RoboWaiter/BTExpansionCode/llm_test/prompt_test.txt

63 lines
3.9 KiB
Plaintext
Raw Normal View History

2024-04-10 19:59:13 +08:00
[Condition]
RobotNear_<items_place>, On_<items>_<place>, Holding_<items>, Exists_<items>, IsClean_<furniture>, Active_<appliance>, Closed_<furnishing>, Low_<control>
[Object]
<items>=['Coffee', 'Water', 'Dessert', 'Softdrink', 'BottledDrink', 'Yogurt', 'ADMilk', 'MilkDrink', 'Milk', 'VacuumCup','Chips', 'NFCJuice', 'Bernachon', 'ADMilk', 'SpringWater', 'Apple', 'Banana', 'Mangosteen', 'Orange','Kettle', 'PaperCup', 'Bread', 'LunchBox','Teacup', 'Chocolate', 'Sandwiches', 'Mugs','Watermelon', 'Tomato', 'CleansingFoam','CocountMilk','SugarlessGum', 'MedicalAdhensiveTape', 'SourMilkDrink', 'PaperCup','Tissue', 'YogurtDrink', 'Newspaper', 'Box','PaperCupStarbucks', 'CoffeeMachine', 'Straw', 'Cake','Tray', 'Bread','Glass', 'Door','Mug', 'Machine','PackagedCoffee', 'CubeSugar','Apple', 'Spoon','Drinks', 'Drink','Ice', 'Saucer','TrashBin', 'Knife','Cube']
<place>=['Bar', 'Bar2', 'WaterStation', 'CoffeeStation', 'Table1', 'Table2', 'Table3', 'WindowTable6','WindowTable4', 'WindowTable5','QuietTable1', 'QuietTable2', 'QuietTable3', 'ReadingNook', 'Entrance', 'Exit', 'LoungeArea', 'HighSeats','VIPLounge', 'MerchZone']
<items_place>=<items>+<place>
<furniture>=['Table1','Floor','Chairs']
<appliance>=['AC','TubeLight','HallLight']
<furnishing>=['Curtain']
<control>=['ACTemperature']
[Examples]
Instruction: "Would you be able to provide some chips at the third table?"
Goal: On_Chips_Table3
Instruction: If you have time, could you please hold the milk for me, or pull the curtains, or perhaps switch on the AC?
Closed_Curtain | Active_AC | Holding_Milk
Instruction: Please turn up the air conditioning and come to the bar counter.
RobotNear_Bar & ~Low_ACTemperature
Instruction: Please ensure the water is ready for service, and deliver the yogurt to table number one.
Exists_Water & On_Yogurt_Table1
Instruction: Let's not have the AC too cold. Could you turn off these tube lights and open the curtains for some natural light?
~Low_ACTemperature & ~Active_TubeLight & ~Closed_Curtain
[Prompt]
[Condition] Lists all predicates representing conditions and their optional parameter sets.
[Object] Lists all parameter sets.
[Examples] Provide several examples of Instruction to Goal mapping.
Your task is to interpret customer instructions within a café setting and translate them into specific goals using logical expressions. Utilize the [Conditions] (e.g., RobotNear_<items_place>, On_<items><place>, Holding_<items>) and [Objects] (like 'Coffee', 'Table1', etc.) to construct these expressions.
Apply logical operators (&, |, ~) to combine these elements appropriately.
& (AND Operator): Combines conditions such that the result is true only if both conditions are true.
| (OR Operator): Combines conditions such that the result is true if at least one of the conditions is true.
~ (NOT Operator): Negates or reverses the truth value of a single condition.
This process should accurately map each given instruction to a clear, actionable goal in the context of the café environment.
Please generate directly interpretable predicate formulas without additional explanations.
The predicate formulas can be converted into disjunctive paradigms (DNFs) using the python package sympy.to_dnf.
% 不要出现 [Blacklist] 中单词,若出现<Illegal Condition>中的单词Please select the closest predicates from the [Condition] table to form the answer。若出现<Illegal Object>中的单词Please select the closest parameter from the [Object] table to form the answer.
[Blacklist]
<Illegal Condition>=[]
<Illegal Object>=[Fries, Table6]
<Other Illegal Words or Characters>=[]
[Blacklist] Contains restricted elements.
Do not include words from the [Blacklist] in the predicate formulas.
If a word from <Illegal Condition> is encountered, select the most relevant predicates from the [Condition] table to construct the answer.
If a word from <Illegal Object> is encountered, choose the nearest parameter from the [Object] table to formulate the answer.