Leela Analysis Script

This script is a modified version of scripts originally from: https://bitbucket.org/mkmatlock/sgftools

Currently, it's designed to work with Leela 0.10.0, or 0.11.0, no guarantees about compatibility with any past or future versions. It runs on python 2.7.

Added features and modifications from the original scripts:

• Leela directly finds your game mistakes on its own and gives you alternate variations.
• Supports SGFs with handicap and komi. (NOTE: Leela only uses Chinese rules, so if your game was really in Japanese rules, in very close endgames or with certain kinds of sekis the analysis may not be correct).
• A variety of minor tweaks to the script interface and the information output to the SGF.
• Cache takes into account the search limits (Leela will not use its past results if they were done with less search time).
• Removed dependence on fcntl library - scripts now work on Windows!

WARNING: It is not uncommon for Leela to mess up on tactical situations and give poor suggestions, particularly when it hasn't realized the life or death of a group yet that is actually already alive or dead. Like all MC bots, it also has a somewhat different notion than humans of how to wrap up a won game and what moves (despite still being winning) are "mistakes". Take the analysis with many grains of salt.

How to Use

First, download and install the "engine only"/"commandline"/"GTP" version of Leela 0.10.0 from: https://sjeng.org/leela.html

Clone this repository to a local directory:

git clone https://github.com/lightvector/leela-analysis
cd leela-analysis

Then simply run the script to analyze a game, providing the command to run to the leela executable, such as ./Leela0100GTP.exe or ./leela_0100_linux_x64.

sgfanalyze.py my_game.sgf --leela /PATH/TO/LEELA.exe > my_game_analyzed.sgf

By default, Leela will go through every position in the provided game and find what it considers to be all the mistakes by both players, producing an SGF file where it highlights those mistakes and provides alternative variations it would have expected. It will probably take an hour or two to run.

Run the script with --help to see other options you can configure. You can change the amount of time Leela will analyze for, change how much effort it puts in to making variations versus just analyzing the main game, or select just a subrange of the game to analyze.

Troubleshooting

If you get an "OSError: [Errno 2] No such file or directory" error or you get an "OSError: [Errno 8] Exec format error" originating from "subprocess.py", check to make sure the command you provided for running Leela is correct. The former usually happens if you provided the wrong path, the latter if you provided the wrong Leela executable for your OS.

If get an error like "WARNING: analysis stats missing data" that causes the analysis to consistently fail at a particular spot in a given sgf file and only output partial results, there is probably a bug in the script that causes it not to be able to parse a particular output by Leela in that position. Feel free to open an issue and provide the SGF file that causes the failure. You can also run with "-v 3" to enable super-verbose output and see exactly what Leela is outputting on that position.