Showdown

MariBot, an extension of pmariglia's Pokémon battle-bot that can play battles on Pokemon Showdown. The bot can play single battles in generations 1 through 9.

This is a modification for pmariglia's original Pokemon battle-bot which contains some experimental modifications and extensions. The project was born out of an idea to try to extend the capability of the bot to play gen 1 and gen 2 at a basic support level, which then snowballed into a variety of small-ish changes to the code base which, in my subjective opinion, enhance the experience with the bot in a variety of ways. A lot of the code is experimental and isn't meant to be written in the cleanest way possible, it may also contain bugs, so please be warned :) There's no support for any of these changes at the moment, as this is mostly meant as a personal "sandbox" fork for my friends and me. Please don't bother the original bot author with bug reports as well if you're playing this fork and experiencing any issues.

Core features of this fork (not meant to be an extensive list):

• The bot differentiates between generations when it comes to the type chart and various other (mostly minor) ways. Not comprehensive.
• Gen 1 and Gen 2 singles are supported at a basic level, the AI will not crash upon entering the match, and some quirks of these early generations are understood by the bot (such as e.g. type differences), but definitely not all, so many things are still seen by the bot through the lens of the newer generations (leading to suboptimal choices).
• Additional configuration options are available in the "env" file, namely "STATE_SEARCH_DEPTH" (which defines the search depth for the AI, defaulting to 2 and allowing the value of 0 to set up the dynamic search depth which increases the depth as the game goes longer), "STATE_SEARCH_PRUNE_TREE" (experimental option that removes the pruning of the state search tree, used to test the bot behavior with and without pruning), "DYNAMIC_SEARCH_OPTS_FOR_MAX" (how many total options are allowed for the AI to go for the maximum state search depth of 4 when dynamic search depth is enabled), "DYNAMIC_SEARCH_BATTLE_THRESHOLD" (how many battles can there be for the dynamic search to increase the depth to 3), "BATTLE_TIMER" (can be disabled so that the 5-minute timer is not used, helpful for higher search depths), "EXPECTED_MODS" (see below), "PREFERRED_AVATAR" (the avatar that the AI will set for itself), "LOCAL_INSECURE_LOGIN" (if set to True, will try to connect to the server with no authentication or password - this should only be done when using the patched local server/client which allows unregistered insecure login!).
• Support for more random formats. Namely, Battle Factory and BSS Factory will now work as expected without crashing, and there are now sets for randbats in every generation, so hopefully the bot will be a little bit more discerning in which sets to expect depending on which gen it is playing.
• Support for some mods, namely, Scalemons, Camomons and 350 Cup, through the "EXPECTED_MODS" option (these mods can be listed by name in the EXPECTED_MODS variable, then the AI will try to account for the mon stat changes in the relevant modded formats).
• Some improvements and fixes are attempted to the AI (namely, the AI properly understands Sleep/Rest Talk now, understands Trick Room teams better, understands Counter and Mirror Coat better for all those Wobbuffet shenanigans, will try not to throw extra layers of hazards when already at max, etc.)
• The bot will no longer crash if it fails to load the Smogon stats from the online source (which is possible for some obscure formats like gen7nu which is no longer played much, if at all). Also, the bot can load a local Smogon stats file as a fallback, from the "smogon" folder (put e.g. the gen7nu-0.json file in there, downloaded from the "chaos" subfolder of one of the folders at https://www.smogon.com/stats/), this allows you to use the local Smogon stats base, including some obscure older formats that are no longer played today and which the bot can't detect and download stats for automatically.
• Bot profiles/personalities (name + avatar combinations) can be set up for the local server/client, but this is mostly undocumented for now, and it ONLY works on the local server which requires some patching to support password-less insecure login. If you're interested, check out the source code (e.g. for LOCAL_INSECURE_LOGIN, PS_USERNAME and PS_PASSWORD, as well as PREFERRED_AVATAR) and keep in mind that you're on your own if you decide to try it! Don't try it on the official Pokemon Showdown, it won't work and you may possibly even be banned!
• Some minor features from other forks are integrated here, e.g. the Random battle bot, courtesy of its original author SwagMander.
• There's now a rudimentary GUI configuration tool which allows you to set up the "env" file, check out configure_tk.py. Note that it requires TkInter, and for some reason it's reported not to work on MacOS - sadly, can't test or comment on this since I don't own a Mac personally.
• There's now an option (ALLOW_TERA_TO_STELLAR_TYPE) that enables (default) or disables the ability of the bot to terastallize into Stellar type. Disabling this option may be a good way to practice for a Draft where tera captains are chosen, which terastallize to specific types, and other mon are usually marked with Stellar to show that they can't terastallize.
• Probably some other (mostly minor) stuff that I forgot.

This fork may be modified further as time goes by in case I find something interesting to try, test, and improve. Anyone is welcome to use and reuse the code from this fork if you find anything here useful. Any of this code can also be backported into the original repo, with any needed changes, if the author of the bot finds anything here of interest.

