jarviscore-framework 0.1.1__py3-none-any.whl → 0.2.1__py3-none-any.whl

tests/test_07_distributed_single_node.py

@@ -0,0 +1,578 @@
+"""
+Test 7: Distributed Mode - Single Node Tests
+
+Tests the core distributed mode functionality on a single node:
+- P2P coordinator + Workflow engine working together
+- Workflow execution with dependency resolution
+- Step result broadcasting
+- Status tracking and memory management
+
+This file uses MOCKED agents (no real LLM) to test the infrastructure.
+Real LLM tests are in test_09 and test_10.
+
+Run with: pytest tests/test_07_distributed_single_node.py -v
+"""
+import asyncio
+import sys
+import pytest
+
+sys.path.insert(0, '.')
+
+from jarviscore import Mesh, MeshMode, Agent
+
+
+# ═══════════════════════════════════════════════════════════════════════════════
+# TEST AGENTS (Mocked - No LLM)
+# ═══════════════════════════════════════════════════════════════════════════════
+
+class DataGeneratorAgent(Agent):
+    """Generates sample data."""
+    role = "generator"
+    capabilities = ["data_generation", "sampling"]
+
+    def __init__(self, agent_id=None):
+        super().__init__(agent_id)
+        self.tasks_executed = []
+
+    async def execute_task(self, task):
+        self.tasks_executed.append(task)
+        return {
+            "status": "success",
+            "output": {
+                "data": [
+                    {"id": 1, "name": "Alice", "score": 85},
+                    {"id": 2, "name": "Bob", "score": 92},
+                    {"id": 3, "name": "Charlie", "score": 78}
+                ],
+                "count": 3
+            },
+            "agent": self.agent_id
+        }
+
+
+class DataAnalyzerAgent(Agent):
+    """Analyzes data and produces insights."""
+    role = "analyzer"
+    capabilities = ["analysis", "statistics"]
+
+    def __init__(self, agent_id=None):
+        super().__init__(agent_id)
+        self.tasks_executed = []
+        self.context_received = []
+
+    async def execute_task(self, task):
+        self.tasks_executed.append(task)
+
+        # Capture context from dependencies
+        context = task.get("context", {})
+        self.context_received.append(context)
+
+        # Use previous step results if available
+        previous_results = context.get("previous_step_results", {})
+
+        return {
+            "status": "success",
+            "output": {
+                "analysis": "Positive trend detected",
+                "mean_score": 85.0,
+                "received_context": bool(previous_results)
+            },
+            "agent": self.agent_id
+        }
+
+
+class DataStorageAgent(Agent):
+    """Stores data to a destination."""
+    role = "storage"
+    capabilities = ["storage", "persistence"]
+
+    def __init__(self, agent_id=None):
+        super().__init__(agent_id)
+        self.tasks_executed = []
+
+    async def execute_task(self, task):
+        self.tasks_executed.append(task)
+        return {
+            "status": "success",
+            "output": {"saved": True, "records": 3},
+            "agent": self.agent_id
+        }
+
+
+class FailingAgent(Agent):
+    """Agent that always fails (for error testing)."""
+    role = "failing"
+    capabilities = ["failure"]
+
+    async def execute_task(self, task):
+        raise RuntimeError("Intentional failure for testing")
+
+
+# ═══════════════════════════════════════════════════════════════════════════════
+# TEST CLASS: Distributed Mode Initialization
+# ═══════════════════════════════════════════════════════════════════════════════
+
+class TestDistributedModeInitialization:
+    """Tests for distributed mode setup and initialization."""
+
+    def test_mesh_creates_in_distributed_mode(self):
+        """Mesh should initialize with distributed mode."""
+        mesh = Mesh(mode="distributed")
+
+        assert mesh.mode == MeshMode.DISTRIBUTED
+        assert mesh._started is False
+        assert mesh._p2p_coordinator is None  # Not started yet
+        assert mesh._workflow_engine is None  # Not started yet
+
+    @pytest.mark.asyncio
+    async def test_p2p_coordinator_initialized(self):
+        """P2P coordinator should be initialized on start."""
