Problem in TD Learner (trained) and a question that how to customize AI - play first at Human Mode

[albert88as]

Hi @dengl11,
I am so impressed your project and trying to learn your code. As the title, can you guide to customize AI plays first and we can customize manually AI plays in Red Side or Black Side.
Additional, at the TD learner(Trained), i faced with the problem that responsed " TypeError: Cannot read property '0' of undefined"
TypeError: Cannot read property '0' of undefined [2] at TDLearner.getValueOfState (D:\88Projects\AlphaZero\npm\ChineseChessAI\server\Strategy\TDLearner\TDLearner.js:38:53) [2] at TDLearner.ABPruning.recurseEvaluation (D:\88Projects\AlphaZero\npm\ChineseChessAI\server\Strategy\ABPruning\ABPruning.js:38:26) [2] at TDLearner.ABPruning.recurseEvaluation (D:\88Projects\AlphaZero\npm\ChineseChessAI\server\Strategy\ABPruning\ABPruning.js:46:37) [2] at TDLearner.ABPruning.recurseEvaluation (D:\88Projects\AlphaZero\npm\ChineseChessAI\server\Strategy\ABPruning\ABPruning.js:46:37) [2] at TDLearner.ABPruning.comptuteNextMove (D:\88Projects\AlphaZero\npm\ChineseChessAI\server\Strategy\ABPruning\ABPruning.js:19:31) [2] at State.nextMove (D:\88Projects\AlphaZero\npm\ChineseChessAI\server\Strategy\State\State.js:102:27) [2] at D:\88Projects\AlphaZero\npm\ChineseChessAI\server\bin\www.js:45:22 [2] at Layer.handle [as handle_request] (D:\88Projects\AlphaZero\npm\ChineseChessAI\node_modules\express\lib\router\layer.js:95:5) [2] at next (D:\88Projects\AlphaZero\npm\ChineseChessAI\node_modules\express\lib\router\route.js:137:13) [2] at Route.dispatch (D:\88Projects\AlphaZero\npm\ChineseChessAI\node_modules\express\lib\router\route.js:112:3)
Can you help me to fixed this error?
Thank you in advance!
Albert,