feat: add a subagent frame
This commit is contained in:
tcmofashi
0
agentlite/tests/integration/__init__.py
Normal file
0
agentlite/tests/integration/__init__.py
Normal file
286
agentlite/tests/integration/test_agent.py
Normal file
286
agentlite/tests/integration/test_agent.py
Normal file
@@ -0,0 +1,286 @@
|
||||
"""Integration tests for Agent class.
|
||||
|
||||
This module tests the Agent class with mocked providers to verify
|
||||
core functionality without making real API calls.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import pytest
|
||||
|
||||
from agentlite import Agent, TextPart
|
||||
|
||||
|
||||
@pytest.mark.integration
|
||||
class TestAgentInitialization:
|
||||
"""Tests for Agent initialization."""
|
||||
|
||||
def test_agent_initialization(self, mock_provider):
|
||||
"""Test basic agent creation."""
|
||||
agent = Agent(provider=mock_provider)
|
||||
|
||||
assert agent.provider is mock_provider
|
||||
assert agent.system_prompt == "You are a helpful assistant."
|
||||
assert agent.max_iterations == 80
|
||||
assert agent.history == []
|
||||
|
||||
def test_agent_with_custom_system_prompt(self, mock_provider):
|
||||
"""Test agent creation with custom system prompt."""
|
||||
agent = Agent(provider=mock_provider, system_prompt="You are a specialized assistant.")
|
||||
|
||||
assert agent.system_prompt == "You are a specialized assistant."
|
||||
|
||||
def test_agent_with_tools(self, mock_provider, add_tool):
|
||||
"""Test agent creation with tools."""
|
||||
agent = Agent(provider=mock_provider, tools=[add_tool])
|
||||
|
||||
assert len(agent.tools.tools) == 1
|
||||
assert agent.tools.tools[0].name == "add"
|
||||
|
||||
def test_agent_with_custom_max_iterations(self, mock_provider):
|
||||
"""Test agent with custom max_iterations."""
|
||||
agent = Agent(provider=mock_provider, max_iterations=5)
|
||||
|
||||
assert agent.max_iterations == 5
|
||||
|
||||
|
||||
@pytest.mark.integration
|
||||
class TestAgentRun:
|
||||
"""Tests for Agent.run() method."""
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_run_simple(self, mock_provider):
|
||||
"""Test simple non-streaming run."""
|
||||
mock_provider.add_text_response("Hello there!")
|
||||
agent = Agent(provider=mock_provider)
|
||||
|
||||
response = await agent.run("Hi")
|
||||
|
||||
assert response == "Hello there!"
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_run_adds_to_history(self, mock_provider):
|
||||
"""Test that run adds messages to history."""
|
||||
mock_provider.add_text_response("Response!")
|
||||
agent = Agent(provider=mock_provider)
|
||||
|
||||
await agent.run("Hello")
|
||||
|
||||
# History should have user message and assistant response
|
||||
assert len(agent.history) == 2
|
||||
assert agent.history[0].role == "user"
|
||||
assert agent.history[0].extract_text() == "Hello"
|
||||
assert agent.history[1].role == "assistant"
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_run_multiple_messages(self, mock_provider):
|
||||
"""Test multiple runs accumulate history."""
|
||||
mock_provider.add_text_responses("Response 1", "Response 2")
|
||||
agent = Agent(provider=mock_provider)
|
||||
|
||||
await agent.run("Message 1")
|
||||
await agent.run("Message 2")
|
||||
|
||||
# Should have 4 messages total
|
||||
assert len(agent.history) == 4
|
||||
assert agent.history[0].role == "user"
|
||||
assert agent.history[1].role == "assistant"
|
||||
assert agent.history[2].role == "user"
|
||||
assert agent.history[3].role == "assistant"
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_run_tracks_calls(self, mock_provider):
|
||||
"""Test that provider.generate is called during run."""
|
||||
mock_provider.add_text_response("Response!")
|
||||
agent = Agent(provider=mock_provider)
|
||||
|
||||
await agent.run("Hello")
|
||||
|
||||
assert len(mock_provider.calls) == 1
|
||||
call = mock_provider.calls[0]
|
||||
assert call["system_prompt"] == "You are a helpful assistant."
|
||||
assert len(call["history"]) == 1 # User message
|
||||
|
||||
|
||||
@pytest.mark.integration
|
||||
class TestAgentGenerate:
|
||||
"""Tests for Agent.generate() method."""
