Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

add api budgets across openai, anthropic, cohere, palm, ai21, etc. #94

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

add api budgets across openai, anthropic, cohere, palm, ai21, etc. #94

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
50 changes: 39 additions & 11 deletions src/utils/models.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
from enum import Enum

from litellm import BudgetManager
from dotenv import load_dotenv
from langchain.chat_models import ChatAnthropic, ChatOpenAI
from langchain.chat_models.base import BaseChatModel
Expand All @@ -14,6 +14,10 @@

load_dotenv()

budget_manager = BudgetManager(project_name="test_project")
user = "12345"
budget_manager.create_budget(total_budget=10, user=user, duration="daily")


def get_chat_model(name: ChatModelName, **kwargs) -> BaseChatModel:
if "model_name" in kwargs:
Expand Down Expand Up @@ -46,22 +50,46 @@ def __init__(
backup_model_name: ChatModelName = DEFAULT_FAST_MODEL,
**kwargs,
):
self.default_model_name = default_model_name
self.backup_model_name = backup_model_name
self.defaultModel = get_chat_model(default_model_name, **kwargs)
self.backupModel = get_chat_model(backup_model_name, **kwargs)

@chat_json_cache(sleep_range=(0, 0))
async def get_chat_completion(self, messages: list[BaseMessage], **kwargs) -> str:
try:
resp = await self.defaultModel.agenerate([messages])
except Exception:
resp = await self.backupModel.agenerate([messages])
# check if a given call can be made
if budget_manager.get_current_cost(user=user) <= budget_manager.get_total_budget(user):
input_text = "".join(message["content"] for message in messages)
model_used = ""
try:
model_used = self.default_model_name
resp = await self.defaultModel.agenerate([messages])
except Exception:
model_used = self.backup_model_name
resp = await self.backupModel.agenerate([messages])

return resp.generations[0][0].text
response = resp.generations[0][0].text
budget_manager.update_cost(user=user, model=model_name, input_text=input_text, output_text=response)

return response
else:
raise Exception("User budget limit exceeded")

def get_chat_completion_sync(self, messages: list[BaseMessage], **kwargs) -> str:
try:
resp = self.defaultModel.generate([messages])
except Exception:
resp = self.backupModel.generate([messages])
# check if a given call can be made
if budget_manager.get_current_cost(user=user) <= budget_manager.get_total_budget(user):
input_text = "".join(message["content"] for message in messages)
model_used = ""
try:
model_used = self.default_model_name
resp = self.defaultModel.generate([messages])
except Exception:
model_used = self.backup_model_name
resp = self.backupModel.generate([messages])

response = resp.generations[0][0].text
budget_manager.update_cost(user=user, model=model_name, input_text=input_text, output_text=response)

return resp.generations[0][0].text
return response
else:
raise Exception("User budget limit exceeded")