phoenix/litellm-mirror
0
0
Fork 1
mirror of https://github.com/BerriAI/litellm.git synced 2025-04-25 02:34:29 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

updates

Browse source
This commit is contained in:
Krrish Dholakia 2023-08-03 06:57:30 -07:00
parent 231a3e727b
commit 6ae8384fb5
14 changed files with 59 additions and 36 deletions
Download patch file Download diff file Expand all files Collapse all files

2
build/lib/litellm/__init__.py
View file

@ -2,7 +2,7 @@ success_callback = []
failure_callback = []
set_verbose=False
telemetry=True
max_tokens = 256 # OpenAI Defaults
####### PROXY PARAMS ################### configurable params if you use proxy models like Helicone
api_base = None
headers = None

2
build/lib/litellm/main.py
View file

@ -189,7 +189,7 @@ def completion(
if max_tokens != float('inf'):
max_tokens_to_sample = max_tokens
else:
max_tokens_to_sample = 300 # default in Anthropic docs https://docs.anthropic.com/claude/reference/client-libraries
max_tokens_to_sample = litellm.max_tokens # default in Anthropic docs https://docs.anthropic.com/claude/reference/client-libraries
## LOGGING
logging(model=model, input=prompt, azure=azure, additional_args={"max_tokens": max_tokens}, logger_fn=logger_fn)
## COMPLETION CALL

22
build/lib/litellm/utils.py
View file

@ -2,6 +2,7 @@ import dotenv, json, traceback, threading
import subprocess, os
import litellm, openai
import random, uuid, requests
import datetime
from openai.error import AuthenticationError, InvalidRequestError, RateLimitError, ServiceUnavailableError, OpenAIError
####### ENVIRONMENT VARIABLES ###################
dotenv.load_dotenv() # Loading env variables using dotenv
@ -11,6 +12,7 @@ add_breadcrumb = None
posthog = None
slack_app = None
alerts_channel = None
heliconeLogger = None
callback_list = []
user_logger_fn = None
additional_details = {}
@ -68,7 +70,7 @@ def client(original_function):
global callback_list, add_breadcrumb
if (len(litellm.success_callback) > 0 or len(litellm.failure_callback) > 0) and len(callback_list) == 0:
callback_list = list(set(litellm.success_callback + litellm.failure_callback))
set_callbacks(callback_list=callback_list)
set_callbacks(callback_list=callback_list,)
if add_breadcrumb:
add_breadcrumb(
category="litellm.llm_call",
@ -83,9 +85,11 @@ def client(original_function):
try:
function_setup(args, kwargs)
## MODEL CALL
start_time = datetime.datetime.now()
result = original_function(*args, **kwargs)
end_time = datetime.datetime.now()
## LOG SUCCESS
my_thread = threading.Thread(target=handle_success, args=(args, kwargs)) # don't interrupt execution of main thread
my_thread = threading.Thread(target=handle_success, args=(args, kwargs, result, start_time, end_time)) # don't interrupt execution of main thread
my_thread.start()
return result
except Exception as e:
@ -97,7 +101,7 @@ def client(original_function):
####### HELPER FUNCTIONS ################
def set_callbacks(callback_list):
global sentry_sdk_instance, capture_exception, add_breadcrumb, posthog, slack_app, alerts_channel
global sentry_sdk_instance, capture_exception, add_breadcrumb, posthog, slack_app, alerts_channel, heliconeLogger
try:
for callback in callback_list:
if callback == "sentry":
@ -134,6 +138,10 @@ def set_callbacks(callback_list):
)
alerts_channel = os.environ["SLACK_API_CHANNEL"]
print_verbose(f"Initialized Slack App: {slack_app}")
elif callback == "helicone":
from .integrations.helicone import HeliconeLogger
heliconeLogger = HeliconeLogger()
except:
pass
@ -200,7 +208,8 @@ def handle_failure(exception, traceback_exception, args, kwargs):
except:
pass
def handle_success(*args, **kwargs):
def handle_success(args, kwargs, result, start_time, end_time):
global heliconeLogger
try:
success_handler = additional_details.pop("success_handler", None)
failure_handler = additional_details.pop("failure_handler", None)
@ -223,6 +232,11 @@ def handle_success(*args, **kwargs):
for detail in additional_details:
slack_msg += f"{detail}: {additional_details[detail]}\n"
slack_app.client.chat_postMessage(channel=alerts_channel, text=slack_msg)
elif callback == "helicone":
print_verbose("reaches helicone for logging!")
model = args[0] if len(args) > 0 else kwargs["model"]
messages = args[1] if len(args) > 1 else kwargs["messages"]
heliconeLogger.log_success(model=model, messages=messages, response_obj=result, start_time=start_time, end_time=end_time)
except:
pass

BIN
dist/litellm-0.1.220-py3-none-any.whl vendored
View file

Binary file not shown.

BIN
dist/litellm-0.1.220.tar.gz vendored
View file

Binary file not shown.

