-
Notifications
You must be signed in to change notification settings - Fork 3
Configuration
Settings in Pantella are all usually handled in the config.json file in the root of the Pantella repository directory. The specifications that the config file is generated from are specified in the ./src/config_loader.py file, and this is where all settings are parsed and loaded.
The config.json file is the main configuration file for Pantella. It is located in the root of the Pantella repository directory. The config.json file contains all the settings for Pantella, including the game settings, the addon settings, the summarizing memory settings, the chromadb memory settings, the speech to text settings, the whisper settings, the character database settings, the language model settings, the prompt style settings, the inference options settings, the vision settings, the LLM API settings, the anthropic API settings, the OpenAI API settings, the llama_cpp_python settings, the llava_cpp_python settings, the transformers settings, the speech settings, the xTTS settings, the piperTTS settings, the xVASynth settings, the xTTS_api settings, the ChatTTS settings, the ParlerTTS settings, the StyleTTS2 settings, the F5_TTS settings, the E2_TTS settings, the Oute_TTS settings, the GPT-SoVITS settings, the debugging settings, the errors settings, and the config settings. The config.json file is a JSON file that contains a dictionary of settings. The settings in the config.json file are organized into sections, with each section containing a dictionary of settings. The settings in the config.json file are used to configure the behavior of Pantella, including the behavior of the characters in games, the behavior of the conversation manager, the behavior of the game interfaces, the behavior of the inference engines, the behavior of the memory managers, the behavior of the speech to text engines, the behavior of the thought processes, the behavior of the tokenizers, and the behavior of the text to speech engines. The settings in the config.json file can be changed to customize the behavior of Pantella to suit your needs.
We'll now break down all the available settings, sections and expected values in the config.json file.
The game settings in the config.json file are used to configure the behavior of the game interfaces. The game settings include the game ID, the conversation manager type, the interface type, the behavior manager, the memory manager, and the character manager type. The game settings are used to configure the behavior of the game interfaces, including the behavior of the conversation manager, the behavior of the game interfaces, the behavior of the behavior managers, the behavior of the memory managers, and the behavior of the character managers.
The game ID setting in the config.json file is used to specify the game that Pantella is running. The game ID setting can be set to one of the following values: skyrim, skyrimvr, fallout4, or fallout4vr. The game ID setting is used to configure the behavior of the game interfaces for the specified game.
The conversation manager type setting in the config.json file is used to specify the type of conversation manager that Pantella is using. The conversation manager type setting can be set to one of the following values: auto or set it to the slug of the conversation manager you want to use.
The interface type setting in the config.json file is used to specify the type of interface that Pantella is using. The interface type setting can be set to one of the following values: auto or set it to the slug of the interface you want to use.
The behavior manager setting in the config.json file is used to specify the behavior manager that Pantella is using. The behavior manager setting can be set to one of the following values: auto or set it to the slug of the behavior manager you want to use.
The memory manager setting in the config.json file is used to specify the memory manager that Pantella is using. The memory manager setting can be set to one of the following values: auto or set it to the slug of the memory manager you want to use.
The character manager type setting in the config.json file is used to specify the type of character manager that Pantella is using. The character manager type setting can be set to one of the following values: auto or set it to the slug of the character manager you want to use.
The addons settings in the config.json file are used to configure the behavior of the addons in Pantella. The addons settings include the disabled addons setting.
The disabled addons setting in the config.json file is used to specify the addons that are disabled in Pantella. The disabled addons setting is a list of addon names that are disabled in Pantella.
The summarizing memory settings in the config.json file are used to configure the behavior of the summarizing memory in Pantella. The summarizing memory settings include the summary limit percentage, the summarizing memory direction, and the summarizing memory depth.
The summary limit percentage setting in the config.json file is used to specify the percentage of the summarizing memory that is used for summarizing. The summary limit percentage setting is a float value between 0 and 1 that specifies the percentage of the summarizing memory that is used for summarizing.
The summarizing memory direction setting in the config.json file is used to specify the direction of the summarizing memory. The summarizing memory direction setting can be set to one of the following values: topdown or bottomup.
The summarizing memory depth setting in the config.json file is used to specify the depth of the summarizing memory. The summarizing memory depth setting is an integer value that specifies the depth of the summarizing memory.
The chromadb memory settings in the config.json file are used to configure the behavior of the chromadb memory in Pantella. The chromadb memory settings include the memory update interval, the logical memories, the emotional memories, the torchmoji max length, the empathy, the chromadb memory messages before, the chromadb memory messages after, the emotional decay min, the emotional decay max, the emotion composition, the chromadb memory depth, the chromadb memory direction, the chromadb query size, and the chromadb memory editor enabled.
The memory update interval setting in the config.json file is used to specify the interval at which the chromadb memory is updated. The memory update interval setting is an integer value that specifies the interval at which the chromadb memory is updated.
The logical memories setting in the config.json file is used to specify the number of logical memories in the chromadb memory. The logical memories setting is an integer value that specifies the number of logical memories in the chromadb memory.
The emotional memories setting in the config.json file is used to specify the number of emotional memories in the chromadb memory. The emotional memories setting is an integer value that specifies the number of emotional memories in the chromadb memory.
The torchmoji max length setting in the config.json file is used to specify the maximum length of the torchmoji in the chromadb memory. The torchmoji max length setting is an integer value that specifies the maximum length of the torchmoji in the chromadb memory.
The empathy setting in the config.json file is used to specify the empathy in the chromadb memory. The empathy setting is a float value between 0 and 1 that specifies the empathy in the chromadb memory. The empathy setting is used to configure the behavior of the chromadb memory in Pantella. Empathy is a measure of how much the emotion of other characters messages affects the emotion of the character.
The chromadb memory messages before setting in the config.json file is used to specify the number of messages before the current message that are stored in the chromadb memory. The chromadb memory messages before setting is an integer value that specifies the number of messages before the current message that are stored in the chromadb memory.
The chromadb memory messages after setting in the config.json file is used to specify the number of messages after the current message that are stored in the chromadb memory. The chromadb memory messages after setting is an integer value that specifies the number of messages after the current message that are stored in the chromadb memory.
The emotional decay min setting in the config.json file is used to specify the minimum emotional decay in the chromadb memory. The emotional decay min setting is a float value between 0 and 1 that specifies the minimum emotional decay in the chromadb memory.
The emotional decay max setting in the config.json file is used to specify the maximum emotional decay in the chromadb memory. The emotional decay max setting is a float value between 0 and 1 that specifies the maximum emotional decay in the chromadb memory.