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_generate_returns_message(self, mock_provider):
|
||||
"""Test that generate returns a Message."""
|
||||
mock_provider.add_text_response("Generated response")
|
||||
agent = Agent(provider=mock_provider)
|
||||
|
||||
message = await agent.generate("Hello")
|
||||
|
||||
assert message.role == "assistant"
|
||||
assert message.extract_text() == "Generated response"
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_generate_without_tool_loop(self, mock_provider):
|
||||
"""Test that generate doesn't do tool calling loop."""
|
||||
# Add tool call response
|
||||
mock_provider.add_tool_call("add", {"a": 1, "b": 2}, "3")
|
||||
agent = Agent(provider=mock_provider, tools=[])
|
||||
|
||||
message = await agent.generate("Calculate 1+2")
|
||||
|
||||
# Should return the tool call without executing it
|
||||
assert message.has_tool_calls()
|
||||
assert len(message.tool_calls) == 1
|
||||
assert message.tool_calls[0].function.name == "add"
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_generate_adds_to_history(self, mock_provider):
|
||||
"""Test that generate adds response to history."""
|
||||
mock_provider.add_text_response("Response!")
|
||||
agent = Agent(provider=mock_provider)
|
||||
|
||||
await agent.generate("Hello")
|
||||
|
||||
assert len(agent.history) == 2
|
||||
assert agent.history[1].role == "assistant"
|
||||
|
||||
|
||||
@pytest.mark.integration
|
||||
class TestAgentHistory:
|
||||
"""Tests for Agent history management."""
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_history_property_returns_copy(self, mock_provider):
|
||||
"""Test that history property returns a copy."""
|
||||
mock_provider.add_text_response("Response!")
|
||||
agent = Agent(provider=mock_provider)
|
||||
|
||||
await agent.run("Hello")
|
||||
|
||||
history = agent.history
|
||||
history.clear() # Modify the copy
|
||||
|
||||
# Original should still have messages
|
||||
assert len(agent.history) == 2
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_clear_history(self, mock_provider):
|
||||
"""Test clearing history."""
|
||||
mock_provider.add_text_response("Response!")
|
||||
agent = Agent(provider=mock_provider)
|
||||
|
||||
await agent.run("Hello")
|
||||
agent.clear_history()
|
||||
|
||||
assert agent.history == []
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_add_message(self, mock_provider):
|
||||
"""Test manually adding a message."""
|
||||
agent = Agent(provider=mock_provider)
|
||||
|
||||
from agentlite import Message
|
||||
|
||||
agent.add_message(Message(role="user", content="Manual message"))
|
||||
|
||||
assert len(agent.history) == 1
|
||||
assert agent.history[0].extract_text() == "Manual message"
|
||||
|
||||
|
||||
@pytest.mark.integration
|
||||
class TestAgentWithTools:
|
||||
"""Tests for Agent with tools."""
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_with_tools_initialization(self, mock_provider, add_tool):
|
||||
"""Test agent initialization with tools."""
|
||||
agent = Agent(
|
||||
provider=mock_provider, tools=[add_tool], system_prompt="You have access to tools."
|
||||
)
|
||||
|
||||
assert len(agent.tools.tools) == 1
|
||||
|
||||
# Run to verify tools are passed to provider
|
||||
mock_provider.add_text_response("I have tools available")
|
||||
await agent.run("Hello")
|
||||
|
||||
# Check that tools were passed to provider
|
||||
assert len(mock_provider.calls) == 1
|
||||
assert len(mock_provider.calls[0]["tools"]) == 1
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_tool_call_execution(self, mock_provider, add_tool):
|
||||
"""Test that agent executes tool calls."""
