Spaces:
Configuration error
Configuration error
File size: 14,022 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 |
import httpx
import json
import pytest
import sys
from typing import Any, Dict, List
from unittest.mock import MagicMock, Mock, patch
import os
import uuid
import time
import base64
sys.path.insert(
0, os.path.abspath("../..")
) # Adds the parent directory to the system path
import litellm
from abc import ABC, abstractmethod
from litellm.integrations.custom_logger import CustomLogger
import json
from litellm.types.utils import StandardLoggingPayload
from litellm.types.llms.openai import (
ResponseCompletedEvent,
ResponsesAPIResponse,
ResponseTextConfig,
ResponseAPIUsage,
IncompleteDetails,
)
from litellm.llms.custom_httpx.http_handler import AsyncHTTPHandler
def validate_responses_api_response(response, final_chunk: bool = False):
"""
Validate that a response from litellm.responses() or litellm.aresponses()
conforms to the expected ResponsesAPIResponse structure.
Args:
response: The response object to validate
Raises:
AssertionError: If the response doesn't match the expected structure
"""
# Validate response structure
print("response=", json.dumps(response, indent=4, default=str))
assert isinstance(
response, ResponsesAPIResponse
), "Response should be an instance of ResponsesAPIResponse"
# Required fields
assert "id" in response and isinstance(
response["id"], str
), "Response should have a string 'id' field"
assert "created_at" in response and isinstance(
response["created_at"], (int, float)
), "Response should have a numeric 'created_at' field"
assert "output" in response and isinstance(
response["output"], list
), "Response should have a list 'output' field"
assert "parallel_tool_calls" in response and isinstance(
response["parallel_tool_calls"], bool
), "Response should have a boolean 'parallel_tool_calls' field"
# Optional fields with their expected types
optional_fields = {
"error": (dict, type(None)), # error can be dict or None
"incomplete_details": (IncompleteDetails, type(None)),
"instructions": (str, type(None)),
"metadata": dict,
"model": str,
"object": str,
"temperature": (int, float, type(None)),
"tool_choice": (dict, str),
"tools": list,
"top_p": (int, float, type(None)),
"max_output_tokens": (int, type(None)),
"previous_response_id": (str, type(None)),
"reasoning": dict,
"status": str,
"text": ResponseTextConfig,
"truncation": (str, type(None)),
"usage": ResponseAPIUsage,
"user": (str, type(None)),
}
if final_chunk is False:
optional_fields["usage"] = type(None)
for field, expected_type in optional_fields.items():
if field in response:
assert isinstance(
response[field], expected_type
), f"Field '{field}' should be of type {expected_type}, but got {type(response[field])}"
# Check if output has at least one item
if final_chunk is True:
assert (
len(response["output"]) > 0
), "Response 'output' field should have at least one item"
return True # Return True if validation passes
class BaseResponsesAPITest(ABC):
"""
Abstract base test class that enforces a common test across all test classes.
"""
@abstractmethod
def get_base_completion_call_args(self) -> dict:
"""Must return the base completion call args"""
pass
@pytest.mark.parametrize("sync_mode", [True, False])
@pytest.mark.asyncio
async def test_basic_openai_responses_api(self, sync_mode):
litellm._turn_on_debug()
litellm.set_verbose = True
base_completion_call_args = self.get_base_completion_call_args()
try:
if sync_mode:
response = litellm.responses(
input="Basic ping", max_output_tokens=20,
**base_completion_call_args
)
else:
response = await litellm.aresponses(
input="Basic ping", max_output_tokens=20,
**base_completion_call_args
)
except litellm.InternalServerError:
pytest.skip("Skipping test due to litellm.InternalServerError")
print("litellm response=", json.dumps(response, indent=4, default=str))
# Use the helper function to validate the response
validate_responses_api_response(response, final_chunk=True)
@pytest.mark.parametrize("sync_mode", [True, False])
@pytest.mark.asyncio
async def test_basic_openai_responses_api_streaming(self, sync_mode):
litellm._turn_on_debug()
base_completion_call_args = self.get_base_completion_call_args()
collected_content_string = ""
response_completed_event = None
if sync_mode:
response = litellm.responses(
input="Basic ping",
stream=True,
**base_completion_call_args
)
for event in response:
print("litellm response=", json.dumps(event, indent=4, default=str))
if event.type == "response.output_text.delta":
collected_content_string += event.delta
elif event.type == "response.completed":
response_completed_event = event
else:
response = await litellm.aresponses(
input="Basic ping",
stream=True,
**base_completion_call_args
)
async for event in response:
print("litellm response=", json.dumps(event, indent=4, default=str))
if event.type == "response.output_text.delta":
collected_content_string += event.delta
elif event.type == "response.completed":
response_completed_event = event
# assert the delta chunks content had len(collected_content_string) > 0
# this content is typically rendered on chat ui's
assert len(collected_content_string) > 0
# assert the response completed event is not None
assert response_completed_event is not None
# assert the response completed event has a response
assert response_completed_event.response is not None
# assert the response completed event includes the usage
assert response_completed_event.response.usage is not None
