Spaces:
Running
Running
File size: 10,202 Bytes
9024ad9 0497d92 2c658eb 9024ad9 0497d92 9024ad9 0497d92 2c658eb 0497d92 2c658eb 0497d92 2c658eb 0497d92 2c658eb 0497d92 2c658eb 0497d92 2c658eb 0497d92 |
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 |
"""
Tests for service layer.
"""
import pytest
from unittest.mock import patch, MagicMock
import httpx
from app.services.summarizer import OllamaService
class StubAsyncResponse:
"""A minimal stub of an httpx.Response-like object for testing."""
def __init__(self, json_data=None, status_code=200, raise_for_status_exc=None):
self._json_data = json_data or {}
self.status_code = status_code
self._raise_for_status_exc = raise_for_status_exc
def json(self):
return self._json_data
def raise_for_status(self):
if self._raise_for_status_exc is not None:
raise self._raise_for_status_exc
class StubAsyncClient:
"""An async context manager stub that mimics httpx.AsyncClient for tests."""
def __init__(self, post_result=None, post_exc=None, get_result=None, get_exc=None, *args, **kwargs):
self._post_result = post_result
self._post_exc = post_exc
self._get_result = get_result
self._get_exc = get_exc
async def __aenter__(self):
return self
async def __aexit__(self, exc_type, exc, tb):
return False
async def post(self, *args, **kwargs):
if self._post_exc is not None:
raise self._post_exc
return self._post_result or StubAsyncResponse()
async def get(self, *args, **kwargs):
if self._get_exc is not None:
raise self._get_exc
return self._get_result or StubAsyncResponse(status_code=200)
class TestOllamaService:
"""Test Ollama service."""
@pytest.fixture
def ollama_service(self):
"""Create Ollama service instance."""
return OllamaService()
def test_service_initialization(self, ollama_service):
"""Test service initialization."""
assert ollama_service.base_url == "http://127.0.0.1:11434"
assert ollama_service.model == "llama3.2:latest" # Updated to match current config
assert ollama_service.timeout == 60 # Updated to match current config
@pytest.mark.asyncio
async def test_summarize_text_success(self, ollama_service, mock_ollama_response):
"""Test successful text summarization."""
stub_response = StubAsyncResponse(json_data=mock_ollama_response)
with patch('httpx.AsyncClient', return_value=StubAsyncClient(post_result=stub_response)):
result = await ollama_service.summarize_text("Test text")
assert result["summary"] == mock_ollama_response["response"]
assert result["model"] == "llama3.2:latest" # Updated to match current config
assert result["tokens_used"] == mock_ollama_response["eval_count"]
assert "latency_ms" in result
@pytest.mark.asyncio
async def test_summarize_text_with_custom_params(self, ollama_service, mock_ollama_response):
"""Test summarization with custom parameters."""
stub_response = StubAsyncResponse(json_data=mock_ollama_response)
# Patch with a factory to capture payload for assertion
captured = {}
class CapturePostClient(StubAsyncClient):
async def post(self, *args, **kwargs):
captured['json'] = kwargs.get('json')
return await super().post(*args, **kwargs)
with patch('httpx.AsyncClient', return_value=CapturePostClient(post_result=stub_response)):
result = await ollama_service.summarize_text(
"Test text",
max_tokens=512,
prompt="Custom prompt"
)
assert result["summary"] == mock_ollama_response["response"]
# Verify captured payload
payload = captured['json']
assert payload["options"]["num_predict"] == 512
assert "Custom prompt" in payload["prompt"]
@pytest.mark.asyncio
async def test_summarize_text_timeout(self, ollama_service):
"""Test timeout handling."""
with patch('httpx.AsyncClient', return_value=StubAsyncClient(post_exc=httpx.TimeoutException("Timeout"))):
with pytest.raises(httpx.HTTPError, match="Ollama API timeout"):
await ollama_service.summarize_text("Test text")
@pytest.mark.asyncio
async def test_summarize_text_http_error(self, ollama_service):
"""Test HTTP error handling."""
http_error = httpx.HTTPStatusError("Bad Request", request=MagicMock(), response=MagicMock())
stub_response = StubAsyncResponse(raise_for_status_exc=http_error)
with patch('httpx.AsyncClient', return_value=StubAsyncClient(post_result=stub_response)):
with pytest.raises(httpx.HTTPError):
await ollama_service.summarize_text("Test text")
@pytest.mark.asyncio
async def test_check_health_success(self, ollama_service):
"""Test successful health check."""
