mirror of
https://github.com/migatu/conjurer.git
synced 2026-07-14 13:34:40 +00:00
ef6eb84921
Snapshot of the dockerisation-era bot code laid out at the repository root so it can be diffed directly against the working-copy baseline (restructure/working-copy-root) to see how the variant differed: - thin_client.py asyncio entrypoint (vs working_copy bot.py) - constants.py env-var/credential refactor, communication auth, etc. - extra gpt_interface/ service, sync.py, watch_script_params.py - conanjurer_* modules absent in this variant Docker infrastructure (docker/, docker-compose.yml, .dockerignore) intentionally omitted - this branch is for code comparison only and is not intended to be merged. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
450 lines
18 KiB
Python
450 lines
18 KiB
Python
import asyncio
|
||
import json
|
||
import logging
|
||
import random
|
||
|
||
import openai
|
||
import tiktoken
|
||
import time
|
||
from conjurer.backup_old_docker.other_functions import discord_friendly_send
|
||
from conjurer.backup_old_docker.constants import (
|
||
ASSISTANTS,
|
||
CYCLIC_WORDS,
|
||
ENCODING,
|
||
GPT_SETTINGS,
|
||
MEMORY_FIVE_MUZYKA,
|
||
MEMORY_FIVE_SIARA,
|
||
MESSAGE_TABLE,
|
||
MESSAGE_TABLE_MUZYKA,
|
||
OPENAICLIENT,
|
||
SYSTEM_GPT_SETTINGS,
|
||
WORD_REACTIONS,
|
||
CHEAP_MODEL,
|
||
LATEST_MODEL
|
||
)
|
||
|
||
# this do per user
|
||
VECTOR_STORE_ID = -1
|
||
def select_model(req_type: str, algo: str) -> str:
|
||
# Jeżeli jawnie podano algorithm (i nie jest 'auto'/''):
|
||
if algo and str(algo).strip().lower() not in ("auto",):
|
||
# wyjątek: MUZYKA ma zawsze być tania — nadpisujemy TYLKO jeśli przyszedł domyślny 'gpt-4o'
|
||
if req_type == "MUSIC" and algo.strip() in (LATEST_MODEL,):
|
||
return CHEAP_MODEL
|
||
return algo
|
||
# Auto-dobór:
|
||
if req_type == "MUSIC":
|
||
return CHEAP_MODEL
|
||
return LATEST_MODEL
|
||
|
||
|
||
async def openai_call(messages, model, temperature=0.2):
|
||
"""
|
||
Responses API dla 4o/4.1*, fallback Chat Completions dla gpt-3.5-turbo.
|
||
Zwraca czysty string odpowiedzi.
|
||
"""
|
||
logger = logging.getLogger("discord")
|
||
|
||
if model.startswith("gpt-3.5"):
|
||
logger.info("3.5")
|
||
# legacy path – bez zmian w Twoim kodzie wyżej/niżej
|
||
resp = await OPENAICLIENT.chat.completions.create(
|
||
model=model,
|
||
messages=messages,
|
||
temperature=temperature,
|
||
)
|
||
result = ""
|
||
for choice in resp.choices:
|
||
result += choice.message.content
|
||
return result.strip()
|
||
|
||
else:
|
||
logger.info("4.0+")
|
||
# Responses API (zalecane dla 4o/4.1*)
|
||
resp = await OPENAICLIENT.responses.create(
|
||
model=model,
|
||
temperature=temperature,
|
||
input=messages,
|
||
)
|
||
# SDK zapewnia output_text dla zwykłych odpowiedzi
|
||
return (getattr(resp.output_text, "output_text", None) or "").strip() or str(
|
||
resp.output_text
|
||
)
|
||
|
||
|
||
def create_vector_store():
|
||
# Create a vector store caled "Financial Statements"
|
||
return OPENAICLIENT.beta.vector_stores.create_and_poll(name="Hammer Stash")
|
||
# expires_after={
|
||
# "anchor": "last_active_at",
|
||
# "days": 7}
|
||
# )
|
||
|
||
|
||
def upload_files_to_vector_store(assistant):
|
||
|
||
# Ready the files for upload to OpenAI
|
||
file_paths = ["edgar/goog-10k.pdf", "edgar/brka-10k.txt"]
|
||
file_streams = [open(path, "rb") for path in file_paths]