The original readme by pmariglia follows. Please note that it's unmodified and doesn't contain any additional information about the modifications in this fork. Any additional documentation regarding the changes and improvements in this fork may be committed at a later date given enough time.

Python version

Developed and tested using Python 3.8.

Getting Started

Configuration

Environment variables are used for configuration. You may either set these in your environment before running, or populate them in the env file.

The configurations available are:

Config Name	Type	Required	Description
BATTLE_BOT	string	yes	The BattleBot module to use. More on this below in the Battle Bots section
WEBSOCKET_URI	string	yes	The address to use to connect to the Pokemon Showdown websocket
PS_USERNAME	string	yes	Pokemon Showdown username
PS_PASSWORD	string	yes	Pokemon Showdown password
BOT_MODE	string	yes	The mode the the bot will operate in. Options are CHALLENGE_USER, SEARCH_LADDER, or ACCEPT_CHALLENGE
POKEMON_MODE	string	yes	The type of game this bot will play: gen8ou, gen7randombattle, etc.
USER_TO_CHALLENGE	string	only if BOT_MODE is CHALLENGE_USER	If BOT_MODE is CHALLENGE_USER, this is the name of the user you want your bot to challenge
RUN_COUNT	int	no	The number of games the bot will play before quitting
TEAM_NAME	string	no	The name of the file that contains the team you want to use. More on this below in the Specifying Teams section.
ROOM_NAME	string	no	If BOT_MODE is ACCEPT_CHALLENGE, the bot will join this chatroom while waiting for a challenge.
SAVE_REPLAY	boolean	no	Specifies whether or not to save replays of the battles (True / False)
LOG_LEVEL	string	no	The Python logging level (DEBUG, INFO, etc.)

Running without Docker

1. Clone

Clone the repository with git clone https://github.com/pmariglia/showdown.git

2. Install Requirements

Install the requirements with pip install -r requirements.txt.

3. Configure your env file

Here is a sample:

BATTLE_BOT=safest
WEBSOCKET_URI=wss://sim3.psim.us/showdown/websocket
PS_USERNAME=MyUsername
PS_PASSWORD=MyPassword
BOT_MODE=SEARCH_LADDER
POKEMON_MODE=gen7randombattle
RUN_COUNT=1

4. Run

Run with python run.py

Running with Docker

This requires Docker 17.06 or higher.

1. Clone the repository

git clone https://github.com/pmariglia/showdown.git

2. Build the Docker image

docker build . -t showdown

3. Run with an environment variable file

docker run --env-file env showdown

Battle Bots

This project has a few different battle bot implementations. Each of these battle bots use a different method to determine which move to use.

Safest

use BATTLE_BOT=safest

The bot searches through the game-tree for two turns and selects the move that minimizes the possible loss for a turn.

For decisions with random outcomes a weighted average is taken for all possible end states. For example: If using draco meteor versus some arbitrary other move results in a score of 1000 if it hits (90%) and a score of 900 if it misses (10%), the overall score for using draco meteor is (0.9 * 1000) + (0.1 * 900) = 990.

This is equivalent to the Expectiminimax strategy.

This decision type is deterministic - the bot will always make the same move given the same situation again.

Nash-Equilibrium (experimental)

use BATTLE_BOT=nash_equilibrium

Using the information it has, plus some assumptions about the opponent, the bot will attempt to calculate the Nash-Equilibrium with the highest payoff and select a move from that distribution.

The Nash Equilibrium is calculated using command-line tools provided by the Gambit project. This decision method should only be used when running with Docker and will fail otherwise.

This decision method is not deterministic. The bot may make a different move if presented with the same situation again.

Team Datasets (experimental)

use BATTLE_BOT=team_datasets

Using a file of sets & teams, this battle-bot is meant to have a better understanding of Pokeon sets that may appear. Populate this dataset by editing data/team_datasets.json.

Still uses the safest decision making method for picking a move, but in theory the knowledge of sets should result in better decision making.

Most Damage

use BATTLE_BOT=most_damage

Selects the move that will do the most damage to the opponent

Does not switch

Write your own bot

Create a package in showdown/battle_bots with a module named main.py. In this module, create a class named BattleBot, override the Battle class, and implement your own find_best_move function.

Set the BATTLE_BOT environment variable to the name of your package and your function will be called each time PokemonShowdown prompts the bot for a move

The Battle Engine

The bots in the project all use a Pokemon battle engine to determine all possible transpositions that may occur from a pair of moves.

For more information, see ENGINE.md

Specifying Teams

You can specify teams by setting the TEAM_NAME environment variable. Examples can be found in teams/teams/.

Passing in a directory will cause a random team to be selected from that directory.

The path specified should be relative to teams/teams/.

Examples

Specify a file:

TEAM_NAME=gen8/ou/clef_sand

Specify a directory:

TEAM_NAME=gen8/ou