+        mesh = Mesh(mode="distributed", config={'bind_port': 7950})
+        mesh.add(DataGeneratorAgent)
+
+        await mesh.start()
+
+        try:
+            assert mesh._p2p_coordinator is not None
+            assert mesh._p2p_coordinator._started is True
+        finally:
+            await mesh.stop()
+
+    @pytest.mark.asyncio
+    async def test_workflow_engine_initialized(self):
+        """Workflow engine should be initialized on start."""
+        mesh = Mesh(mode="distributed", config={'bind_port': 7951})
+        mesh.add(DataGeneratorAgent)
+
+        await mesh.start()
+
+        try:
+            assert mesh._workflow_engine is not None
+            assert mesh._workflow_engine._started is True
+        finally:
+            await mesh.stop()
+
+    @pytest.mark.asyncio
+    async def test_both_components_initialized(self):
+        """Both P2P coordinator AND workflow engine should be present."""
+        mesh = Mesh(mode="distributed", config={'bind_port': 7952})
+        mesh.add(DataGeneratorAgent)
+
+        await mesh.start()
+
+        try:
+            # This is the KEY difference from other modes
+            assert mesh._p2p_coordinator is not None, "Distributed needs P2P"
+            assert mesh._workflow_engine is not None, "Distributed needs Workflow"
+        finally:
+            await mesh.stop()
+
+    @pytest.mark.asyncio
+    async def test_capabilities_announced(self):
+        """Agent capabilities should be announced to P2P mesh."""
+        mesh = Mesh(mode="distributed", config={'bind_port': 7953})
+        mesh.add(DataGeneratorAgent)
+        mesh.add(DataAnalyzerAgent)
+
+        await mesh.start()
+
+        try:
+            # Check capability map was populated
+            cap_map = mesh._p2p_coordinator._capability_map
+            assert "data_generation" in cap_map or len(cap_map) > 0
+        finally:
+            await mesh.stop()
+
+
+# ═══════════════════════════════════════════════════════════════════════════════
+# TEST CLASS: Workflow Execution in Distributed Mode
+# ═══════════════════════════════════════════════════════════════════════════════
+
+class TestDistributedWorkflowExecution:
+    """Tests for workflow execution in distributed mode."""
+
+    @pytest.fixture
+    async def distributed_mesh(self):
+        """Create a distributed mesh with multiple agents."""
+        mesh = Mesh(mode="distributed", config={'bind_port': 7954})
+        generator = mesh.add(DataGeneratorAgent)
+        analyzer = mesh.add(DataAnalyzerAgent)
+        storage = mesh.add(DataStorageAgent)
+
+        await mesh.start()
+
+        yield mesh, generator, analyzer, storage
+
+        await mesh.stop()
+
+    @pytest.mark.asyncio
+    async def test_single_step_workflow(self, distributed_mesh):
+        """Execute a single-step workflow."""
+        mesh, generator, analyzer, storage = distributed_mesh
+
+        results = await mesh.workflow("test-single", [
+            {"agent": "generator", "task": "Generate sample data"}
+        ])
+
+        assert len(results) == 1
+        assert results[0]["status"] == "success"
+        assert "data" in results[0]["output"]
+        assert len(generator.tasks_executed) == 1
+
+    @pytest.mark.asyncio
+    async def test_multi_step_workflow(self, distributed_mesh):
+        """Execute a multi-step workflow without dependencies."""
+        mesh, generator, analyzer, storage = distributed_mesh
+
+        results = await mesh.workflow("test-multi", [
+            {"agent": "generator", "task": "Generate data"},
+            {"agent": "analyzer", "task": "Analyze data"},
+            {"agent": "storage", "task": "Store results"}
+        ])
+
+        assert len(results) == 3
+        assert all(r["status"] == "success" for r in results)
+        assert len(generator.tasks_executed) == 1
+        assert len(analyzer.tasks_executed) == 1
+        assert len(storage.tasks_executed) == 1
+
+    @pytest.mark.asyncio
+    async def test_workflow_with_dependencies(self, distributed_mesh):
+        """Execute workflow where steps depend on previous steps."""