# basic test assert the usage seems reasonable
print("response_completed_event.response.usage=", response_completed_event.response.usage)
assert response_completed_event.response.usage.input_tokens > 0 and response_completed_event.response.usage.input_tokens < 100
assert response_completed_event.response.usage.output_tokens > 0 and response_completed_event.response.usage.output_tokens < 1000
assert response_completed_event.response.usage.total_tokens > 0 and response_completed_event.response.usage.total_tokens < 1000
# total tokens should be the sum of input and output tokens
assert response_completed_event.response.usage.total_tokens == response_completed_event.response.usage.input_tokens + response_completed_event.response.usage.output_tokens
@pytest.mark.parametrize("sync_mode", [False, True])
@pytest.mark.asyncio
async def test_basic_openai_responses_delete_endpoint(self, sync_mode):
litellm._turn_on_debug()
litellm.set_verbose = True
base_completion_call_args = self.get_base_completion_call_args()
if sync_mode:
response = litellm.responses(
input="Basic ping", max_output_tokens=20,
**base_completion_call_args
)
# delete the response
if isinstance(response, ResponsesAPIResponse):
litellm.delete_responses(
response_id=response.id,
**base_completion_call_args
)
else:
raise ValueError("response is not a ResponsesAPIResponse")
else:
response = await litellm.aresponses(
input="Basic ping", max_output_tokens=20,
**base_completion_call_args
)
# async delete the response
if isinstance(response, ResponsesAPIResponse):
await litellm.adelete_responses(
response_id=response.id,
**base_completion_call_args
)
else:
raise ValueError("response is not a ResponsesAPIResponse")
@pytest.mark.parametrize("sync_mode", [True, False])
@pytest.mark.asyncio
async def test_basic_openai_responses_streaming_delete_endpoint(self, sync_mode):
#litellm._turn_on_debug()
#litellm.set_verbose = True
base_completion_call_args = self.get_base_completion_call_args()
response_id = None
if sync_mode:
response_id = None
response = litellm.responses(
input="Basic ping", max_output_tokens=20,
stream=True,
**base_completion_call_args
)
for event in response:
print("litellm response=", json.dumps(event, indent=4, default=str))
if "response" in event:
response_obj = event.get("response")
if response_obj is not None:
response_id = response_obj.get("id")
print("got response_id=", response_id)
# delete the response
assert response_id is not None
litellm.delete_responses(
response_id=response_id,
**base_completion_call_args
)
else:
response = await litellm.aresponses(
input="Basic ping", max_output_tokens=20,
stream=True,
**base_completion_call_args
)
async for event in response:
print("litellm response=", json.dumps(event, indent=4, default=str))
if "response" in event:
response_obj = event.get("response")
if response_obj is not None:
response_id = response_obj.get("id")
print("got response_id=", response_id)
# delete the response
assert response_id is not None
await litellm.adelete_responses(
response_id=response_id,
**base_completion_call_args
)
@pytest.mark.parametrize("sync_mode", [False, True])
@pytest.mark.asyncio
async def test_basic_openai_responses_get_endpoint(self, sync_mode):
litellm._turn_on_debug()
litellm.set_verbose = True
base_completion_call_args = self.get_base_completion_call_args()
if sync_mode:
response = litellm.responses(
input="Basic ping", max_output_tokens=20,
**base_completion_call_args
)
# get the response
if isinstance(response, ResponsesAPIResponse):
result = litellm.get_responses(
response_id=response.id,
**base_completion_call_args
)
assert result is not None
assert result.id == response.id
assert result.output == response.output
else:
raise ValueError("response is not a ResponsesAPIResponse")
else:
response = await litellm.aresponses(
input="Basic ping", max_output_tokens=20,
**base_completion_call_args
)
# async get the response
if isinstance(response, ResponsesAPIResponse):
result = await litellm.aget_responses(
response_id=response.id,
**base_completion_call_args
)
assert result is not None
assert result.id == response.id
assert result.output == response.output
else:
raise ValueError("response is not a ResponsesAPIResponse")
@pytest.mark.asyncio
async def test_basic_openai_list_input_items_endpoint(self):
"""Test that calls the OpenAI List Input Items endpoint"""
litellm._turn_on_debug()
response = await litellm.aresponses(
model="gpt-4o",
input="Tell me a three sentence bedtime story about a unicorn.",
)
print("Initial response=", json.dumps(response, indent=4, default=str))
response_id = response.get("id")
assert response_id is not None, "Response should have an ID"
print(f"Got response_id: {response_id}")
list_items_response = await litellm.alist_input_items(
response_id=response_id,
limit=20,
order="desc",
)
print(
"List items response=",
json.dumps(list_items_response, indent=4, default=str),
)
@pytest.mark.asyncio
async def test_multiturn_responses_api(self):
litellm._turn_on_debug()
litellm.set_verbose = True
base_completion_call_args = self.get_base_completion_call_args()
response_1 = await litellm.aresponses(
input="Basic ping", max_output_tokens=20, **base_completion_call_args
)
# follow up with a second request
response_1_id = response_1.id
response_2 = await litellm.aresponses(
input="Basic ping",
max_output_tokens=20,
previous_response_id=response_1_id,
**base_completion_call_args
)
# assert the response is not None
assert response_1 is not None
assert response_2 is not None
|