stub_response = StubAsyncResponse(status_code=200)
with patch('httpx.AsyncClient', return_value=StubAsyncClient(get_result=stub_response)):
result = await ollama_service.check_health()
assert result is True
@pytest.mark.asyncio
async def test_check_health_failure(self, ollama_service):
"""Test health check failure."""
with patch('httpx.AsyncClient', return_value=StubAsyncClient(get_exc=httpx.HTTPError("Connection failed"))):
result = await ollama_service.check_health()
assert result is False
# Tests for Dynamic Timeout System
@pytest.mark.asyncio
async def test_dynamic_timeout_small_text(self, ollama_service, mock_ollama_response):
"""Test dynamic timeout calculation for small text (should use base timeout)."""
stub_response = StubAsyncResponse(json_data=mock_ollama_response)
captured_timeout = None
class TimeoutCaptureClient(StubAsyncClient):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.timeout = None
async def __aenter__(self):
return self
async def post(self, *args, **kwargs):
return await super().post(*args, **kwargs)
with patch('httpx.AsyncClient') as mock_client:
mock_client.return_value = TimeoutCaptureClient(post_result=stub_response)
mock_client.return_value.timeout = 120 # Base timeout
result = await ollama_service.summarize_text("Short text")
# Verify the client was called with the base timeout
mock_client.assert_called_once()
call_args = mock_client.call_args
assert call_args[1]['timeout'] == 120
@pytest.mark.asyncio
async def test_dynamic_timeout_large_text(self, ollama_service, mock_ollama_response):
"""Test dynamic timeout calculation for large text (should extend timeout)."""
stub_response = StubAsyncResponse(json_data=mock_ollama_response)
large_text = "A" * 5000 # 5000 characters
with patch('httpx.AsyncClient') as mock_client:
mock_client.return_value = StubAsyncClient(post_result=stub_response)
result = await ollama_service.summarize_text(large_text)
# Verify the client was called with extended timeout
# Expected: 30s base + (5000-1000)/1000 * 10 = 30 + 40 = 70s
mock_client.assert_called_once()
call_args = mock_client.call_args
expected_timeout = 60 + (5000 - 1000) // 1000 * 5 # 80 seconds
assert call_args[1]['timeout'] == expected_timeout
@pytest.mark.asyncio
async def test_dynamic_timeout_maximum_cap(self, ollama_service, mock_ollama_response):
"""Test that dynamic timeout is capped at 2 minutes (120 seconds)."""
stub_response = StubAsyncResponse(json_data=mock_ollama_response)
very_large_text = "A" * 50000 # 50000 characters (should exceed 120s cap)
with patch('httpx.AsyncClient') as mock_client:
mock_client.return_value = StubAsyncClient(post_result=stub_response)
result = await ollama_service.summarize_text(very_large_text)
# Verify the timeout is capped at 120 seconds
mock_client.assert_called_once()
call_args = mock_client.call_args
assert call_args[1]['timeout'] == 120 # Maximum cap
@pytest.mark.asyncio
async def test_dynamic_timeout_logging(self, ollama_service, mock_ollama_response, caplog):
"""Test that dynamic timeout calculation is logged correctly."""
stub_response = StubAsyncResponse(json_data=mock_ollama_response)
test_text = "A" * 2500 # 2500 characters
with patch('httpx.AsyncClient', return_value=StubAsyncClient(post_result=stub_response)):
await ollama_service.summarize_text(test_text)
# Check that the logging message contains the correct information
log_messages = [record.message for record in caplog.records]
timeout_log = next((msg for msg in log_messages if "Processing text of" in msg), None)
assert timeout_log is not None
assert "2500 characters" in timeout_log
assert "timeout of" in timeout_log
@pytest.mark.asyncio
async def test_timeout_error_message_improvement(self, ollama_service):
"""Test that timeout errors now include dynamic timeout and text length info."""
test_text = "A" * 2000 # 2000 characters
expected_timeout = 60 + (2000 - 1000) // 1000 * 5 # 65 seconds
with patch('httpx.AsyncClient', return_value=StubAsyncClient(post_exc=httpx.TimeoutException("Timeout"))):
with pytest.raises(httpx.HTTPError) as exc_info:
await ollama_service.summarize_text(test_text)
# Verify the error message includes the dynamic timeout and text length
error_message = str(exc_info.value)
assert f"timeout after {expected_timeout}s" in error_message
assert "Text may be too long or complex" in error_message
|