The emotion composition setting in the config.json file is used to specify the composition of emotions in the chromadb memory. The emotion composition setting is a dictionary of emotions and their corresponding emojis that are used to compose the emotions in the chromadb memory.
The chromadb memory depth setting in the config.json file is used to specify the depth of the chromadb memory. The chromadb memory depth setting is an integer value that specifies the depth of the chromadb memory.
The chromadb memory direction setting in the config.json file is used to specify the direction of the chromadb memory. The chromadb memory direction setting can be set to one of the following values: topdown or bottomup.
The chromadb query size setting in the config.json file is used to specify the size of the chromadb query. The chromadb query size setting is an integer value that specifies the size of the chromadb query. The chromadb query size setting is used to configure the behavior of the chromadb memory in Pantella.
The chromadb memory editor enabled setting in the config.json file is used to specify whether the chromadb memory editor is enabled. The chromadb memory editor is a gradio interface that allows you to edit the chromadb memory.
The speech to text settings in the config.json file are used to configure the behavior of the speech to text engines in Pantella. The speech to text settings include the speech to text enabled setting, the speech to text engine setting, the speech to text language setting, the audio threshold setting, the pause threshold setting, and the listen timeout setting.
The speech to text enabled setting in the config.json file is used to specify whether the speech to text engine is enabled. The speech to text enabled setting is a boolean value that specifies whether the speech to text engine is enabled.
The speech to text engine setting in the config.json file is used to specify the speech to text engine that Pantella is using. The speech to text engine setting can be set to one of the following values: faster_whisper or openai_whisper.
The speech to text language setting in the config.json file is used to specify the language of the speech to text engine. The speech to text language setting is a string value that specifies the language of the speech to text engine.
The audio threshold setting in the config.json file is used to specify the audio threshold of the speech to text engine. The audio threshold setting is a string value that specifies the audio threshold of the speech to text engine.
The pause threshold setting in the config.json file is used to specify the pause threshold of the speech to text engine. The pause threshold setting is a float value that specifies the pause threshold of the speech to text engine.
The listen timeout setting in the config.json file is used to specify the listen timeout of the speech to text engine. The listen timeout setting is an integer value that specifies the listen timeout of the speech to text engine.
The whisper settings in the config.json file are used to configure the behavior of the whisper engine in Pantella. The whisper settings include the whisper model setting.
The whisper model setting in the config.json file is used to specify the whisper model that Pantella is using. The whisper model setting is a string value that specifies the whisper model that Pantella is using.
The faster whisper settings in the config.json file are used to configure the behavior of the faster whisper engine in Pantella. The faster whisper settings include the speech to text translate setting, the whisper process device setting, the whisper CPU threads setting, the whisper compute type setting, the beam size setting, and the VAD filter setting.
The speech to text translate setting in the config.json file is used to specify whether the speech to text engine translates the speech to text. The speech to text translate setting is a boolean value that specifies whether the speech to text engine translates the speech to text.
The whisper process device setting in the config.json file is used to specify the device that the whisper engine is using. The whisper process device setting can be set to one of the following values: cpu or cuda.
The whisper CPU threads setting in the config.json file is used to specify the number of CPU threads that the whisper engine is using. The whisper CPU threads setting is an integer value that specifies the number of CPU threads that the whisper engine is using. The default value is 4.
The whisper compute type setting in the config.json file is used to specify the compute type of the whisper engine. The whisper compute type setting can be set to one of the following values: auto or cuda.
The beam size setting in the config.json file is used to specify the beam size of the whisper engine. The beam size setting is an integer value that specifies the beam size of the whisper engine. The default value is 5.
The VAD filter setting in the config.json file is used to specify whether the whisper engine uses a VAD filter. The VAD filter setting is a boolean value that specifies whether the whisper engine uses a VAD filter. The default value is True.
The OpenAI whisper settings in the config.json file are used to configure the behavior of the OpenAI whisper engine in Pantella. The OpenAI whisper settings include the whisper URL setting.
The whisper URL setting in the config.json file is used to specify the URL of the whisper engine. The whisper URL setting is a string value that specifies the URL of the whisper engine.
The character database settings in the config.json file are used to configure the behavior of the character database in Pantella. The character database settings include the allow name matching setting, the allow ID matching setting, the allow exact base ID matching setting, the allow greedy base ID matching setting, the override voice model with simple predictions setting, and the auto save generated characters setting.
The allow name matching setting in the config.json file is used to specify whether the character database allows name matching. The allow name matching setting is a boolean value that specifies whether the character database allows name matching. The default value is True.
The allow ID matching setting in the config.json file is used to specify whether the character database allows ID matching. The allow ID matching setting is a boolean value that specifies whether the character database allows ID matching. The default value is True.
The allow exact base ID matching setting in the config.json file is used to specify whether the character database allows exact base ID matching. The allow exact base ID matching setting is a boolean value that specifies whether the character database allows exact base ID matching. The default value is True.
The allow greedy base ID matching setting in the config.json file is used to specify whether the character database allows greedy base ID matching. The allow greedy base ID matching setting is a boolean value that specifies whether the character database allows greedy base ID matching. The default value is True.
The override voice model with simple predictions setting in the config.json file is used to specify whether the character database overrides the voice model with simple predictions. The override voice model with simple predictions setting is a boolean value that specifies whether the character database overrides the voice model with simple predictions. The default value is False.
The auto save generated characters setting in the config.json file is used to specify whether the character database auto saves generated characters. The auto save generated characters setting is a boolean value that specifies whether the character database auto saves generated characters. The default value is True.
The language model settings in the config.json file are used to configure the behavior of the language model in Pantella. The language model settings include the language model type setting, the language model model setting, the language model device setting, the language model temperature setting, the language model top_k setting, the language model top_p setting, the language model length penalty setting, the language model repetition penalty setting, the language model speed setting, the language model num beams setting, the language model use cached latents setting, the language model use preloaded latents setting, the language model use text splitting setting, the language model stream chunk size setting, the language model banned voice models setting, and the language model default model setting.
The inference_engine setting is used to specify the inference engine to use for the language model. You can read more about the inference engines over here. The inference_engine setting can be set to one of the following values:
-
default: Use the default inference engine for the language model. This isopenai_apiby default. -
openai_api: Use the OpenAI API Standard for inference. This is what you use for interfacing with OpenAI and OpenRouter. -
anthropic_api: Use the Anthropic API Standard for inference. This is what you use for interfacing with Anthropic. -
llama_cpp_python: Use the Llama C++ Python API for inference. This is what you use for interfacing with Llama and Llava locally without a server. -
transformers: Use the Hugging Face Transformers API for inference. This is what you use for interfacing with Hugging Face Transformers. This is what you use for interfacing with Transformer models without a server.