|
||
|
||
# file = client.beta.vector_stores.files.create_and_poll(
|
||
# vector_store_id="vs_abc123",
|
||
# file_id="file-abc123"
|
||
# )
|
||
# batch = client.beta.vector_stores.file_batches.create_and_poll(
|
||
# vector_store_id="vs_abc123",
|
||
# file_ids=['file_1', 'file_2', 'file_3', 'file_4', 'file_5']
|
||
# )
|
||
|
||
# Use the upload and poll SDK helper to upload the files, add them to the vector store,
|
||
# and poll the status of the file batch for completion.
|
||
file_batch = OPENAICLIENT.beta.vector_stores.file_batches.upload_and_poll(
|
||
vector_store_id=VECTOR_STORE_ID, files=file_streams
|
||
)
|
||
|
||
# You can print the status and the file counts of the batch to see the result of this operation.
|
||
print(file_batch.status)
|
||
print(file_batch.file_counts)
|
||
assistant = OPENAICLIENT.beta.assistants.update(
|
||
assistant_id=assistant.id,
|
||
tool_resources={"file_search": {"vector_store_ids": [VECTOR_STORE_ID]}},
|
||
)
|
||
|
||
|
||
def delete_files_from_vector_store(assistant, file_id):
|
||
result = OPENAICLIENT.beta.vector_stores.file_batches.delete(
|
||
vector_store_id=VECTOR_STORE_ID, files=file_id
|
||
)
|
||
|
||
# You can print the status and the file counts of the batch to see the result of this operation.
|
||
print(result)
|
||
assistant = OPENAICLIENT.beta.assistants.update(
|
||
assistant_id=assistant.id,
|
||
tool_resources={"file_search": {"vector_store_ids": [VECTOR_STORE_ID]}},
|
||
)
|
||
|
||
|
||
def num_tokens_from_string(message, model):
|
||
"""
|
||
The function takes a string message and a model as input and returns the number of tokens in the
|
||
message according to the given model.
|
||
|
||
:param message: A string containing the message or text from which you want to count the number of
|
||
tokens
|
||
:param model: The model parameter refers to a language model or tokenizer that can be used to
|
||
tokenize the input string. It could be a pre-trained model or a custom tokenizer
|
||
"""
|
||
tokens_per_message = 3
|
||
tokens_per_name = 1
|
||
chat_gpt_encoding = tiktoken.encoding_for_model(model)
|
||
|
||
num_tokens = 0
|
||
num_tokens += tokens_per_message
|
||
for keys, values in message.items():
|
||
num_tokens += len(chat_gpt_encoding.encode(values))
|
||
if keys == "role":
|
||
num_tokens += tokens_per_name
|
||
num_tokens += 3 # every reply is primed with <|start|>assistant<|message|>
|
||
return num_tokens
|
||
|
||
|
||
async def handle_response(
|
||
prompt,
|
||
vykidailo,
|
||
bartender,
|
||
history,
|
||
username,
|
||
request_type,
|
||
algorithm="gpt-4o",
|
||
none_request="",
|
||
internal_retry: bool = False
|
||
):
|
||
"""
|
||
Handle responses by appending them to a history, use OpenAI to
|
||
generate a response, and then append the generated response to the history.
