You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
urta=InnKeeper("Urta", "f", age=44, descr="A gruff, curvy woman with a long brown coat and bushy hair that reaches her waist. When not serving, she keeps polishing jugs with a dirty rag.", personality="She's the owner of The Prancing Llama, and of few words. But the words she speak are kind. She knows a little about all the patrons in her establishment.", short_descr="A curvy woman with long brown coat standing behind the bar.")
brim=Maid("Brim", "f", age=22, descr="A timid girl with long dark blonde hair in a braid down her back. She carries trays and dishes back and forth.", personality="She's shy and rarely looks anyone in the eye. When she speaks, it is merely a whisper. She traveled up the mountain looking for work, and ended up serving at the Inn. She dreams of big adventures.", short_descr="A timid maid with a braid wearing a tunic and dirty apron.")
that based on this message string. That will make it quite hard because you need to
643
643
parse the string again to figure out what happened... Use handle_verb / notify_action instead.
644
644
"""
645
+
#alt_prompt ="### Instruction: Location: [" + str(self.look(short=True)) + "]. Rewrite the following text in your own words using vivid language, use 'location' for context. Text:\n\n [{input_text}] \n\nEnd of text.\n\n### Response:\n"
combat_prompt=f'### Instruction: Rewrite the following combat between user {name} and {victim.title} and result into a vivid description in less than 300 words. Location: {self.location}, {self.location.short_description}. Write one to two paragraphs, ending in either death, or a stalemate. Combat Result: {attacker_msg} ### Response:\n\n'
PRE_PROMPT: 'Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n'
6
-
BASE_PROMPT: '### Instruction: Rewrite the following text in your own words using vivid language. Text:\n\n {input_text} \n\nEnd of text.\n\n### Response:\n\n'
7
-
DIALOGUE_PROMPT: 'The following is a conversation between {character1} and {character2}. {character2_description}. Chat history: {previous_conversation}\n\n### Instruction: Write a single response as {character2}, using {character2} description.\n\n### Response:\n'
8
-
ITEM_PROMPT: '### Instruction: Perform an analysis of the following text. Consider whether an item has been given, taken, dropped or put somewhere. Text:\n\n{text}\n\nEnd of text.\n\n Write your response in json format.\n\n Example {\n\n"thoughts":your thoughts on whether an item has been given, taken, dropped or put, and to whom/where.\n\n"result": {"item":"what item?", "from":"?", "to":"?"}\n\n}\n\nInsert the results from your analysis in "item", "from" and "to", or make them empty if no item was given, taken, dropped or put somewhere. Make sure the answer is valid json\n\n### Response:\n\n'
9
-
COMBAT_PROMPT: '### Instruction: Rewrite the following combat result into a vivid description. Write one to two paragraphs, ending in either death, or a stalemate. Combat Result: {result}\n\n### Response:\n\n'
11
+
BASE_PROMPT: "History: [{history}]. ### Instruction: Rewrite the following text in your own words using vivid language. 'History' can be used to create a context for what you write. Text:\n\n [{input_text}] \n\nEnd of text.\n\n### Response:\n"
12
+
DIALOGUE_PROMPT: 'The following is a conversation between {character1} and {character2}. {character2_description}. Chat history: {previous_conversation}\n\n {character2}s sentiment towards {character1}: {sentiment}. ### Instruction: Write a single response as {character2}, using {character2} description.\n\n### Response:\n'
13
+
ITEM_PROMPT: '### Instruction: Items:[{items}];Characters:[{character1},{character2}] Text:[{text}] \n\nIn the supplied text, was an item explicitly given, taken, dropped or put somewhere? Insert your thoughts about it in [my thoughts], and the results in "item", "from" and "to". Insert {character1}s sentiment towards {character2} in a single word in [sentiment assessment]. Write your response in JSON format. Example: {{ "thoughts":"[my thoughts]", "result": {{ "item":"", "from":"", "to":""}}, {{"sentiment":"[sentiment assessment]"}} }} End of example. \n\n Make sure the response is valid JSON\n\n### Response:\n'
14
+
15
+
COMBAT_PROMPT: '### Instruction: Rewrite the following combat result into a vivid description. Write one to two paragraphs, ending in either death, or a stalemate. Combat Result: {result}\n\n### Response:\n'
0 commit comments