The tokenizer_type setting is used to specify the tokenizer type to use for the language model. You can read more about the tokenizer types over here. The tokenizer_type setting can be set to one of the following values:
-
default: Use the default tokenizer for the language model. This is variable based on the model, and it's recommended to leave this at default unless you know what you're doing. -
tiktoken: Use the TikToken tokenizer for the language model. This is what you use for interfacing with TikToken. -
embeddingThis is a sketchy tokenizer that can be used for remote models that don't expose their tokenizer. It's not recommended to use this unless you know what you're doing. It's much slower than the other tokenizers and doesn't always work. -
koboldcpp: Use the KoboldCPP API for tokenization. This is what you use for interfacing with KoboldCPP locally or via Google Colab. -
huggingface: Use a tokenizer from Huggingface.
The maximum_local_tokens setting is used to specify the maximum number of tokens to generate locally before sending the output to the game. This setting is used to prevent the language model from generating too many tokens locally and causing performance issues. The maximum_local_tokens setting can be set to any positive integer value. The default value is 8192.
The max_response_sentences setting is used to specify the maximum number of sentences to generate in a single response. This setting is used to prevent the language model from generating too many sentences in a single response. The max_response_sentences setting can be set to any positive integer value. The default value is 999 which is effectively unlimited.
The wait_time_buffer setting is used to specify the amount of time to wait before sending the next message to the game. This setting is used to prevent the language model from generating messages too quickly and causing performance issues. The wait_time_buffer setting can be set to any positive float value. The default value is 1.0.
The same_output_limit setting is used to specify the maximum number of times the language model can generate the same output in a row before changing the input. This setting is used to prevent the language model from getting stuck in a loop generating the same output. The same_output_limit setting can be set to any positive integer value. The default value is 30.
The conversation_limit_pct setting is used to specify the percentage of the context the conversation has to reach before it clears it and reloads. This is used when estimating token counts. The conversation_limit_pct setting can be set to any positive float value. The default value is 0.9. If you have an exact tokenizer, you can (probably) set this to 1.0.
The min_conversation_length setting is used to specify the minimum number of messages in a conversation before it clears the context and reloads. This is used when estimating token counts. The min_conversation_length setting can be set to any positive integer value. The default value is 5.
The reload_buffer setting is used to specify the number of messages to keep in the context before clearing it and reloading. This is used when estimating token counts. The reload_buffer setting can be set to any positive integer value. The default value is 20.
The prompt_style setting is used to specify the prompt style to use for the conversation manager. The prompt_style setting can be set to any valid prompt style. The default value is default. The default is normal_en which is the default prompt style for the English language. This can be set to the prompt_style slug of any prompt style you have installed. See more about prompt styles over here.
The behavior_style setting is used to specify the behavior style to use for the conversation manager. The behavior_style setting can be set to any valid behavior style. The default value is normal.
The conversation_start_type setting is used to specify the conversation start type to use for the conversation manager. The conversation_start_type setting can be set to any valid conversation start type. The default value is force_npc_greeting_for_first_meeting_then_llm_choice. This can be one of the following values:
-
always_llm_choice: Always let the LLM choose the next message. -
always_force_npc_greeting: Always force the NPC to greet the player. -
always_player_greeting: Always force the player to greet the NPC. -
implicit_predetermined_player_greeting: Use a predetermined player greeting if the player doesn't greet the NPC. -
predetermined_npc_greeting: Use a predetermined NPC greeting if the NPC doesn't greet the player. -
predetermined_npc_greeting_for_first_meeting_then_llm_choice: Use a predetermined NPC greeting for the first meeting, then let the LLM choose the next message. -
force_npc_greeting_for_first_meeting_then_llm_choice: Force the NPC to greet the player for the first meeting, then let the LLM choose the next message.
The strip_smalls setting is used to specify whether to strip small messages. The strip_smalls setting can be set to True or False. The default value is True.
The small_size setting is used to specify the length of a small message. The small_size setting can be set to any positive integer value. The default value is 3.
The assure_grammar setting is used to specify whether to assure grammar in the messages. The assure_grammar setting can be set to True or False. The default value is True.
The assist_check setting is used to specify whether to check for assistance in the messages. The assist_check setting can be set to True or False. The default value is False.
The break_on_time_announcements setting is used to specify whether to break on time announcements. The break_on_time_announcements setting can be set to True or False. The default value is True.
The as_a_check setting is used to specify whether to check for "as a" in the messages. The as_a_check setting can be set to True or False. The default value is False.
The meet_string_game_events setting is used to specify whether to meet string game events. The meet_string_game_events setting can be set to True or False. The default value is False.
The message_reformatting setting is used to specify whether to reformat messages. The message_reformatting setting can be set to True or False. The default value is False.
The game_update_pruning setting is used to specify whether to prune game updates. The game_update_pruning setting can be set to True or False. The default value is True.
The game_update_prune_count setting is used to specify the number of game updates to prune. The game_update_prune_count setting can be set to any positive integer value. The default value is 5.
The conversation_start_role setting is used to specify the role to start the conversation. The conversation_start_role setting can be set to system, user, or assistant. The default value is system.
The start_as_stranger setting is used to specify whether to start as a stranger. The start_as_stranger setting can be set to True or False. The default value is True.
The custom_possible_player_aliases setting is used to specify additional names the player goes by. The custom_possible_player_aliases setting can be set to a list of strings. The default value is [].
The InferenceOptions settings are used to specify the inference options for the conversation manager.
The thought_type setting is used to specify the thought process type to use for the conversation manager. The thought_type setting can be set to any valid thought process slug. The default value is simple. This can be one of the following values:
-
basic_planner: Use a basic planner for the thought process. -
branching: Use branching for the thought process. -
combined_theory_of_mind: Use a combined theory of mind for the thought process. -
freudian: Use a freudian thought process. -
jungian: Use a jungian thought process. -
simple: Use a simple thought process.
The character_type setting is used to specify the character type to use for the conversation manager. The character_type setting can be set to any valid character type slug. The default value is auto. This can be one of the following values:
-
auto: Automatically determine the character type. -
skyrim: Use the Skyrim character type.
The cot_enabled setting is used to specify whether the CoT is enabled. The cot_enabled setting can be set to True or False. The default value is False. You can read more about the CoT here.