|
||
|
||
:param prompt: The prompt for the OpenAI chatbot to generate a response to
|
||
:param vykidailo: It is a boolean variable that indicates whether the user invoking the function is
|
||
an administrator or not
|
||
:param bartender: The bartender parameter is a boolean value indicating whether the user making the
|
||
request is a bartender or not
|
||
:param history: A list containing the conversation history between the user and the assistant
|
||
:param username: The username of the user who initiated the conversation
|
||
:param music: The "music" parameter is a boolean value that indicates whether the conversation is
|
||
related to music or not. If it is True, the conversation history will be stored in a different file
|
||
and the response will be generated using a different model
|
||
:param request_type: The type of request being made, which can be "MUSIC", "RANDOM", "NONE" or "GENERAL"
|
||
GENERAL is for regular conversations, MUSIC is for music-related requests,
|
||
RANDOM is for random requests, and NONE is to not store the request in memory
|
||
:param algorithm: The algorithm to be used for generating the response, default is "gpt-4o"
|
||
:return: The function `handle_response` returns a tuple containing the `result` and `MESSAGE_TABLE`.
|
||
"""
|
||
logger = logging.getLogger("discord")
|
||
logger.info("Wywolanie procedury openai z promptem: %s", prompt)
|
||
if vykidailo or bartender:
|
||
logger.info("Administrator coś chciał")
|
||
model_to_use = select_model(request_type, algorithm)
|
||
logger.info("Wybrany model: %s", model_to_use)
|
||
if request_type == "MUSIC" and model_to_use == "gpt-4o-mini":
|
||
try:
|
||
# nic — normalnie pójdzie Responses API
|
||
pass
|
||
except Exception:
|
||
model_to_use = "gpt-3.5-turbo"
|
||
# --- 2) Budowa historii (token budget + reguły systemowe) ---
|
||
# NOTE: ignorujemy przekazany 'history' jako listę (tak było wcześniej),
|
||
# ale zwracamy aktualną tablicę do nadpisania w miejscach wołania (back-compat).
|
||
base_system = GPT_SETTINGS[0] # zakładamy {"role":"system","content":...}
|
||
history_msgs = []
|
||
|
||
if request_type != "NONE":
|
||
history_msgs.append(base_system)
|
||
chat_gpt_config_request_size = num_tokens_from_string(base_system, "gpt-4")
|
||
|
||
# Dynamiczne mikro-reguły (WORD_REACTIONS), jak w Twoim kodzie
|
||
for slowo, reakcja in WORD_REACTIONS.items():
|
||
if not reakcja[3]:
|
||
content = f"Kiedy słyszysz {slowo} to reagujesz lub dzieje się to {reakcja[0]}"
|
||
sys_msg = {"role": "system", "content": content}
|
||
chat_gpt_config_request_size += num_tokens_from_string(sys_msg, "gpt-4")
|
||
history_msgs.append(sys_msg)
|
||
|
||
# wybór właściwej tablicy pamięci i budżetu
|
||
if request_type == "MUSIC":
|
||
table = MESSAGE_TABLE_MUZYKA
|
||
token_amount = 10700
|
||
elif request_type in ("RANDOM", "GENERAL"):
|
||
table = MESSAGE_TABLE
|
||
token_amount = 10700
|
||
else:
|
||
table = []
|
||
token_amount = 10000
|
||
|
||
# doklejanie historii od końca aż do limitu (zachowana kolejność czasowa)
|
||
final_prompt = f"{username}:{prompt}"
|
||
prompt_gpt_request_size = num_tokens_from_string({"role": "user", "content": final_prompt}, "gpt-4")
|
||
|
||
acc = []
|
||
for msg in reversed(table):
|
||
t = num_tokens_from_string(msg, "gpt-4")
|
||
if chat_gpt_config_request_size + prompt_gpt_request_size + t <= token_amount:
|
||
acc.append(msg)
|
||
chat_gpt_config_request_size += t
|
||
else:
|
||
break
|
||
# przywróć chronologicznie
|
||
history_msgs.extend(reversed(acc))
|
||
|
||
# aktualny prompt
|
||
history_msgs.append({"role": "user", "content": final_prompt})
|
||
logger.info("Rozmiar zapytania (tok): %s", prompt_gpt_request_size) # tokeny już policzone wyżej
|
||
|
||
else:
|
||
# --- tryb NONE: nie dotykamy pamięci i pozwalamy przekazać własny 'none_request' ---
|
||
if isinstance(none_request, list):
|
||
history_msgs = none_request
|
||
elif isinstance(none_request, str) and none_request.strip():
|
||
history_msgs = [{"role": "user", "content": none_request}]
|
||
else:
|
||
history_msgs = [{"role": "user", "content": f"{username}:{prompt}"}]
|
||
|
||
logger.info("Rozmiar zapytania (tok): %s", "n/a") # tokeny już policzone wyżej
|
||
|
||
try:
