CML TextGen REST API Client.md

import requests
import html
import json

# Define the Texgen API endpoint
HOST = 'cmlllm-textgenuiurl'
URI = f'https://{HOST}/api/v1/chat'


# Function to send a message to Texgen
def send_message(user_input, history, username, continueFlag):
    request = {
        'user_input': user_input,
        'max_new_tokens': 2500,
        'auto_max_new_tokens': False,
        'max_tokens_second': 0,
        'history': history,
        'mode': 'instruct',  # Valid options: 'chat', 'chat-instruct', 'instruct'
        'character': 'Example',
        'instruction_template': 'Vicuna-v1.1',  # Will get autodetected if unset
        'your_name': username,
        # 'name1': 'name of user', # Optional
        # 'name2': 'name of character', # Optional
        # 'context': 'character context', # Optional
        # 'greeting': 'greeting', # Optional
        # 'name1_instruct': 'You', # Optional
        # 'name2_instruct': 'Assistant', # Optional
        # 'context_instruct': 'context_instruct', # Optional
        # 'turn_template': 'turn_template', # Optional
        'regenerate': False,
        '_continue': continueFlag,
        # 'chat_instruct_command': 'Continue the chat dialogue below. Write a single reply for the character "<|character|>".\n\n<|prompt|>',

        # Generation params. If 'preset' is set to different than 'None', the values
        # in presets/preset-name.yaml are used instead of the individual numbers.
        'preset': 'None',
        'do_sample': True,
        'temperature': 0.7,
        'top_p': 0.1,
        'typical_p': 1,
        'epsilon_cutoff': 0,  # In units of 1e-4
        'eta_cutoff': 0,  # In units of 1e-4
        'tfs': 1,
        'top_a': 0,
        'repetition_penalty': 1.18,
        'repetition_penalty_range': 0,
        'top_k': 40,
        'min_length': 0,
        'no_repeat_ngram_size': 0,
        'num_beams': 1,
        'penalty_alpha': 0,
        'length_penalty': 1,
        'early_stopping': False,
        'mirostat_mode': 0,
        'mirostat_tau': 5,
        'mirostat_eta': 0.1,
        'grammar_string': '',
        'guidance_scale': 1,
        'negative_prompt': '',

        'seed': -1,
        'add_bos_token': True,
        'truncation_length': 2048 * 5,
        'ban_eos_token': False,
        'custom_token_bans': '',
        'skip_special_tokens': True,
        'stopping_strings': []
    }
    response = requests.post(URI, json=request)
    return response;

     
# Function to interact with the chatbot
def chat_with_bot():
    chatUsername = input("Please enter your name :")
    print(f'Hello {chatUsername}, Welcome to the Texgen Chatbot! \nType exit to quit')
    
    # Basic example
    user_input = ""
    history = {'internal': [], 'visible': []}
    continueGeneration = 'n'
    
    while True:
        if continueGeneration.lower() == 'n':
            user_input = input(f'{chatUsername}:')
            if user_input.lower() == "exit":
                print("Goodbye!")
                break
        
        continueFlag = False if (continueGeneration.lower() == 'n') else True    
        response = send_message(user_input, history, chatUsername,continueFlag)
    
        if response.status_code == 200:
            data = response.json()
            if data['results'][0]['history']['visible']:
                history = data['results'][0]['history']
                visible_response = history['visible'][-1][1]
                unescaped_string = html.unescape(visible_response)
                print(f'TextGen Bot: {unescaped_string}')
                #If the response ends with a '?', user shoud answer that question first.
                if not unescaped_string.endswith("?"):
                    # Assume the response is partial , and ask user to contuinue with the text generation.
                    continueGeneration = input("Do you want to continue generation of the current response (y/n)?")
                    continue
            else:
                print("TextGen Bot: Response is empty.")
        else:
            print("Error: Unable to get a response from the Texgen API.")
        continueGeneration = 'n'    

if __name__ == "__main__":
    chat_with_bot()