The temperature setting is used to specify the temperature for the conversation manager. The temperature setting can be set to any float value. The default value is 0.8.
The top_p setting is used to specify the top p value for the conversation manager. The top_p setting can be set to any float value. The default value is 1.
The min_p setting is used to specify the min p value for the conversation manager. The min_p setting can be set to any float value. The default value is 0.05.
The typical_p setting is used to specify the typical p value for the conversation manager. The typical_p setting can be set to any float value. The default value is 0.9.
The top_k setting is used to specify the top k value for the conversation manager. The top_k setting can be set to any positive integer value. The default value is 0.
The repeat_penalty setting is used to specify the repeat penalty for the conversation manager. The repeat_penalty setting can be set to any float value. The default value is 1.0.
The tfs_z setting is used to specify the tfs z value for the conversation manager. The tfs_z setting can be set to any float value. The default value is 1.0.
The frequency_penalty setting is used to specify the frequency penalty for the conversation manager. The frequency_penalty setting can be set to any float value. The default value is 0.
The presence_penalty setting is used to specify the presence penalty for the conversation manager. The presence_penalty setting can be set to any float value. The default value is 0.
The mirostat_mode setting is used to specify the mirostat mode for the conversation manager. The mirostat_mode setting can be set to any integer value. The default value is 0.
The mirostat_eta setting is used to specify the mirostat eta value for the conversation manager. The mirostat_eta setting can be set to any float value. The default value is 0.1.
The mirostat_tau setting is used to specify the mirostat tau value for the conversation manager. The mirostat_tau setting can be set to any integer value. The default value is 5.
The max_tokens setting is used to specify the max tokens value for the conversation manager. The max_tokens setting can be set to any positive integer value. The default value is 512.
The logit_bias setting is used to specify the logit bias for the conversation manager. The logit_bias setting can be set to a dictionary of logit bias values. The default value is {}.
The Vision settings are used to specify the vision options for vision enabled inference engines/models.
The vision_enabled setting is used to specify whether vision is enabled. The vision_enabled setting can be set to True or False. The default value is False.
The ocr_lang setting is used to specify the OCR language to use for the vision engine. The ocr_lang setting can be set to any valid OCR language. The default value is en.
The ocr_use_angle_cls setting is used to specify whether to use angle classification for OCR. The ocr_use_angle_cls setting can be set to True or False. The default value is True.
The ocr_filter setting is used to specify the OCR filter for the vision engine. The ocr_filter setting can be set to a list of strings. There are a bunch of default values already, too many to list here. So the default is not [] but it could be!
The append_system_image_near_end setting is used to specify whether to append the system image near the end. The append_system_image_near_end setting can be set to True or False. The default value is True.
The image_message_depth setting is used to specify the image message depth for the vision engine. The image_message_depth setting can be set to any integer value. The default value is -1.
The image_message setting is used to specify the image message for the vision engine. The image_message setting can be set to any valid image message. The default value is The image below is what {player_perspective_name}'s can see from their perspective:\n{image}\n{ocr}.
The paddle_ocr setting is used to specify whether to use paddle OCR. The paddle_ocr setting can be set to True or False. The default value is True.
The ocr_resolution setting is used to specify the OCR resolution for the vision engine. The ocr_resolution setting can be set to any integer value. The default value is 256.
The resize_image setting is used to specify whether to resize the image. The resize_image setting can be set to True or False. The default value is True.
The image_resolution setting is used to specify the image resolution for the vision engine. The image_resolution setting can be set to any integer value. The default value is 672.
The resize_image setting is used to specify whether to resize the image. The resize_image setting can be set to True or False. The default value is True.
The LLM_API_Settings settings are used to specify the LLM API options for LLM enabled inference engines/models.
The log_all_api_requests setting is used to specify whether to log all API requests to a local directory for debugging purposes. The log_all_api_requests setting can be set to True or False. The default value is False.
The reverse_proxy setting is used to specify whether to use a reverse proxy. The reverse_proxy setting can be set to True or False. The default value is False.
The anthropic_api settings are used to specify the Anthropic API options for Anthropic enabled inference engines/models.
The anthropic_model setting is used to specify the Anthropic model to use. The anthropic_model setting can be set to any valid Anthropic model. The default value is undi95/toppy-m-7b:free.
The alternative_anthropic_api_base setting is used to specify the alternative Anthropic API base URL to use. The alternative_anthropic_api_base setting can be set to any valid Anthropic API base URL. The default value is none.
The anthropic_api_key_path setting is used to specify the path to the Anthropic API key file. The anthropic_api_key_path setting can be set to any valid file path. The default value is .\\ANTHROPIC_SECRET_KEY.txt.
The openai_api settings are used to specify the OpenAI API options for OpenAI enabled inference engines/models.
The openai_model setting is used to specify the model slug to use. The openai_model setting can be set to any valid OpenAI model. The default value is undi95/toppy-m-7b:free.
The openai_character_generator_model setting is used to specify the character generator model to use. The openai_character_generator_model setting can be set to any valid model slug. The default value is an empty string.
The openai_completions_type setting is used to specify the completions type to use. The openai_completions_type setting can be set to text or chat. The default value is text.
The alternative_openai_api_base setting is used to specify the alternative OpenAI API base URL to use. The alternative_openai_api_base setting can be set to any valid OpenAI API base URL. The default value is https://openrouter.ai/api/v1/.
The openai_api_key_path setting is used to specify the path to the OpenAI API key file. The openai_api_key_path setting can be set to any valid file path. The default value is .\\GPT_SECRET_KEY.txt.
The banned_samplers setting is used to specify the banned samplers for the OpenAI API. The banned_samplers setting can be set to a list of banned samplers. The default value is an empty list.
The api_log_dir setting is used to specify the API log directory for the OpenAI API. The api_log_dir setting can be set to any valid directory path. The default value is .\\api_logs.
The llama_cpp_python settings are used to specify the llama-cpp-python options for llama-cpp-python enabled inference engines/models.
The model_path setting is used to specify the path to the model file. The model_path setting can be set to any valid model file path. The default value is .\\model.gguf.
The n_gpu_layers setting is used to specify the number of GPU layers to use. The n_gpu_layers setting can be set to any integer value. The default value is 0 (no GPU layers). This can also be set to 99 to fully offload to the GPU.
The n_threads setting is used to specify the number of threads to use. The n_threads setting can be set to any integer value. The default value is 4. If you're fully offloaded, you can set this to 1.