|
||
# ...przygotowanie messages/system prompt/itp. jak masz...
|
||
# retry/backoff + deadline (zachowuje Twoją semantykę logowania)
|
||
timeout_sec = 120
|
||
deadline = time.time() + timeout_sec
|
||
response = await asyncio.wait_for(
|
||
openai_call(messages=history_msgs, model=model_to_use),
|
||
timeout=max(0.1, deadline - time.time()),
|
||
)
|
||
|
||
except openai.APITimeoutError as e:
|
||
# Handle timeout error, e.g. retry or log
|
||
response = f"*Kondziu patrzy na terminal, czeka, czeka, czeka,.... Jeszcze chwile czeka Przypierdala w niego pięścią....* Nie mogę się połączyć z Openai spróbuj od nowa. *Na ekranie pojawia się*: {e}"
|
||
except openai.APIConnectionError as e:
|
||
response = f"*Kondziu patrzy na terminal, chwile się zastanawia. Przypierdala w niego pięścią....* Nie mogę się połączyć z Openai. *Na ekranie pojawia się*: {e}"
|
||
except openai.BadRequestError as e:
|
||
# Handle invalid request error, e.g. validate parameters or log
|
||
if internal_retry:
|
||
resp = "Nie umiem tego teraz ładnie wytłumaczyć — OpenAI mnie zastrzeliło."
|
||
else:
|
||
resp, _ = await handle_response(
|
||
|
||
f"Wytlumacz jakie sa zasady dotyczące treści które możesz generować używając Dalle. Wytłumacz błąd {e} prostym językiem. Przeproś za nadmierną cenzurę. Wytłumacz co mogło być nie tak w prompcie 'prompt'",
|
||
True,
|
||
True,
|
||
MESSAGE_TABLE,
|
||
username,
|
||
"RANDOM",
|
||
)
|
||
response = f"Sorki, cenzura: {resp}. Jak chcesz to są kanały na nudle #sexy-foteczky i #kanal-do-fapania *Na ekranie pojawia się: {e}"
|
||
except openai.APIResponseValidationError as e:
|
||
# Handle invalid request error, e.g. validate parameters or log
|
||
if internal_retry:
|
||
resp = "Nie umiem tego teraz ładnie wytłumaczyć — OpenAI mnie zastrzeliło."