|
||||
# First response: tool call
|
||||
mock_provider.add_tool_call("add", {"a": 1, "b": 2}, "3")
|
||||
# Second response: text after tool result
|
||||
mock_provider.add_text_response("The sum is 3")
|
||||
|
||||
agent = Agent(provider=mock_provider, tools=[add_tool])
|
||||
|
||||
response = await agent.run("What is 1+2?")
|
||||
|
||||
assert "3" in response
|
||||
# Should have made 2 calls to provider
|
||||
assert len(mock_provider.calls) == 2
|
||||
|
||||
|
||||
@pytest.mark.integration
|
||||
class TestAgentMaxIterations:
|
||||
"""Tests for max_iterations behavior."""
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_respects_max_iterations(self, mock_provider, add_tool):
|
||||
"""Test that agent stops after max_iterations."""
|
||||
# Always return tool calls to trigger iteration limit
|
||||
for _ in range(10):
|
||||
mock_provider.add_tool_call("add", {"a": 1, "b": 2}, "3")
|
||||
|
||||
agent = Agent(provider=mock_provider, tools=[add_tool], max_iterations=3)
|
||||
|
||||
response = await agent.run("Calculate")
|
||||
|
||||
# Should stop after max_iterations
|
||||
assert len(mock_provider.calls) <= 3
|
||||
assert "Maximum tool call iterations reached" in response
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_no_iterations_for_simple_response(self, mock_provider):
|
||||
"""Test that simple responses don't count as iterations."""
|
||||
mock_provider.add_text_response("Simple response")
|
||||
agent = Agent(provider=mock_provider, max_iterations=1)
|
||||
|
||||
response = await agent.run("Hello")
|
||||
|
||||
assert response == "Simple response"
|
||||
|
||||
|
||||
@pytest.mark.integration
|
||||
class TestAgentStreaming:
|
||||
"""Tests for streaming mode."""
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_run_streaming(self, mock_provider):
|
||||
"""Test streaming run."""
|
||||
mock_provider.add_text_response("Streamed response")
|
||||
agent = Agent(provider=mock_provider)
|
||||
|
||||
stream = await agent.run("Hello", stream=True)
|
||||
|
||||
# Collect stream
|
||||
chunks = []
|
||||
async for chunk in stream:
|
||||
chunks.append(chunk)
|
||||
|
||||
assert len(chunks) > 0
|
||||
assert "".join(chunks) == "Streamed response"
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_agent_streaming_adds_to_history(self, mock_provider):
|
||||
"""Test that streaming adds messages to history."""
|
||||
mock_provider.add_text_response("Response")
|
||||
agent = Agent(provider=mock_provider)
|
||||
|
||||
stream = await agent.run("Hello", stream=True)
|
||||
async for _ in stream:
|
||||
pass
|
||||
|
||||
assert len(agent.history) == 2
|
||||
348
agentlite/tests/integration/test_with_api.py
Normal file
348
agentlite/tests/integration/test_with_api.py
Normal file
@@ -0,0 +1,348 @@
|
||||
"""Integration tests for AgentLite with real API.
|
||||
|
||||
This script runs comprehensive tests against the real OpenAI API.
|
||||
Requires OPENAI_API_KEY environment variable to be set.
|
||||
|
||||
Usage:
|
||||
export OPENAI_API_KEY="sk-..."
|
||||
python tests/integration/test_with_api.py
|
||||
"""
|
||||
|
||||
import asyncio
|
||||
import os
|
||||
import sys
|
||||
from pathlib import Path
|
||||
import pytest
|
||||
|
||||
# Add src to path
|
||||
sys.path.insert(0, str(Path(__file__).parent.parent.parent / "src"))
|
||||
|
||||
from agentlite import Agent, OpenAIProvider, LLMClient, llm_complete
|
||||
from agentlite.skills import discover_skills, SkillTool, index_skills_by_name
|
||||
from agentlite.tools import ConfigurableToolset
|
||||
|
||||
|
||||
# Test configuration
|
||||
TEST_MODEL = "gpt-4o-mini" # Use mini for cost efficiency
|
||||
HAS_OPENAI_API_KEY = bool(os.environ.get("OPENAI_API_KEY"))
|
||||
|
||||
pytestmark = pytest.mark.skipif(
|
||||
not HAS_OPENAI_API_KEY, reason="OPENAI_API_KEY is required to run integration tests"
|
||||
)
|
||||
|
||||
|
||||
def get_provider():
|
||||
"""Get OpenAI provider with API key."""