The n_batch setting is used to specify the sie of the batch of tokens parsed by the model. The n_batch setting can be set to any integer value. The default value is 512.
The tensor_split setting is used to specify how the model is split across multiple GPUs. The tensor_split setting can be set to a list of floats. The default value is an empty list.
The main_gpu setting is used to specify the main GPU to use. The main_gpu setting can be set to any integer value. The default value is 0 for the first GPU.
The split_mode setting is used to specify the split mode to use. The split_mode setting can be set to 0, 1, or 2. The default value is 0.
The use_mmap setting is used to specify whether to use mmap. The use_mmap setting can be set to True or False. The default value is True.
The use_mlock setting is used to specify whether to use mlock. The use_mlock setting can be set to True or False. The default value is False.
The n_threads_batch setting is used to specify the number of threads per batch. The n_threads_batch setting can be set to any integer value. The default value is 1.
The offload_kqv setting is used to specify whether to offload kqv. The offload_kqv setting can be set to True or False. The default value is True.
The llava_cpp_python settings are used to specify the llava-cpp-python options for llava-cpp-python enabled inference engines/models.
The llava_clip_model_path setting is used to specify the path to the llava-clip model file. The llava_clip_model_path setting can be set to any valid model file path. The default value is .\\clip_model.gguf.
The transformers settings are used to specify the Transformers options for Transformers enabled inference engines/models.
The transformers_model_slug setting is used to specify the model slug to use. The transformers_model_slug setting can be set to any valid Transformers model. The default value is mistralai/Mistral-7B-Instruct-v0.1.
The trust_remote_code setting is used to specify whether to trust remote code. The trust_remote_code setting can be set to True or False. The default value is False.
The device_map setting is used to specify the device map to use. The device_map setting can be set to cuda, cuda:0, cuda:1, or auto. The default value is cuda:0.
The load_in_8bit setting is used to specify whether to load in 8bit. The load_in_8bit setting can be set to True or False. The default value is False.
The Speech settings are used to specify the speech options for how to generate audio dialogue for characters.
The tts_engine setting is used to specify the TTS engine to use. The tts_engine setting can be set to a list of TTS engines or a single TTS engine. The default value is ["piper_binary"]. If you use a list of multiple TTS they will be used in a fallback order. Individual characters can be whitelisted/blacklisted from using certain TTS engines as well.
Available TTS engines are:
piper_binaryxvasynthxtts_apichat_ttsparler_ttsstyle_tts_2f5_ttse2_ttsoute_ttsGPT-SoVITS
You can read more about each TTS in the Available Text To Speech Engines section.
The end_conversation_wait_time setting is used to specify the wait time at the end of a conversation. The end_conversation_wait_time setting can be set to any integer value. The default value is 1.
The sentences_per_voiceline setting is used to specify the number of sentences per voiceline. The sentences_per_voiceline setting can be set to any integer value. The default value is 2.
The narrator_voice setting is used to specify the narrator voice to use. The narrator_voice setting can be set to any valid narrator voice. The default value is MaleKhajiit.
The narrator_volume setting is used to specify the narrator volume to use. The narrator_volume setting can be set to any float value. The default value is 0.5.
The narrator_delay setting is used to specify the narrator delay to use. The narrator_delay setting can be set to any float value. The default value is 0.2.
The xTTS settings are used to specify the xTTS options for xTTS enabled inference engines/models.
The xtts_device setting is used to specify the xTTS device to use. The xtts_device setting can be set to any valid xTTS device. The default value is cuda.
The xtts_voice_samples_dir setting is used to specify the path to the xTTS voice samples directory. The xtts_voice_samples_dir setting can be set to any valid directory path. The default value is .\\data\\voice_samples.
The xtts_preload_latents setting is used to specify whether to preload latents. The xtts_preload_latents setting can be set to True or False. The default value is True.
The xtts_use_cached_latents setting is used to specify whether to use cached latents. The xtts_use_cached_latents setting can be set to True or False. The default value is True.
The xtts_temperature setting is used to specify the xTTS temperature to use. The xtts_temperature setting can be set to any float value. The default value is 0.75.
The xtts_top_k setting is used to specify the xTTS top k to use. The xtts_top_k setting can be set to any integer value. The default value is 50.
The xtts_top_p setting is used to specify the xTTS top p to use. The xtts_top_p setting can be set to any float value. The default value is 0.85.
The xtts_length_penalty setting is used to specify the xTTS length penalty to use. The xtts_length_penalty setting can be set to any float value. The default value is 1.0.
The xtts_repetition_penalty setting is used to specify the xTTS repetition penalty to use. The xtts_repetition_penalty setting can be set to any float value. The default value is 10.0.
The xtts_speed setting is used to specify the xTTS speed to use. The xtts_speed setting can be set to any float value. The default value is 1.0.
The xtts_num_beams setting is used to specify the xTTS number of beams to use. The xtts_num_beams setting can be set to any integer value. The default value is 1.
The piperTTS settings are used to specify the PiperTTS options for PiperTTS enabled inference engines/models.
The piper_binary_dir setting is used to specify the path to the Piper binary directory. The piper_binary_dir setting can be set to any valid directory path. The default value is .\\piper\\.
The piper_models_dir setting is used to specify the path to the Piper models directory. The piper_models_dir setting can be set to any valid directory path. The default value is .\\data\\models\\piper\\.
The piper_tts_banned_voice_models setting is used to specify the banned voice models for PiperTTS. The piper_tts_banned_voice_models setting can be set to a list of banned voice models. The default value is an empty list.
The xVASynth settings are used to specify the xVASynth options for xVASynth enabled inference engines/models.
The xvasynth_path setting is used to specify the path to the xVASynth directory. The xvasynth_path setting can be set to any valid directory path. The default value is C:\\Games\\Steam\\steamapps\\common\\xVASynth.
The xvasynth_process_device setting is used to specify the xVASynth process device to use. The xvasynth_process_device setting can be set to cpu or cuda. The default value is cpu.
The pace setting is used to specify the pace to use. The pace setting can be set to any float value. The default value is 1.0.
The use_cleanup setting is used to specify whether to use cleanup. The use_cleanup setting can be set to True or False. The default value is False.
The use_sr setting is used to specify whether to use SR. The use_sr setting can be set to True or False. The default value is False.
The xvasynth_base_url setting is used to specify the xVASynth base URL to use. The xvasynth_base_url setting can be set to any valid xVASynth base URL. The default value is http://127.0.0.1:8008.
The xvasynth_banned_voice_models setting is used to specify the banned voice models for xVASynth. The xvasynth_banned_voice_models setting can be set to a list of banned voice models. The default value is an empty list.