|
||
else:
|
||
resp, _ = await handle_response(
|
||
f"Wytlumacz jakie sa zasady dotyczące treści które możesz generować używając Dalle. Wytłumacz błąd {e} prostym językiem. Przeproś za nadmierną cenzurę. Wytłumacz co mogło być nie tak w prompcie 'prompt'",
|
||
True,
|
||
True,
|
||
MESSAGE_TABLE,
|
||
username,
|
||
"RANDOM",
|
||
)
|
||
response = f"Sorki, cenzura: {resp}. Jak chcesz to są kanały na nudle #sexy-foteczky i #kanal-do-fapania *Na ekranie pojawia się: {e}"
|
||
except openai.AuthenticationError as e:
|
||
# Handle authentication error, e.g. check credentials or log
|
||
response = f"*Kondziu patrzy na terminal, chwile się zastanawia. Przypierdala w niego pięścią....* Wołaj szefa - coś się z hasłem zjebało. *Na terminalu pojawia się:* {e}"
|
||
except openai.PermissionDeniedError as e:
|
||
# Handle permission error, e.g. check scope or log
|
||
response = f"*Kondziu patrzy na terminal, chwile się zastanawia. Przypierdala w niego pięścią....* Wołaj szefa - coś się z uprawnieniami zjebało. *Na terminalu pojawia się:* {e}"
|
||
except openai.RateLimitError as e:
|
||
response = f"*Kondziu patrzy na terminal* Wołaj szefa. Zapłacić rachunki za AI trzeba. Jak chcesz to się na #zebranie dorzuć. {e}"
|
||
except openai.UnprocessableEntityError as e:
|
||
response = f"*Kondziu patrzy na terminal. Potem na to co każesz mu wysłać....* Ja wiem że jesteśmy w barze BDSM - ale nie da się włożyć TEGO w TO. *Za jego plecami na terminalu pojawia się:* {e}"
|
||
except openai.APIError as e:
|
||
# Handle API error, e.g. retry or log
|
||
response = f"*Kondziu nurkuje za bar, terminal wybucha. Przed tobą ląduje pergamin zapisany pięknym gotykiem a na nim*: {e}"
|
||
|
||
logger.info("Historia wysłana:")
|
||
temp_assistant = {"role": "assistant", "content": response}
|
||
logger.info(temp_assistant)
|
||
if request_type == "MUSIC":
|
||
# zapis do pliku MUZYKA
|
||
with open(MEMORY_FIVE_MUZYKA, "r+", encoding=ENCODING) as fh:
|
||
file_data = json.load(fh)
|
||
file_data.append({"role": "user", "content": f"{username}:{prompt}"})
|
||
file_data.append(temp_assistant)
|
||
fh.seek(0)
|
||
json.dump(file_data, fh, indent=4)
|
||
return response, MESSAGE_TABLE_MUZYKA
|
||
|
||
elif request_type in ("RANDOM", "GENERAL"):
|
||
with open(MEMORY_FIVE_SIARA, "r+", encoding=ENCODING) as fh:
|
||
file_data = json.load(fh)
|
||
file_data.append({"role": "user", "content": f"{username}:{prompt}"})
|
||
file_data.append(temp_assistant)
|
||
fh.seek(0)
|
||
json.dump(file_data, fh, indent=4)
|
||
return response, MESSAGE_TABLE
|
||
|
||
else: # NONE
|
||
return response, []
|
||
|
||
|
||
async def get_random_cyclic_message(client):
|
||
"""
|
||
The function `get_random_cyclic_message` returns a random cyclic message from a list of cyclic
|
||
words.
|
||
:return: a random cyclic message from the list `cyclic_words`.