BIN
dist/litellm-0.1.222-py3-none-any.whl vendored Normal file
View file

Binary file not shown.

BIN
dist/litellm-0.1.222.tar.gz vendored Normal file
View file

Binary file not shown.

28
docs/advanced.md
View file

@ -1,21 +1,27 @@
# Advanced - liteLLM client
# Advanced - Callbacks
## Use liteLLM client to send Output Data to Posthog, Sentry etc
liteLLM allows you to create `completion_client` and `embedding_client` to send successfull / error LLM API call data to Posthog, Sentry, Slack etc
## Use Callbacks to send Output Data to Posthog, Sentry etc
liteLLM provides `success_callbacks` and `failure_callbacks`, making it easy for you to send data to a particular provider depending on the status of your responses.
liteLLM supports:
- [Helicone](https://docs.helicone.ai/introduction)
- [Sentry](https://docs.sentry.io/platforms/python/)
- [PostHog](https://posthog.com/docs/libraries/python)
- [Slack](https://slack.dev/bolt-python/concepts)
### Quick Start
```python
from main import litellm_client
import os
from litellm import completion
# set callbacks
litellm.success_callback=["posthog", "helicone"]
litellm.failure_callback=["sentry"]
## set env variables
os.environ['SENTRY_API_URL'] = ""
os.environ['SENTRY_API_URL'], os.environ['SENTRY_API_TRACE_RATE']= ""
os.environ['POSTHOG_API_KEY'], os.environ['POSTHOG_API_URL'] = "api-key", "api-url"
# init liteLLM client
client = litellm_client(success_callback=["posthog"], failure_callback=["sentry", "posthog"])
completion = client.completion
embedding = client.embedding
os.environ["HELICONE_API_KEY"] = ""
response = completion(model="gpt-3.5-turbo", messages=messages)
```

9
docs/client_integrations.md
View file

@ -1,10 +1,11 @@
# Data Logging Integrations
| Integration | Required OS Variables | How to Use with litellm Client |
| Integration | Required OS Variables | How to Use with callbacks |
|-----------------|--------------------------------------------|-------------------------------------------|
| Sentry | `SENTRY_API_URL` | `client = litellm_client(success_callback=["sentry"], failure_callback=["sentry"])` |
| Posthog | `POSTHOG_API_KEY`,<br>`POSTHOG_API_URL` | `client = litellm_client(success_callback=["posthog"], failure_callback=["posthog"])` |
| Slack | `SLACK_API_TOKEN`,<br>`SLACK_API_SECRET`,<br>`SLACK_API_CHANNEL` | `client = litellm_client(success_callback=["slack"], failure_callback=["slack"])` |
| Sentry | `SENTRY_API_URL` | `litellm.success_callback=["sentry"], litellm.failure_callback=["sentry"]` |
| Posthog | `POSTHOG_API_KEY`,<br>`POSTHOG_API_URL` | `litellm.success_callback=["posthog"], litellm.failure_callback=["posthog"]` |
| Slack | `SLACK_API_TOKEN`,<br>`SLACK_API_SECRET`,<br>`SLACK_API_CHANNEL` | `litellm.success_callback=["slack"], litellm.failure_callback=["slack"]` |
| Helicone | `HELICONE_API_TOKEN` | `litellm.success_callback=["helicone"]` |

6
docs/index.md
View file

@ -4,7 +4,11 @@ a light 100 line package to simplify calling OpenAI, Azure, Cohere, Anthropic AP
###### litellm manages:
* Calling all LLM APIs using the OpenAI format - `completion(model, messages)`
* Consistent output for all LLM APIs, text responses will always be available at `['choices'][0]['message']['content']`
* **[Advanced]** Automatically logging your output to Sentry, Posthog, Slack [see liteLLM Client](https://litellm.readthedocs.io/en/latest/advanced/)
* Consistent Exceptions for all LLM APIs, we map RateLimit, Context Window, and Authentication Error exceptions across all providers to their OpenAI equivalents. [see Code](https://github.com/BerriAI/litellm/blob/ba1079ff6698ef238c5c7f771dd2b698ec76f8d9/litellm/utils.py#L250)
###### observability:
* Logging - see exactly what the raw model request/response is `completion(.., logger_fn=your_logging_fn)`
* Callbacks - automatically send your data to Helicone, Sentry, Posthog, Slack - `litellm.success_callbacks`, `litellm.failure_callbacks` [see Callbacks](https://litellm.readthedocs.io/en/latest/advanced/)
## Quick Start
Go directly to code: [Getting Started Notebook](https://colab.research.google.com/drive/1gR3pY-JzDZahzpVdbGBtrNGDBmzUNJaJ?usp=sharing)