The xTTS_api settings are used to specify the xTTS API options for xTTS API enabled inference engines/models.
The xtts_api_dir setting is used to specify the path to the xTTS API directory. The xtts_api_dir setting can be set to any valid directory path. The default value is .\\xtts-api-server-pantella\\.
The xtts_api_base_url setting is used to specify the xTTS API base URL to use. The xtts_api_base_url setting can be set to any valid xTTS API base URL. The default value is http://127.0.0.1:8020.
The xtts_api_data setting is used to specify the xTTS API data to use. The xtts_api_data setting can be set to any valid xTTS API data. The default value is a dictionary with the following keys and values:
-
temperature:0.75 -
length_penalty:1.0 -
repetition_penalty:3.0 -
top_k:40 -
top_p:0.80 -
speed:1.25 -
enable_text_splitting:True -
stream_chunk_size:200
The xtts_api_banned_voice_models setting is used to specify the banned voice models for xTTS API. The xtts_api_banned_voice_models setting can be set to a list of banned voice models. The default value is an empty list.
The default_xtts_api_model setting is used to specify the default xTTS API model to use. The default_xtts_api_model setting can be set to any valid xTTS API model. The default value is v2.0.2.
The xtts_banned_voice_models setting is used to specify the banned voice models for xTTS. The xtts_banned_voice_models setting can be set to a list of banned voice models. The default value is an empty list.
The ChatTTS settings are used to specify the ChatTTS options for ChatTTS enabled inference engines/models.
The ensure_all_voice_samples_have_inference_settings setting is used to specify whether to ensure all voice samples have inference settings. The ensure_all_voice_samples_have_inference_settings setting can be set to True or False. The default value is True.
The chat_tts_default_infer_code_prompt setting is used to specify the default infer code prompt to use. The chat_tts_default_infer_code_prompt setting can be set to any valid infer code prompt. The default value is [speed_3].
The chat_tts_default_infer_code_temperature setting is used to specify the default infer code temperature to use. The chat_tts_default_infer_code_temperature setting can be set to any float value. The default value is 0.3.
The chat_tts_default_infer_code_repetition_penalty setting is used to specify the default infer code repetition penalty to use. The chat_tts_default_infer_code_repetition_penalty setting can be set to any float value. The default value is 1.05.
The chat_tts_default_refine_text_prompt setting is used to specify the default refine text prompt to use. The chat_tts_default_refine_text_prompt setting can be set to any valid refine text prompt. The default value is an empty string.
The chat_tts_default_refine_text_temperature setting is used to specify the default refine text temperature to use. The chat_tts_default_refine_text_temperature setting can be set to any float value. The default value is 0.7.
The chat_tts_default_refine_text_top_p setting is used to specify the default refine text top p to use. The chat_tts_default_refine_text_top_p setting can be set to any float value. The default value is 0.7.
The chat_tts_default_refine_text_top_k setting is used to specify the default refine text top k to use. The chat_tts_default_refine_text_top_k setting can be set to any integer value. The default value is 20.
The chat_tts_default_refine_text_repetition_penalty setting is used to specify the default refine text repetition penalty to use. The chat_tts_default_refine_text_repetition_penalty setting can be set to any float value. The default value is 1.0.
The chat_tts_banned_voice_models setting is used to specify the banned voice models for ChatTTS. The chat_tts_banned_voice_models setting can be set to a list of banned voice models. The default value is an empty list.
The ParlerTTS settings are used to specify the ParlerTTS options for ParlerTTS enabled inference engines/models.
The parler_tts_model setting is used to specify the ParlerTTS model to use. The parler_tts_model setting can be set to any valid ParlerTTS model. The default value is parler-tts/parler-tts-mini-v1.
The parler_tts_max_length setting is used to specify the ParlerTTS max length to use. The parler_tts_max_length setting can be set to any integer value. The default value is 4096.
The parler_tts_temperature setting is used to specify the ParlerTTS temperature to use. The parler_tts_temperature setting can be set to any float value. The default value is 0.1.
The parler_tts_repetition_penalty setting is used to specify the ParlerTTS repetition penalty to use. The parler_tts_repetition_penalty setting can be set to any float value. The default value is 1.1.
The parler_tts_banned_voice_models setting is used to specify the banned voice models for ParlerTTS. The parler_tts_banned_voice_models setting can be set to a list of banned voice models. The default value is an empty list.
The StyleTTS_2 settings are used to specify the StyleTTS 2 options for StyleTTS 2 enabled inference engines/models.
The style_tts_2_model setting is used to specify the StyleTTS 2 model to use. The style_tts_2_model setting can be set to any valid StyleTTS 2 model. The default value is styletts2/styletts2-ada.
The style_tts_2_temperature setting is used to specify the StyleTTS 2 temperature to use. The style_tts_2_temperature setting can be set to any float value. The default value is 0.7.
The style_tts_2_top_p setting is used to specify the StyleTTS 2 top p to use. The style_tts_2_top_p setting can be set to any float value. The default value is 0.7.
The style_tts_2_top_k setting is used to specify the StyleTTS 2 top k to use. The style_tts_2_top_k setting can be set to any integer value. The default value is 20.
The style_tts_2_repetition_penalty setting is used to specify the StyleTTS 2 repetition penalty to use. The style_tts_2_repetition_penalty setting can be set to any float value. The default value is 1.0.
The style_tts_2_banned_voice_models setting is used to specify the banned voice models for StyleTTS 2. The style_tts_2_banned_voice_models setting can be set to a list of banned voice models. The default value is an empty list.
The F5_TTS settings are used to specify the F5 TTS options for F5 TTS enabled inference engines/models.
The f5_tts_volume setting is used to specify the F5 TTS volume to use. The f5_tts_volume setting can be set to any float value. The default value is 1.5.
The f5_tts_device setting is used to specify the F5 TTS device to use. The f5_tts_device setting can be set to any valid F5 TTS device. The default value is cuda.
The f5_tts_banned_voice_models setting is used to specify the banned voice models for F5 TTS. The f5_tts_banned_voice_models setting can be set to a list of banned voice models. The default value is an empty list.
The E2_TTS settings are used to specify the E2 TTS options for E2 TTS enabled inference engines/models.
The e2_tts_volume setting is used to specify the E2 TTS volume to use. The e2_tts_volume setting can be set to any float value. The default value is 1.5.
The e2_tts_banned_voice_models setting is used to specify the banned voice models for E2 TTS. The e2_tts_banned_voice_models setting can be set to a list of banned voice models. The default value is an empty list.
