9583333333333334,\n", + " 'support': 192}}\n" + ] + } + ] + }, + { + "cell_type": "markdown", + "source": [ + "# Feature Importances" + ], + "metadata": { + "id": "VtGGKLa3LaWF" + } + }, + { + "cell_type": "code", + "source": [ + "from pandas import Series\n", + "\n", + "model = dt\n", + "\n", + "feature_importances = Series(model.feature_importances_, features).sort_values(ascending=False)\n", + "feature_importances.name = \"encoded_feature\"\n", + "feature_importances" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "NhlUStVyLcLa", + "outputId": "d67ad42c-1f44-4957-8444-2e7a20e61213" + }, + "execution_count": 15, + "outputs": [ + { + "output_type": "execute_result", + "data": { + "text/plain": [ + "b2_o 0.113211\n", + "c1_x 0.089668\n", + "c1_o 0.082508\n", + "a3_o 0.074124\n", + "b3_o 0.070662\n", + "a1_o 0.070609\n", + "c3_x 0.070545\n", + "a2_o 0.060734\n", + "c2_o 0.060618\n", + "b1_o 0.052907\n", + "a2_x 0.045175\n", + "b3_x 0.044105\n", + "b1_x 0.036198\n", + "a1_x 0.032857\n", + "c2_x 0.027928\n", + "c3_o 0.026983\n", + "a3_x 0.015001\n", + "c2_b 0.009991\n", + "a3_b 0.004616\n", + "b3_b 0.003228\n", + "a1_b 0.002885\n", + "a2_b 0.002400\n", + "c1_b 0.001539\n", + "b2_b 0.000962\n", + "c3_b 0.000549\n", + "b2_x 0.000000\n", + "b1_b 0.000000\n", + "Name: encoded_feature, dtype: float64" + ] + }, + "metadata": {}, + "execution_count": 15 + } + ] + }, + { + "cell_type": "code", + "source": [ + "from plotly.express import bar\n", + "\n", + "top_features = feature_importances.sort_values(ascending=True)# [0:10]\n", + "\n", + "fig = bar(\n", + " x=top_features.values,\n", + " y=top_features.keys(),\n", + " orientation=\"h\", # horizontal flips x and y params below, but not labels\n", + " labels={\"x\": \"Relative Importance\", \"y\": \"Feature Name\"},\n", + " title=\"Importance of Board Features for X Player in Tic Tac Toe\",\n", + " height=750\n", + ")\n", + "fig.show()" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 767 + }, + "id": "ZtdEBgXXMIjq", + "outputId": "3b6f618b-af39-411a-a194-39d2c4a9d8b3" + }, + "execution_count": 16, + "outputs": [ + { + "output_type": "display_data", + "data": { + "text/html": [ + "\n", + "\n", + "\n", + "
\n", + " \n", + " \n", + " \n", + "
\n", + " \n", + "
\n", + "\n", + "" + ] + }, + "metadata": {} + } + ] + }, + { + "cell_type": "markdown", + "source": [ + "# Decision Tree Visualization" + ], + "metadata": { + "id": "9JjUhZCYN3RO" + } + }, + { + "cell_type": "code", + "source": [ + "# https://scikit-learn.org/stable/modules/generated/sklearn.tree.export_graphviz.html\n", + "from sklearn.tree import export_graphviz\n", + "\n", + "# https://pypi.org/project/graphviz/\n", + "from graphviz import Source \n", + "\n", + "#class_names = [{0:\"Not Survive\", 1:\"Survive\"}[class_name] for class_name in search.classes_] # [0, 1]\n", + "\n", + "dot_data = export_graphviz(dt,\n", + " out_file=None,\n", + " max_depth=3,\n", + " feature_names=features,\n", + " #class_names=class_names, # expects class names to be strings\n", + " impurity=False,\n", + " filled=True,\n", + " proportion=True,\n", + " rounded=True\n", + ")\n", + "graph = Source(dot_data)\n", + "png_bytes = graph.pipe(format=\"png\")\n", + "with open(\"decision_tree.png\", \"wb\") as f:\n", + " f.write(png_bytes)" + ], + "metadata": { + "id": "NBYX71tEN7Mb" + }, + "execution_count": 17, + "outputs": [] + }, + { + "cell_type": "code", + "source": [ + "from IPython.display import Image, display\n", + "\n", + "display(Image(filename=\"decision_tree.png\"))" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 344 + }, + "id": "7nz6DXjAOo0X", + "outputId": "28273b29-3a46-445c-f3a4-6431a86b5ab1" + }, + "execution_count": 18, + "outputs": [ + { + "output_type": "display_data", + "data": { + "image/png": "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\n", + "text/plain": [ + "" + ] + }, + "metadata": {} + } + ] + } + ] +} \ No newline at end of file diff --git a/test/board_test.py b/test/board_test.py index 6f27c10..456ebc9 100644 --- a/test/board_test.py +++ b/test/board_test.py @@ -55,19 +55,19 @@ def test_winner_determination(): -def test_board_state_notation(): - - board = Board() - assert board.notation == "---------" - - board.set_square("A1", "X") - assert board.notation == "X--------" - - board.set_square("A2", "O") - assert board.notation == "X--O-----" - - board.set_square("B1", "X") - assert board.notation == "XX-O-----" - - board.set_square("B2", "O") - assert board.notation == "XX-OO----" +#def test_board_state_notation(): +# +# board = Board() +# assert board.notation == "---------" +# +# board.set_square("A1", "X") +# assert board.notation == "X--------" +# +# board.set_square("A2", "O") +# assert board.notation == "X--O-----" +# +# board.set_square("B1", "X") +# assert board.notation == "XX-O-----" +# +# board.set_square("B2", "O") +# assert board.notation == "XX-OO----"