+        mesh, generator, analyzer, storage = distributed_mesh
+
+        results = await mesh.workflow("test-deps", [
+            {"agent": "generator", "task": "Generate user data"},
+            {"agent": "analyzer", "task": "Analyze the data", "depends_on": [0]},
+            {"agent": "storage", "task": "Store analysis", "depends_on": [1]}
+        ])
+
+        assert len(results) == 3
+        assert all(r["status"] == "success" for r in results)
+
+        # Verify analyzer received context from generator
+        assert len(analyzer.context_received) == 1
+        context = analyzer.context_received[0]
+        assert "previous_step_results" in context
+        assert "step0" in context["previous_step_results"]
+
+    @pytest.mark.asyncio
+    async def test_workflow_with_missing_agent(self, distributed_mesh):
+        """Workflow should fail gracefully when agent not found."""
+        mesh, generator, analyzer, storage = distributed_mesh
+
+        results = await mesh.workflow("test-missing", [
+            {"agent": "nonexistent_agent", "task": "Should fail"}
+        ])
+
+        assert len(results) == 1
+        assert results[0]["status"] == "failure"
+        assert "No agent found" in results[0]["error"]
+
+    @pytest.mark.asyncio
+    async def test_workflow_routes_by_capability(self, distributed_mesh):
+        """Workflow should route by capability when role not found."""
+        mesh, generator, analyzer, storage = distributed_mesh
+
+        results = await mesh.workflow("test-capability", [
+            {"agent": "data_generation", "task": "Use capability routing"}
+        ])
+
+        assert len(results) == 1
+        assert results[0]["status"] == "success"
+        # Should have been routed to generator agent
+        assert len(generator.tasks_executed) == 1
+
+    @pytest.mark.asyncio
+    async def test_step_results_stored_in_memory(self, distributed_mesh):
+        """Step results should be stored in workflow memory."""
+        mesh, generator, analyzer, storage = distributed_mesh
+
+        await mesh.workflow("test-memory", [
+            {"id": "gen-step", "agent": "generator", "task": "Generate"},
+            {"id": "analyze-step", "agent": "analyzer", "task": "Analyze", "depends_on": ["gen-step"]}
+        ])
+
+        # Check workflow engine memory
+        memory = mesh._workflow_engine.get_memory()
+        assert "gen-step" in memory
+        assert "analyze-step" in memory
+
+
+# ═══════════════════════════════════════════════════════════════════════════════
+# TEST CLASS: Step Broadcasting
+# ═══════════════════════════════════════════════════════════════════════════════
+
+class TestDistributedStepBroadcasting:
+    """Tests for P2P step result broadcasting."""
+
+    @pytest.mark.asyncio
+    async def test_broadcaster_initialized(self):
+        """Broadcaster should be initialized with P2P coordinator."""
+        mesh = Mesh(mode="distributed", config={'bind_port': 7955})
+        mesh.add(DataGeneratorAgent)
+
+        await mesh.start()
+
+        try:
+            assert mesh._p2p_coordinator.broadcaster is not None
+        finally:
+            await mesh.stop()
+
+    @pytest.mark.asyncio
+    async def test_status_manager_tracks_steps(self):
+        """Status manager should track step states."""
+        mesh = Mesh(mode="distributed", config={'bind_port': 7956})
+        mesh.add(DataGeneratorAgent)
+
+        await mesh.start()
+
+        try:
+            await mesh.workflow("test-status", [
+                {"id": "tracked-step", "agent": "generator", "task": "Track me"}
+            ])
+
+            # Check status was tracked
+            status = mesh._workflow_engine.get_status("tracked-step")
+            assert status is not None
+            assert status["status"] == "completed"
+        finally:
+            await mesh.stop()
+
+    @pytest.mark.asyncio
+    async def test_failed_step_status_tracked(self):
+        """Failed steps should have failure status."""