The Oute_TTS settings are used to specify the Oute TTS options for Oute TTS enabled inference engines/models.
The oute_tts_temperature setting is used to specify the Oute TTS temperature to use. The oute_tts_temperature setting can be set to any float value. The default value is 0.1.
The oute_tts_repetition_penalty setting is used to specify the Oute TTS repetition penalty to use. The oute_tts_repetition_penalty setting can be set to any float value. The default value is 1.1.
The oute_tts_max_length setting is used to specify the Oute TTS max length to use. The oute_tts_max_length setting can be set to any integer value. The default value is 4096.
The oute_tts_volume setting is used to specify the Oute TTS volume to use. The oute_tts_volume setting can be set to any float value. The default value is 1.5.
The oute_tts_banned_voice_models setting is used to specify the banned voice models for Oute TTS. The oute_tts_banned_voice_models setting can be set to a list of banned voice models. The default value is an empty list.
The GPT-SoVITS settings are used to specify the GPT-SoVITS options for GPT-SoVITS enabled inference engines/models.
The gpt_sovits_is_half setting is used to specify whether to use half precision for GPT-SoVITS. The gpt_sovits_is_half setting can be set to True or False. The default value is True.
The gpt_sovits_device setting is used to specify the GPT-SoVITS device to use. The gpt_sovits_device setting can be set to any valid GPT-SoVITS device. The default value is cuda.
The gpt_sovits_prompt_language setting is used to specify the GPT-SoVITS prompt language to use. The gpt_sovits_prompt_language setting can be set to any valid GPT-SoVITS prompt language. The default value is en.
The gpt_sovits_text_language setting is used to specify the GPT-SoVITS text language to use. The gpt_sovits_text_language setting can be set to any valid GPT-SoVITS text language. The default value is en.
The gpt_sovits_version setting is used to specify the GPT-SoVITS version to use. The gpt_sovits_version setting can be set to any valid GPT-SoVITS version. The default value is v2.
The gpt_sovits_cut_type setting is used to specify the GPT-SoVITS cut type to use. The gpt_sovits_cut_type setting can be set to any valid GPT-SoVITS cut type. The default value is none.
The gpt_sovits_temperature setting is used to specify the GPT-SoVITS temperature to use. The gpt_sovits_temperature setting can be set to any float value. The default value is 1.0.
The gpt_sovits_top_k setting is used to specify the GPT-SoVITS top k to use. The gpt_sovits_top_k setting can be set to any integer value. The default value is 20.
The gpt_sovits_top_p setting is used to specify the GPT-SoVITS top p to use. The gpt_sovits_top_p setting can be set to any float value. The default value is 1.0.
The gpt_sovits_banned_voice_models setting is used to specify the banned voice models for GPT-SoVITS. The gpt_sovits_banned_voice_models setting can be set to a list of banned voice models. The default value is an empty list.
The gpt_sovits_error_on_too_short_or_too_long_audio setting is used to specify whether to error on too short or too long audio for GPT-SoVITS. The gpt_sovits_error_on_too_short_or_too_long_audio setting can be set to True or False. The default value is True.
The Debugging settings are used to specify the debugging options for the application.
The debug_mode setting is used to specify whether to enable debug mode. The debug_mode setting can be set to True or False. The default value is False. Debug mode forces Pantella to use the gradio game interface.
The share_debug_ui setting is used to specify whether to share the debug UI. The share_debug_ui setting can be set to True or False. The default value is False.
The remove_mei_folders setting is used to specify whether to remove MEI folders. The remove_mei_folders setting can be set to True or False. The default value is False.
The play_audio_from_script setting is used to specify whether to play audio from script. The play_audio_from_script setting can be set to True or False. The default value is False.
The debug_character_name setting is used to specify the debug character name. The debug_character_name setting can be set to any valid character name. The default value is Hulda.
The debug_use_mic setting is used to specify whether to use the mic in debug mode. The debug_use_mic setting can be set to True or False. The default value is False.
The default_player_response setting is used to specify the default player response. The default_player_response setting can be set to any valid player response. The default value is Can you tell me something about yourself?.
The tts_boot_annoncements setting is used to specify whether to use TTS boot announcements. The tts_boot_annoncements setting can be set to True or False. The default value is True.
The add_voicelines_to_all_voice_folders setting is used to specify whether to add voicelines to all voice folders. The add_voicelines_to_all_voice_folders setting can be set to True or False. The default value is False.
The play_startup_announcement setting is used to specify whether to play the startup announcement. The play_startup_announcement setting can be set to True or False. The default value is True.
The Errors settings are used to specify the error handling options for the application.
The block_logs_from setting is used to specify the blocked logs from. The block_logs_from setting can be set to a list of blocked logs from. The default value is an empty list.
The block_log_types setting is used to specify the blocked log types. The block_log_types setting can be set to a list of blocked log types. The default value is an empty list.
The error_on_empty_full_reply setting is used to specify whether to error on empty full reply. The error_on_empty_full_reply setting can be set to True or False. The default value is False.
The continue_on_voice_model_error setting is used to specify whether to continue on voice model error. The continue_on_voice_model_error setting can be set to True or False. The default value is False.
The continue_on_missing_character setting is used to specify whether to continue on missing character. The continue_on_missing_character setting can be set to True or False. The default value is False.
The continue_on_llm_api_error setting is used to specify whether to continue on LLM API error. The continue_on_llm_api_error setting can be set to True or False. The default value is True.
The continue_on_failure_to_send_audio_to_game_interface setting is used to specify whether to continue on failure to send audio to game interface. The continue_on_failure_to_send_audio_to_game_interface setting can be set to True or False. The default value is False.
The must_generate_a_sentence setting is used to specify whether to must generate a sentence. The must_generate_a_sentence setting can be set to True or False. The default value is False.
The use_game_event_lines_as_is_if_cannot_parse setting is used to specify whether to use game event lines as is if cannot parse. The use_game_event_lines_as_is_if_cannot_parse setting can be set to True or False. The default value is True.
The bad_author_retries setting is used to specify the bad author retries. The bad_author_retries setting can be set to any integer value. The default value is 5.
The retries setting is used to specify the retries. The retries setting can be set to any integer value. The default value is 3.
The system_loop setting is used to specify the system loop. The system_loop setting can be set to any integer value. The default value is 3.
The Config settings are used to specify the configuration options for the application.
The linux_mode setting is used to specify whether to enable Linux mode. The linux_mode setting can be set to True or False. The default value is False. If you're running Pantella on Linux, set this to True.