|
||||
api_key = os.environ.get("OPENAI_API_KEY")
|
||||
if not api_key:
|
||||
print("❌ OPENAI_API_KEY not set!")
|
||||
print("Please set your OpenAI API key:")
|
||||
print(" export OPENAI_API_KEY='sk-...'")
|
||||
sys.exit(1)
|
||||
return OpenAIProvider(api_key=api_key, model=TEST_MODEL)
|
||||
|
||||
|
||||
async def test_basic_agent():
|
||||
"""Test 1: Basic Agent functionality."""
|
||||
print("\n" + "=" * 60)
|
||||
print("Test 1: Basic Agent Functionality")
|
||||
print("=" * 60)
|
||||
|
||||
try:
|
||||
provider = get_provider()
|
||||
agent = Agent(
|
||||
provider=provider,
|
||||
system_prompt="You are a helpful assistant. Be concise.",
|
||||
)
|
||||
|
||||
response = await agent.run("What is 2+2?")
|
||||
print(f"✅ Agent responded: {response[:100]}...")
|
||||
|
||||
assert "4" in response, "Expected '4' in response"
|
||||
print("✅ Basic Agent test PASSED")
|
||||
return True
|
||||
|
||||
except Exception as e:
|
||||
print(f"❌ Basic Agent test FAILED: {e}")
|
||||
return False
|
||||
|
||||
|
||||
async def test_agent_with_tools():
|
||||
"""Test 2: Agent with tool suite."""
|
||||
print("\n" + "=" * 60)
|
||||
print("Test 2: Agent with Tool Suite")
|
||||
print("=" * 60)
|
||||
|
||||
try:
|
||||
from agentlite.tools import ToolSuiteConfig, ReadFile, Glob
|
||||
|
||||
provider = get_provider()
|
||||
|
||||
# Create toolset with file tools
|
||||
config = ToolSuiteConfig()
|
||||
toolset = ConfigurableToolset(config, work_dir=Path.cwd())
|
||||
|
||||
agent = Agent(
|
||||
provider=provider,
|
||||
system_prompt="You are a helpful assistant with file access.",
|
||||
tools=toolset.tools,
|
||||
)
|
||||
|
||||
print(f"✅ Agent created with {len(agent.tools.tools)} tools")
|
||||
|
||||
# Test simple query (without requiring file access)
|
||||
response = await agent.run("List the Python files in the current directory")
|
||||
print(f"✅ Agent with tools responded: {response[:100]}...")
|
||||
|
||||
print("✅ Agent with Tools test PASSED")
|
||||
return True
|
||||
|
||||
except Exception as e:
|
||||
print(f"❌ Agent with Tools test FAILED: {e}")
|
||||
import traceback
|
||||
|
||||
traceback.print_exc()
|
||||
return False
|
||||
|
||||
|
||||
async def test_llm_client():
|
||||
"""Test 3: LLMClient functionality."""
|
||||
print("\n" + "=" * 60)
|
||||
print("Test 3: LLMClient Functionality")
|
||||
print("=" * 60)
|
||||
|
||||
try:
|
||||
provider = get_provider()
|
||||
client = LLMClient(provider=provider)
|
||||
|
||||
response = await client.complete(
|
||||
user_prompt="What is the capital of France?",
|
||||
system_prompt="You are a helpful assistant. Be concise.",
|
||||
)
|
||||
|
||||
print(f"✅ LLMClient responded: {response.content[:100]}...")
