Spaces:
Configuration error
Configuration error
File size: 9,105 Bytes
447ebeb |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 |
import json
import os
import sys
from datetime import datetime
from unittest.mock import AsyncMock
sys.path.insert(
0, os.path.abspath("../..")
) # Adds the parent directory to the system-path
import pytest
import litellm
from litellm.integrations.opentelemetry import OpenTelemetry, OpenTelemetryConfig, Span
import asyncio
import logging
from opentelemetry.sdk.trace.export.in_memory_span_exporter import InMemorySpanExporter
from litellm._logging import verbose_logger
from litellm.proxy._types import SpanAttributes
verbose_logger.setLevel(logging.DEBUG)
EXPECTED_SPAN_NAMES = ["litellm_request", "raw_gen_ai_request"]
exporter = InMemorySpanExporter()
@pytest.mark.asyncio
@pytest.mark.parametrize("streaming", [True, False])
async def test_async_otel_callback(streaming):
litellm.set_verbose = True
litellm.callbacks = [OpenTelemetry(config=OpenTelemetryConfig(exporter=exporter))]
response = await litellm.acompletion(
model="gpt-3.5-turbo",
messages=[{"role": "user", "content": "hi"}],
temperature=0.1,
user="OTEL_USER",
stream=streaming,
)
if streaming is True:
async for chunk in response:
print("chunk", chunk)
await asyncio.sleep(4)
spans = exporter.get_finished_spans()
print("spans", spans)
assert len(spans) == 2
_span_names = [span.name for span in spans]
print("recorded span names", _span_names)
assert set(_span_names) == set(EXPECTED_SPAN_NAMES)
# print the value of a span
for span in spans:
print("span name", span.name)
print("span attributes", span.attributes)
if span.name == "litellm_request":
validate_litellm_request(span)
# Additional specific checks
assert span._attributes["gen_ai.request.model"] == "gpt-3.5-turbo"
assert span._attributes["gen_ai.system"] == "openai"
assert span._attributes["gen_ai.request.temperature"] == 0.1
assert span._attributes["llm.is_streaming"] == str(streaming)
assert span._attributes["llm.user"] == "OTEL_USER"
elif span.name == "raw_gen_ai_request":
if streaming is True:
validate_raw_gen_ai_request_openai_streaming(span)
else:
validate_raw_gen_ai_request_openai_non_streaming(span)
# clear in memory exporter
exporter.clear()
def validate_litellm_request(span):
expected_attributes = [
"gen_ai.request.model",
"gen_ai.system",
"gen_ai.request.temperature",
"llm.is_streaming",
"llm.user",
"gen_ai.response.id",
"gen_ai.response.model",
"llm.usage.total_tokens",
"gen_ai.usage.completion_tokens",
"gen_ai.usage.prompt_tokens",
]
# get the str of all the span attributes
print("span attributes", span._attributes)
for attr in expected_attributes:
value = span._attributes[attr]
print("value", value)
assert value is not None, f"Attribute {attr} has None value"
def validate_raw_gen_ai_request_openai_non_streaming(span):
expected_attributes = [
"llm.openai.messages",
"llm.openai.temperature",
"llm.openai.user",
"llm.openai.extra_body",
"llm.openai.id",
"llm.openai.choices",
"llm.openai.created",
"llm.openai.model",
"llm.openai.object",
"llm.openai.service_tier",
"llm.openai.system_fingerprint",
"llm.openai.usage",
]
print("span attributes", span._attributes)
for attr in span._attributes:
print(attr)
for attr in expected_attributes:
assert span._attributes[attr] is not None, f"Attribute {attr} has None"
def validate_raw_gen_ai_request_openai_streaming(span):
expected_attributes = [
"llm.openai.messages",
"llm.openai.temperature",
"llm.openai.user",
"llm.openai.extra_body",
"llm.openai.model",
]
print("span attributes", span._attributes)
for attr in span._attributes:
print(attr)
for attr in expected_attributes:
assert span._attributes[attr] is not None, f"Attribute {attr} has None"
@pytest.mark.parametrize(
"model",