21
docs/supported.md
View file

@ -5,6 +5,8 @@
| Model Name | Function Call | Required OS Variables |
|------------------|----------------------------------------|--------------------------------------|
| gpt-3.5-turbo | `completion('gpt-3.5-turbo', messages)` | `os.environ['OPENAI_API_KEY']` |
| gpt-3.5-turbo-16k | `completion('gpt-3.5-turbo-16k', messages)` | `os.environ['OPENAI_API_KEY']` |
| gpt-3.5-turbo-16k-0613 | `completion('gpt-3.5-turbo-16k-0613', messages)` | `os.environ['OPENAI_API_KEY']` |
| gpt-4 | `completion('gpt-4', messages)` | `os.environ['OPENAI_API_KEY']` |
## Azure OpenAI Chat Completion Models
@ -26,16 +28,11 @@
|------------------|--------------------------------------------|--------------------------------------|
| command-nightly | `completion('command-nightly', messages)` | `os.environ['COHERE_API_KEY']` |
### OpenRouter Models
| Model Name | Function Call | Required OS Variables |
|----------------------------------|----------------------------------------------------------------------|---------------------------------------------------------------------------|
| google/palm-2-codechat-bison | `completion('google/palm-2-codechat-bison', messages)` | `os.environ['OPENROUTER_API_KEY']`,<br>`os.environ['OR_SITE_URL']`,<br>`os.environ['OR_APP_NAME']` |
| google/palm-2-chat-bison | `completion('google/palm-2-chat-bison', messages)` | `os.environ['OPENROUTER_API_KEY']`,<br>`os.environ['OR_SITE_URL']`,<br>`os.environ['OR_APP_NAME']` |
| openai/gpt-3.5-turbo | `completion('openai/gpt-3.5-turbo', messages)` | `os.environ['OPENROUTER_API_KEY']`,<br>`os.environ['OR_SITE_URL']`,<br>`os.environ['OR_APP_NAME']` |
| openai/gpt-3.5-turbo-16k | `completion('openai/gpt-3.5-turbo-16k', messages)` | `os.environ['OPENROUTER_API_KEY']`,<br>`os.environ['OR_SITE_URL']`,<br>`os.environ['OR_APP_NAME']` |
| openai/gpt-4-32k | `completion('openai/gpt-4-32k', messages)` | `os.environ['OPENROUTER_API_KEY']`,<br>`os.environ['OR_SITE_URL']`,<br>`os.environ['OR_APP_NAME']` |
| anthropic/claude-2 | `completion('anthropic/claude-2', messages)` | `os.environ['OPENROUTER_API_KEY']`,<br>`os.environ['OR_SITE_URL']`,<br>`os.environ['OR_APP_NAME']` |
| anthropic/claude-instant-v1 | `completion('anthropic/claude-instant-v1', messages)` | `os.environ['OPENROUTER_API_KEY']`,<br>`os.environ['OR_SITE_URL']`,<br>`os.environ['OR_APP_NAME']` |
| meta-llama/llama-2-13b-chat | `completion('meta-llama/llama-2-13b-chat', messages)` | `os.environ['OPENROUTER_API_KEY']`,<br>`os.environ['OR_SITE_URL']`,<br>`os.environ['OR_APP_NAME']` |
| meta-llama/llama-2-70b-chat | `completion('meta-llama/llama-2-70b-chat', messages)` | `os.environ['OPENROUTER_API_KEY']`,<br>`os.environ['OR_SITE_URL']`,<br>`os.environ['OR_APP_NAME']` |
### Anthropic Models
| Model Name | Function Call | Required OS Variables |
|------------------|--------------------------------------------|--------------------------------------|
| claude-instant-1 | `completion('claude-instant-1', messages)` | `os.environ['ANTHROPIC_API_KEY']` |
| claude-v2 | `completion('claude-v2', messages)` | `os.environ['ANTHROPIC_API_KEY']` |

2
litellm.egg-info/PKG-INFO
View file

@ -1,6 +1,6 @@
Metadata-Version: 2.1
Name: litellm
Version: 0.1.220
Version: 0.1.222
Summary: Library to easily interface with LLM API providers
Author: BerriAI
License-File: LICENSE

1
litellm/utils.py
View file

@ -112,6 +112,7 @@ def set_callbacks(callback_list):
subprocess.check_call([sys.executable, '-m', 'pip', 'install', 'sentry_sdk'])
import sentry_sdk
sentry_sdk_instance = sentry_sdk
sentry_trace_rate = os.environ.get("SENTRY_API_TRACE_RATE") if "SENTRY_API_TRACE_RATE" in os.environ else "1.0"
sentry_sdk_instance.init(dsn=os.environ.get("SENTRY_API_URL"), traces_sample_rate=float(os.environ.get("SENTRY_API_TRACE_RATE")))
capture_exception = sentry_sdk_instance.capture_exception
add_breadcrumb = sentry_sdk_instance.add_breadcrumb

2
setup.py
View file

@ -2,7 +2,7 @@ from setuptools import setup, find_packages
setup(
name='litellm',
version='0.1.221',
version='0.1.222',
description='Library to easily interface with LLM API providers',
author='BerriAI',
packages=[