The python_binary setting is used to specify the Python binary to use. The python_binary setting can be set to any valid Python binary. The default value is ../../python-3.10.11-embed/python.exe. Change to python or python3 for use with a system Python installation. If you want to use a different Python binary, you can set it here.
The character_database_file setting is used to specify the character database file. The character_database_file setting can be set to any valid character database file. The default value is .\\characters\\. This can be a csv file path, a directory file path, or a list of csv file paths and directory file paths.
The conversation_data_directory setting is used to specify the conversation data directory. The conversation_data_directory setting can be set to any valid conversation data directory. The default value is .\\data\\conversations.
The voice_model_ref_ids_file setting is used to specify the voice model ref IDs file. The voice_model_ref_ids_file setting can be set to any valid voice model ref IDs file. The default value is .\\skyrim_voice_model_ids.json.
The logging_file_path setting is used to specify the logging file path. The logging_file_path setting can be set to any valid logging file path. The default value is .\\logging.log.
The config_port setting is used to specify the config port. The config_port setting can be set to any integer value. The default value is 8021.
The memory_editor_port setting is used to specify the memory editor port. The memory_editor_port setting can be set to any integer value. The default value is 8022.
The debug_ui_port setting is used to specify the debug UI port. The debug_ui_port setting can be set to any integer value. The default value is 8023.
The ConfigLoader class is responsible for loading the settings from the config.json file and providing them to the rest of the application. The ConfigLoader class also provides a web server that can be used to change the settings in the config.json file while the application is running. The class is defined in the ./src/config_loader.py file and includes the following methods and properties:
-
__init__(self): The constructor for the ConfigLoader class. This method initializes the ConfigLoader object and loads the settings from theconfig.jsonfile. -
game_path: A property that returns the path to the game directory specified in theconfig.jsonfile. -
mod_path: A property that returns the path to the mod directory specified in theconfig.jsonfile. -
save(self): A method that saves the current settings to theconfig.jsonfile. -
load(self): A method that loads the settings from theconfig.jsonfile. -
get_prompt_styles(self): A method that returns a list of available prompt styles. -
load_addons(self): A method that loads the addons from the./addons/directory. See more about addons over here. -
get_addon(self, addon_slug): A method that returns the addon with the specified name. -
get_behavior_styles(self): A method that returns a list of available behavior styles. -
prompts: A property that returns the prompts defined in your currently selected prompt style. See more about prompt styles over here. -
stop: A property that returns the stop values defined in your currently selected prompt style. See more about prompt styles over here. -
BOS_token: A property that returns the BOS_token values defined in your currently selected prompt style. See more about prompt styles over here. -
EOS_token: A property that returns the EOS_token values defined in your currently selected prompt style. See more about prompt styles over here. -
message_signifier: A property that returns the message_signifier values defined in your currently selected prompt style. See more about prompt styles over here. -
role_seperator: A property that returns the role_seperator values defined in your currently selected prompt style. See more about prompt styles over here. -
message_separator: A property that returns the message_separator values defined in your currently selected prompt style. See more about prompt styles over here. -
message_format: A property that returns the message_format values defined in your currently selected prompt style. See more about prompt styles over here. -
system_name: A property that returns the system_name values defined in your currently selected prompt style. See more about prompt styles over here. -
user_name: A property that returns the user_name values defined in your currently selected prompt style. See more about prompt styles over here. -
assistant_name: A property that returns the assistant_name values defined in your currently selected prompt style. See more about prompt styles over here. -
language: A property that returns the language values defined in your currently selected prompt style. See more about prompt styles over here. -
racial_language: A property that returns the racial_language values defined in your currently selected prompt style. See more about prompt styles over here. -
set_prompt_style(self, llm):- Set the prompt style for the conversation manager -
set_behavior_style(self, behavior_style):- Set the behavior style for the conversation manager -
default(self): A method that returns the default settings for the application. -
_unique(self): A method that returns a dictionary of settings that have been changed from the default settings. -
unique(self): A method that returns a dictionary of settings that have been changed from the default settings. -
descriptions(self): A method that returns a dictionary of descriptions for each setting. -
export(self): A method that returns the current settings in a dictionary format. -
default_types(self): A method that returns the types of each setting in the default settings. -
host_config_server(self): A method that starts a web server that can be used to change the settings in theconfig.jsonfile while the application is running.
The constructor for the ConfigLoader class. This method initializes the ConfigLoader object and loads the settings from the config.json file. This is where the order of which the settings are loaded is defined.
A property that returns the path to the game directory specified in the config.json file.
A property that returns the path to the mod directory specified in the config.json file.
A method that saves the current settings to the config.json file.
A method that loads the settings from the config.json file.
A method that returns a list of available prompt styles.
A method that loads the addons from the ./addons/ directory. See more about addons over here.
A method that returns the addon with the specified name.
A method that returns a list of available behavior styles.
A property that returns the prompts defined in your currently selected prompt style. See more about prompt styles over here.
A property that returns the stop values defined in your currently selected prompt style. See more about prompt styles over here.
A property that returns the BOS_token values defined in your currently selected prompt style. See more about prompt styles over here.
A property that returns the EOS_token values defined in your currently selected prompt style. See more about prompt styles over here.
A property that returns the message_signifier values defined in your currently selected prompt style. See more about prompt styles over here.
A property that returns the role_seperator values defined in your currently selected prompt style. See more about prompt styles over here.
A property that returns the message_separator values defined in your currently selected prompt style. See more about prompt styles over here.
A property that returns the message_format values defined in your currently selected prompt style. See more about prompt styles over here.
A property that returns the system_name values defined in your currently selected prompt style. See more about prompt styles over here.
A property that returns the user_name values defined in your currently selected prompt style. See more about prompt styles over here.
A property that returns the assistant_name values defined in your currently selected prompt style. See more about prompt styles over here.
A property that returns the language values defined in your currently selected prompt style. See more about prompt styles over here.
A property that returns the racial_language values defined in your currently selected prompt style. See more about prompt styles over here.
Set the prompt style for the conversation manager
Set the behavior style for the conversation manager
A method that returns the default settings for the application.
A method that returns a dictionary of settings that have been changed from the default settings.
A method that returns a dictionary of settings that have been changed from the default settings.
A method that returns a dictionary of descriptions for each setting.
A method that returns the current settings in a dictionary format.
A method that returns the types of each setting in the default settings.
A method that starts a web server that can be used to change the settings in the config.json file while the application is running.