+        mesh = Mesh(mode="distributed", config={'bind_port': 7957})
+        mesh.add(FailingAgent)
+
+        await mesh.start()
+
+        try:
+            await mesh.workflow("test-fail-status", [
+                {"id": "failing-step", "agent": "failing", "task": "Will fail"}
+            ])
+
+            status = mesh._workflow_engine.get_status("failing-step")
+            assert status is not None
+            assert status["status"] == "failed"
+        finally:
+            await mesh.stop()
+
+
+# ═══════════════════════════════════════════════════════════════════════════════
+# TEST CLASS: Distributed Mode Lifecycle
+# ═══════════════════════════════════════════════════════════════════════════════
+
+class TestDistributedModeLifecycle:
+    """Tests for distributed mode start/stop lifecycle."""
+
+    @pytest.mark.asyncio
+    async def test_graceful_shutdown(self):
+        """Mesh should shutdown gracefully."""
+        mesh = Mesh(mode="distributed", config={'bind_port': 7958})
+        mesh.add(DataGeneratorAgent)
+
+        await mesh.start()
+        assert mesh._started is True
+
+        await mesh.stop()
+        assert mesh._started is False
+
+    @pytest.mark.asyncio
+    async def test_workflow_fails_before_start(self):
+        """Workflow should fail if mesh not started."""
+        mesh = Mesh(mode="distributed", config={'bind_port': 7959})
+        mesh.add(DataGeneratorAgent)
+
+        # Don't start the mesh
+        with pytest.raises(RuntimeError) as exc_info:
+            await mesh.workflow("test", [{"agent": "generator", "task": "fail"}])
+
+        assert "not started" in str(exc_info.value)
+
+    @pytest.mark.asyncio
+    async def test_agent_setup_called(self):
+        """Agent setup should be called on mesh start."""
+        setup_called = []
+
+        class TrackingAgent(Agent):
+            role = "tracker"
+            capabilities = ["tracking"]
+
+            async def setup(self):
+                await super().setup()
+                setup_called.append(self.agent_id)
+
+            async def execute_task(self, task):
+                return {"status": "success"}
+
+        mesh = Mesh(mode="distributed", config={'bind_port': 7960})
+        agent = mesh.add(TrackingAgent)
+
+        await mesh.start()
+
+        try:
+            assert agent.agent_id in setup_called
+        finally:
+            await mesh.stop()
+
+    @pytest.mark.asyncio
+    async def test_agent_teardown_called(self):
+        """Agent teardown should be called on mesh stop."""
+        teardown_called = []
+
+        class TrackingAgent(Agent):
+            role = "tracker"
+            capabilities = ["tracking"]
+
+            async def teardown(self):
+                await super().teardown()
+                teardown_called.append(self.agent_id)
+
+            async def execute_task(self, task):
+                return {"status": "success"}
+
+        mesh = Mesh(mode="distributed", config={'bind_port': 7961})
+        agent = mesh.add(TrackingAgent)
+
+        await mesh.start()
+        await mesh.stop()
+
+        assert agent.agent_id in teardown_called
+
+
+# ═══════════════════════════════════════════════════════════════════════════════
+# TEST CLASS: Comparison with Other Modes
+# ═══════════════════════════════════════════════════════════════════════════════
+
+class TestDistributedVsOtherModes:
+    """Tests that verify distributed mode differs from other modes."""
+
+    def test_autonomous_mode_has_no_p2p(self):
+        """Autonomous mode should NOT have P2P coordinator."""
+        mesh = Mesh(mode="autonomous")
+        mesh.add(DataGeneratorAgent)
+
+        # Before start - no P2P
+        assert mesh._p2p_coordinator is None
+
+    @pytest.mark.asyncio
+    async def test_autonomous_starts_without_p2p(self):
+        """Autonomous mode should start without P2P."""
+        mesh = Mesh(mode="autonomous")
+        mesh.add(DataGeneratorAgent)
+
+        await mesh.start()
+
+        try:
+            # Autonomous has workflow engine but NOT P2P
+            assert mesh._workflow_engine is not None
+            assert mesh._p2p_coordinator is None
+        finally:
+            await mesh.stop()
+
+    @pytest.mark.asyncio
+    async def test_p2p_mode_has_no_workflow_engine(self):
+        """P2P mode should NOT have workflow engine."""