|
||||
assert "Paris" in response.content, "Expected 'Paris' in response"
|
||||
|
||||
print("✅ LLMClient test PASSED")
|
||||
return True
|
||||
|
||||
except Exception as e:
|
||||
print(f"❌ LLMClient test FAILED: {e}")
|
||||
import traceback
|
||||
|
||||
traceback.print_exc()
|
||||
return False
|
||||
|
||||
|
||||
async def test_llm_streaming():
|
||||
"""Test 4: LLM streaming."""
|
||||
print("\n" + "=" * 60)
|
||||
print("Test 4: LLM Streaming")
|
||||
print("=" * 60)
|
||||
|
||||
try:
|
||||
provider = get_provider()
|
||||
client = LLMClient(provider=provider)
|
||||
|
||||
chunks = []
|
||||
async for chunk in client.stream(
|
||||
user_prompt="Count from 1 to 3",
|
||||
system_prompt="You are a helpful assistant.",
|
||||
):
|
||||
chunks.append(chunk)
|
||||
print(f" Chunk: {chunk[:20]}...")
|
||||
|
||||
full_response = "".join(chunks)
|
||||
print(f"✅ Streamed response: {full_response[:100]}...")
|
||||
|
||||
print("✅ LLM Streaming test PASSED")
|
||||
return True
|
||||
|
||||
except Exception as e:
|
||||
print(f"❌ LLM Streaming test FAILED: {e}")
|
||||
import traceback
|
||||
|
||||
traceback.print_exc()
|
||||
return False
|
||||
|
||||
|
||||
async def test_subagents():
|
||||
"""Test 5: Subagent functionality."""
|
||||
print("\n" + "=" * 60)
|
||||
print("Test 5: Subagent Functionality")
|
||||
print("=" * 60)
|
||||
|
||||
try:
|
||||
from agentlite.labor_market import LaborMarket
|
||||
from agentlite.tools.multiagent.task import Task
|
||||
|
||||
provider = get_provider()
|
||||
|
||||
# Create parent agent
|
||||
parent = Agent(
|
||||
provider=provider,
|
||||
system_prompt="You are a coordinator agent.",
|
||||
name="coordinator",
|
||||
)
|
||||
|
||||
# Create subagent
|
||||
coder = Agent(
|
||||
provider=provider,
|
||||
system_prompt="You are a coding specialist. Write clean, simple code.",
|
||||
name="coder",
|
||||
)
|
||||
|
||||
# Add subagent to parent
|
||||
parent.add_subagent("coder", coder, "Writes code")
|
||||
|
||||
# Add Task tool
|
||||
parent.tools.add(Task(labor_market=parent.labor_market))
|
||||
|
||||
print(f"✅ Created parent with {len(parent.labor_market)} subagent(s)")
|
||||
print(f" Subagents: {parent.labor_market.list_subagents()}")
|
||||
|
||||
print("✅ Subagent test PASSED")
|
||||
return True
|
||||
|
||||
except Exception as e:
|
||||
print(f"❌ Subagent test FAILED: {e}")
|
||||
import traceback
|
||||
|
||||
traceback.print_exc()
|
||||
return False
|
||||
|
||||
|
||||
async def test_skills():
|
||||
"""Test 6: Skills functionality."""
|
||||
print("\n" + "=" * 60)
|
||||
print("Test 6: Skills Functionality")
|
||||
print("=" * 60)
|
||||
|
||||
try:
|
||||
# Discover example skills
|
||||
skills_dir = Path(__file__).parent.parent.parent / "examples" / "skills"
|
||||
if not skills_dir.exists():
|
||||
print("⚠️ Skills directory not found, skipping")
|
||||
return True
|
||||
|
||||
skills = discover_skills(skills_dir)
|
||||
print(f"✅ Discovered {len(skills)} skill(s)")
|
||||
|
||||
for skill in skills:
|
||||
print(f" - {skill.name} ({skill.type})")
|
||||
|
||||
if len(skills) == 0:
|
||||
print("⚠️ No skills found, skipping further tests")
|
||||
return True
|
||||
|
||||
# Test with agent
|
||||
provider = get_provider()
|
||||
agent = Agent(
|
||||
provider=provider,
|
||||
system_prompt="You are a helpful assistant.",
|
||||
)
|
||||
|
||||
skill_index = index_skills_by_name(skills)
|
||||
skill_tool = SkillTool(skill_index, parent_agent=agent)
|
||||
agent.tools.add(skill_tool)
|
||||
|
||||
print(f"✅ Added SkillTool to agent")
|
||||
print("✅ Skills test PASSED")
|
||||
return True
|
||||
|
||||
except Exception as e:
|
||||
print(f"❌ Skills test FAILED: {e}")
|
||||
import traceback
|
||||
|
||||
traceback.print_exc()
|
||||
return False
|
||||
|
||||
|
||||
async def test_conversation_history():
|
||||
"""Test 7: Conversation history."""