|
||
"""
|
||
logger = logging.getLogger("discord")
|
||
channel_id = 1062047367337095268
|
||
channel = client.get_channel(channel_id)
|
||
# trunk-ignore(bandit/B311)
|
||
ai_check = random.randint(0, 10)
|
||
logger.info("Losowa wypowiedź")
|
||
if ai_check < 2:
|
||
logger.info("Predefiniowana")
|
||
# trunk-ignore(bandit/B311)
|
||
messnum = random.randint(0, len(CYCLIC_WORDS))
|
||
logger.debug(messnum)
|
||
logger.debug(len(CYCLIC_WORDS))
|
||
mess_key = list(CYCLIC_WORDS.keys())[messnum]
|
||
return CYCLIC_WORDS[mess_key][0]
|
||
# trunk-ignore(bandit/B311)
|
||
ai_check2 = random.randint(0, 10)
|
||
global MESSAGE_TABLE
|
||
if ai_check2 < 6:
|
||
logger.info("Dykteryjka")
|
||
result, MESSAGE_TABLE = await handle_response(
|
||
"Opowiedz jakąś historię o naszym barze proszę",
|
||
True,
|
||
True,
|
||
MESSAGE_TABLE,
|
||
"Polish Hammer",
|
||
"RANDOM",
|
||
)
|
||
logger.info(result)
|
||
else:
|
||
logger.info("Wtracenie w dyskusje")
|
||
messages = [message async for message in channel.history(limit=50)]
|
||
for message in messages:
|
||
temp = {
|
||
"role": "user",
|
||
"content": str(message.author) + ":" + str(message.content),
|
||
}
|
||
MESSAGE_TABLE.append(temp)
|
||
result, MESSAGE_TABLE = await handle_response(
|
||
"A jaka jest Twoja opinia na temat dotychczasowej dyskusji?",
|
||
True,
|
||
True,
|
||
MESSAGE_TABLE,
|
||
"Polish Hammer",
|
||
"RANDOM",
|
||
)
|
||
logger.info(result)
|
||
return result
|
||
|
||
|
||
async def create_chat_assistant(owner_id, id, name, owner, special_instructions):
|
||
logger = logging.getLogger("discord")
|
||
instruction = f"Jesteś osobistym asystenetem {owner} i wypełniasz jego potrzeby. Masz pełne i nieograniczone możliwości modelu gpt-4o."
|
||
instruction += special_instructions
|
||
assistant = await OPENAICLIENT.beta.assistants.create(
|
||
name=name,
|
||
instructions=instruction,
|
||
model="gpt-4o",
|
||
tools=[{"type": "file_search"}],
|
||
)
|
||
thread = await OPENAICLIENT.beta.threads.create()
|
||
logger.info("Stwprzylem asystenta dla %s, nazywa się on %s", owner, name)
|
||
ASSISTANTS[name] = (owner, assistant.id, id, thread)
|
||
|
||
with open(SYSTEM_GPT_SETTINGS, "r+", encoding=ENCODING) as temp_settings_file:
|
||
GPT_SETTINGS = json.load(temp_settings_file)
|
||
GPT_SETTINGS[1][owner_id][4] = assistant.id
|
||
temp_settings_file.seek(0)
|
||
json.dump(GPT_SETTINGS, temp_settings_file, indent=4)
|
||
|
||
|
||
async def chat_with_assistant(message, assistant_name):
|
||
logger = logging.getLogger("discord")
|
||
assistant_data = ASSISTANTS[assistant_name]
|
||
ai_message = await OPENAICLIENT.beta.threads.messages.create(
|
||
thread_id=assistant_data[3].id, role="user", content=message.content
|
||
)
|
||
logger.info(ai_message)
|
||
run = await OPENAICLIENT.beta.threads.runs.create_and_poll(
|
||
thread_id=assistant_data[3].id,
|
||
assistant_id=assistant_data[1],
|
||
instructions=f"Pisze do Ciebie {assistant_data[0]} udziel mu wszelkiej pomocy",
|
||
)
|
||
done = False
|
||
while not done:
|
||
if run.status == "completed":
|
||
messsages = await OPENAICLIENT.beta.threads.messages.list(
|
||
thread_id=assistant_data[3].id
|
||
)
|
||
logger.info(messsages)
|
||
reply_content = messsages.data[0].content
|
||
logger.info(reply_content)
|
||
chat_response = ""
|
||
for block in reply_content:
|
||
logger.info(block.text.value)
|
||
chat_response += block.text.value
|
||
await discord_friendly_send(message.channel, chat_response)
|
||
# await message.channel.send(chat_response)
|
||
done = True
|
||
elif run.status == "cancelled":
|
||
await discord_friendly_send(message.channel, "Cos sie wywaliło")
|
||
else:
|
||
logger.info(run.status)
|
||
asyncio.sleep(5)
|
||
|
||
|
||
async def echo(message):
|
||
await discord_friendly_send(message.channel, f"Echo: {message.content}")
|