+        mesh = Mesh(mode="p2p", config={'bind_port': 7962})
+
+        class P2PAgent(Agent):
+            role = "p2p_agent"
+            capabilities = ["p2p"]
+
+            async def execute_task(self, task):
+                return {"status": "success"}
+
+            async def run(self):
+                while not self.shutdown_requested:
+                    await asyncio.sleep(0.1)
+
+        mesh.add(P2PAgent)
+
+        await mesh.start()
+
+        try:
+            # P2P has P2P coordinator but NOT workflow engine
+            assert mesh._p2p_coordinator is not None
+            assert mesh._workflow_engine is None
+        finally:
+            await mesh.stop()
+
+    @pytest.mark.asyncio
+    async def test_distributed_has_both(self):
+        """Distributed mode should have BOTH P2P and workflow engine."""
+        mesh = Mesh(mode="distributed", config={'bind_port': 7963})
+        mesh.add(DataGeneratorAgent)
+
+        await mesh.start()
+
+        try:
+            # Distributed has BOTH
+            assert mesh._p2p_coordinator is not None, "Missing P2P coordinator"
+            assert mesh._workflow_engine is not None, "Missing workflow engine"
+        finally:
+            await mesh.stop()
+
+
+# ═══════════════════════════════════════════════════════════════════════════════
+# MANUAL RUN
+# ═══════════════════════════════════════════════════════════════════════════════
+
+async def run_manual_test():
+    """Run a manual demonstration of distributed mode."""
+    print("\n" + "="*70)
+    print("DISTRIBUTED MODE - SINGLE NODE DEMONSTRATION")
+    print("="*70)
+
+    # Create mesh
+    mesh = Mesh(mode="distributed", config={'bind_port': 7964})
+
+    generator = mesh.add(DataGeneratorAgent)
+    analyzer = mesh.add(DataAnalyzerAgent)
+    storage = mesh.add(DataStorageAgent)
+
+    print(f"\n[SETUP] Created mesh with {len(mesh.agents)} agents")
+    print(f"  - Generator: {generator.role} ({generator.capabilities})")
+    print(f"  - Analyzer: {analyzer.role} ({analyzer.capabilities})")
+    print(f"  - Storage: {storage.role} ({storage.capabilities})")
+
+    # Start mesh
+    await mesh.start()
+    print(f"\n[START] Mesh started")
+    print(f"  - P2P Coordinator: {mesh._p2p_coordinator is not None}")
+    print(f"  - Workflow Engine: {mesh._workflow_engine is not None}")
+
+    # Execute workflow
+    print(f"\n[WORKFLOW] Executing pipeline with dependencies...")
+
+    results = await mesh.workflow("demo-pipeline", [
+        {"id": "generate", "agent": "generator", "task": "Generate user data"},
+        {"id": "analyze", "agent": "analyzer", "task": "Analyze patterns", "depends_on": ["generate"]},
+        {"id": "store", "agent": "storage", "task": "Save results", "depends_on": ["analyze"]}
+    ])
+
+    print(f"\n[RESULTS]")
+    for i, result in enumerate(results):
+        status = result.get("status", "unknown")
+        agent = result.get("agent", "unknown")
+        print(f"  Step {i+1}: {status} (executed by {agent})")
+
+    # Show memory
+    print(f"\n[MEMORY] Workflow memory contains:")
+    for step_id, data in mesh._workflow_engine.get_memory().items():
+        print(f"  - {step_id}: {type(data).__name__}")
+
+    # Show context injection
+    print(f"\n[CONTEXT] Analyzer received context: {analyzer.context_received[0].get('previous_step_results', {}).keys()}")
+
+    # Stop mesh
+    await mesh.stop()
+    print(f"\n[STOP] Mesh stopped gracefully")
+
+    print("\n" + "="*70)
+    print("DEMONSTRATION COMPLETE")
+    print("="*70)
+
+
+if __name__ == "__main__":
+    asyncio.run(run_manual_test())