["anthropic/claude-3-opus-20240229"],
)
@pytest.mark.flaky(retries=6, delay=2)
def test_completion_claude_3_function_call_with_otel(model):
litellm.set_verbose = True
litellm.callbacks = [OpenTelemetry(config=OpenTelemetryConfig(exporter=exporter))]
tools = [
{
"type": "function",
"function": {
"name": "get_current_weather",
"description": "Get the current weather in a given location",
"parameters": {
"type": "object",
"properties": {
"location": {
"type": "string",
"description": "The city and state, e.g. San Francisco, CA",
},
"unit": {"type": "string", "enum": ["celsius", "fahrenheit"]},
},
"required": ["location"],
},
},
}
]
messages = [
{
"role": "user",
"content": "What's the weather like in Boston today in Fahrenheit?",
}
]
try:
# test without max tokens
response = litellm.completion(
model=model,
messages=messages,
tools=tools,
tool_choice={
"type": "function",
"function": {"name": "get_current_weather"},
},
drop_params=True,
)
print("response from LiteLLM", response)
except litellm.InternalServerError:
pass
except Exception as e:
pytest.fail(f"Error occurred: {e}")
finally:
# clear in memory exporter
exporter.clear()
@pytest.mark.asyncio
@pytest.mark.parametrize("streaming", [True, False])
@pytest.mark.parametrize("global_redact", [True, False])
async def test_awesome_otel_with_message_logging_off(streaming, global_redact):
"""
No content should be logged when message logging is off
tests when litellm.turn_off_message_logging is set to True
tests when OpenTelemetry(message_logging=False) is set
"""
litellm.set_verbose = True
litellm.callbacks = [OpenTelemetry(config=OpenTelemetryConfig(exporter=exporter))]
if global_redact is False:
otel_logger = OpenTelemetry(
message_logging=False, config=OpenTelemetryConfig(exporter="console")
)
else:
# use global redaction
litellm.turn_off_message_logging = True
otel_logger = OpenTelemetry(config=OpenTelemetryConfig(exporter="console"))
litellm.callbacks = [otel_logger]
litellm.success_callback = []
litellm.failure_callback = []
response = await litellm.acompletion(
model="gpt-3.5-turbo",
messages=[{"role": "user", "content": "hi"}],
mock_response="hi",
stream=streaming,
)
print("response", response)
if streaming is True:
async for chunk in response:
print("chunk", chunk)
await asyncio.sleep(1)
spans = exporter.get_finished_spans()
print("spans", spans)
assert len(spans) == 1
_span = spans[0]
print("span attributes", _span.attributes)
validate_redacted_message_span_attributes(_span)
# clear in memory exporter
exporter.clear()
if global_redact is True:
litellm.turn_off_message_logging = False
def validate_redacted_message_span_attributes(span):
expected_attributes = [
"gen_ai.request.model",
"gen_ai.system",
"llm.is_streaming",
"llm.request.type",
"gen_ai.response.id",
"gen_ai.response.model",
"llm.usage.total_tokens",
"metadata.prompt_management_metadata",
"gen_ai.usage.completion_tokens",
"gen_ai.usage.prompt_tokens",
"metadata.user_api_key_hash",
"metadata.requester_ip_address",
"metadata.user_api_key_team_alias",
"metadata.requester_metadata",
"metadata.user_api_key_team_id",
"metadata.spend_logs_metadata",
"metadata.usage_object",
"metadata.user_api_key_alias",
"metadata.user_api_key_user_id",
"metadata.user_api_key_org_id",
"metadata.user_api_key_end_user_id",
"metadata.user_api_key_user_email",
"metadata.applied_guardrails",
"metadata.mcp_tool_call_metadata",
"metadata.vector_store_request_metadata",
"metadata.requester_custom_headers",
]
_all_attributes = set(
[
name.value if isinstance(name, SpanAttributes) else name
for name in span.attributes.keys()
]
)
print("all_attributes", _all_attributes)
for attr in _all_attributes:
print(f"attr: {attr}, type: {type(attr)}")
assert _all_attributes == set(expected_attributes)
pass
|