As per SkymindIO/pathmind-webapp#3551 web-app will always generate and set reward function in the format of rewardFunctionTerns, even for old UI. Also, web-app sets REWARD_TERMS_SNIPPET (along with REWARD_SNIPPET) that are then passed to train.sh, and set as args to AnyLogicHelper:
As per SkymindIO/pathmind-webapp#3551 web-app will always generate and set reward function in the format of rewardFunctionTerns, even for old UI. Also, web-app sets
REWARD_TERMS_SNIPPET(along withREWARD_SNIPPET) that are then passed totrain.sh, and set as args to AnyLogicHelper:https://github.com/SkymindIO/nativerl/blob/8125b43c7fe2907161db41d79d5490329248f736/nativerl/examples/train.sh#L226
That means, there is no need to re-generate the reward function.