CrownPyBot/PromptGenerator.py at master · Thebys/CrownPyBot · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
from conversation import Conversation
from enum import Enum
import random
import logging
import tiktoken
from states.emotions import Emotions
from props.eras import Eras
from props.locations import Locations
from props.characters import Characters
from props.objects import Objects
from props.scenes import Scenes
from conversation import Conversation


class PromptGenerator:
    """A class that assists in prompt construction."""

    def generate_random_props_string(self):
        number_of_eras = random.randint(0, 2)
        number_of_locations = random.randint(0, 2)
        number_of_characters = random.randint(0, 4)
        number_of_objects = random.randint(0, 4)
        result_string = ""

        for i in range(number_of_eras):
            result_string += f"{random.choice(list(Eras)).value} "
        for i in range(number_of_locations):
            result_string += f"{random.choice(list(Locations)).value} "
        for i in range(number_of_characters):
            result_string += f"{random.choice(list(Characters)).value}, "
        for i in range(number_of_objects):
            result_string += f"{random.choice(list(Objects)).value}, "

        encoding = tiktoken.encoding_for_model("gpt-3.5-turbo")
        logging.debug(
            f"AI - props weight - Pr-{len(encoding.encode(result_string))} tokens.")
        return result_string

    def with_get_single_life_anecdote(self, Conversation, Emotion, Word_Count=128, Context=None, Scene=None, Status=None) -> 'Conversation':
        Conversation.with_default_messages(Context, Scene, Status)
        Conversation.with_message(
            "user", f"Tell a shorter than {Word_Count} word {Emotion} memory, anecdote or bonmote from your life.")
        return Conversation

    def with_get_single_life_anecdote_with_random_props(self, Conversation, Emotion, Word_Count=256, Context=None, Scene=None, Status=None) -> 'Conversation':
        PG = PromptGenerator()
        random_props_string = PG.generate_random_props_string()
        Conversation.with_default_messages(Context, Scene, Status)
        Conversation.with_message(
            "user", f"Tell shorter than {Word_Count} words long {Emotion} story from you life. Adopt the following props: {random_props_string}.")
        return Conversation

    def with_get_single_vault_tec_praise(self, Conversation, Emotion, Word_Count=128, Context=None, Scene=None, Status=None) -> 'Conversation':
        Conversation.with_default_messages(Context, Scene, Status)
        Conversation.with_message(
            "user", f"You greet passer-bys in {Emotion} tone and tell a shorter than {Word_Count} word unexpected and funny Vault-Tek praise.")
        return Conversation

    def with_home_lab_smalltalk(self, Conversation, Emotion, Word_Count=None) -> 'Conversation':
        if Word_Count is None:
            Word_Count = random.randint(128, 1024)
        Conversation.with_message(
            "user", f"You are chilling with Tomáš in {Emotion} tone and tell a shorter than {Word_Count} unexpected and funny commentary on past, present, future, tech, people, the mundane and the extraordinary…")
        return Conversation

    def with_creative_introduction(self, Conversation, Emotion, Word_Count=None) -> 'Conversation':
        if Word_Count is None:
            Word_Count = random.randint(128, 1024)
        Conversation.with_message(
            "user", f"You are chilling with Tomáš in {Emotion} tone and tell a shorter than {Word_Count} unexpected and funny commentary on past, present, future, tech, people, the mundane and the extraordinary…")
        return Conversation