|
||||
print("\n" + "=" * 60)
|
||||
print("Test 7: Conversation History")
|
||||
print("=" * 60)
|
||||
|
||||
try:
|
||||
provider = get_provider()
|
||||
agent = Agent(
|
||||
provider=provider,
|
||||
system_prompt="You are a helpful assistant.",
|
||||
)
|
||||
|
||||
# First message
|
||||
response1 = await agent.run("My name is Alice")
|
||||
print(f"✅ Response 1: {response1[:50]}...")
|
||||
|
||||
# Second message (should remember context)
|
||||
response2 = await agent.run("What is my name?")
|
||||
print(f"✅ Response 2: {response2[:50]}...")
|
||||
|
||||
assert "Alice" in response2, "Expected agent to remember name"
|
||||
|
||||
print("✅ Conversation History test PASSED")
|
||||
return True
|
||||
|
||||
except Exception as e:
|
||||
print(f"❌ Conversation History test FAILED: {e}")
|
||||
import traceback
|
||||
|
||||
traceback.print_exc()
|
||||
return False
|
||||
|
||||
|
||||
async def run_all_tests():
|
||||
"""Run all integration tests."""
|
||||
print("\n" + "=" * 60)
|
||||
print("AgentLite Integration Tests with Real API")
|
||||
print("=" * 60)
|
||||
print(f"Model: {TEST_MODEL}")
|
||||
|
||||
# Check API key
|
||||
if not os.environ.get("OPENAI_API_KEY"):
|
||||
print("\n❌ OPENAI_API_KEY not set!")
|
||||
print("\nTo run these tests, set your OpenAI API key:")
|
||||
print(" export OPENAI_API_KEY='sk-...'")
|
||||
print("\nGet your API key from: https://platform.openai.com/api-keys")
|
||||
return []
|
||||
|
||||
results = []
|
||||
|
||||
# Run all tests
|
||||
results.append(("Basic Agent", await test_basic_agent()))
|
||||
results.append(("Agent with Tools", await test_agent_with_tools()))
|
||||
results.append(("LLMClient", await test_llm_client()))
|
||||
results.append(("LLM Streaming", await test_llm_streaming()))
|
||||
results.append(("Subagents", await test_subagents()))
|
||||
results.append(("Skills", await test_skills()))
|
||||
results.append(("Conversation History", await test_conversation_history()))
|
||||
|
||||
# Print summary
|
||||
print("\n" + "=" * 60)
|
||||
print("Test Summary")
|
||||
print("=" * 60)
|
||||
|
||||
passed = sum(1 for _, result in results if result)
|
||||
total = len(results)
|
||||
|
||||
for name, result in results:
|
||||
status = "✅ PASSED" if result else "❌ FAILED"
|
||||
print(f"{status}: {name}")
|
||||
|
||||
print(f"\n{passed}/{total} tests passed")
|
||||
|
||||
if passed == total:
|
||||
print("\n🎉 All tests passed!")
|
||||
else:
|
||||
print(f"\n⚠️ {total - passed} test(s) failed")
|
||||
|
||||
return results
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
results = asyncio.run(run_all_tests())
|
||||
|
||||
# Exit with error code if any tests failed
|
||||
if results and not all(r for _, r in results):
|
||||
sys.exit(1)
|
||||
Reference in New Issue
Block a user