refactor(agents): migrate code to hub (#1397, #1102) #2799

Workflow file for this run

.github/workflows/test_gaia_cli_linux.yml at be066a2

	# Copyright(C) 2025-2026 Advanced Micro Devices, Inc. All rights reserved.
	# SPDX-License-Identifier: MIT

	# This workflow tests the GAIA CLI on Linux (Ubuntu) with Lemonade server support
	# Tests include: CLI installation, Lemonade server setup, core commands, and evaluation tools
	# Platform: Linux/Ubuntu (Full CLI functionality with Lemonade server)

	name: GAIA CLI Tests (Linux)

	on:
	workflow_call:
	push:
	branches: [ main ]
	pull_request:
	branches: [ main ]
	types: [opened, synchronize, reopened, ready_for_review]
	merge_group:
	workflow_dispatch:

	# Cancel in-progress runs when a new run is triggered
	concurrency:
	group: ${{ github.workflow }}-${{ github.head_ref \|\| github.ref }}
	cancel-in-progress: true

	permissions:
	contents: read

	jobs:
	test-gaia-cli-linux:
	name: Test GAIA CLI on Linux (Full Integration)
	runs-on: ubuntu-latest
	timeout-minutes: 15
	if: github.event_name != 'pull_request' \|\| github.event.pull_request.draft == false \|\| contains(github.event.pull_request.labels.*.name, 'ready_for_ci')

	steps:
	- name: Checkout repository
	uses: actions/checkout@v6

	- name: Free disk space
	uses: ./.github/actions/free-disk-space

	- name: Install system dependencies
	run: \|
	sudo apt-get update
	sudo apt-get install -y build-essential curl jq libportaudio2

	- name: Setup Python environment
	uses: ./.github/actions/setup-venv
	with:
	python-version: '3.12'
	# Note: 'blender' excluded as bpy requires Blender runtime
	install-package: '.[dev,talk,rag,mcp,api]'
	extra-index-url: 'https://download.pytorch.org/whl/cpu'

	- name: Install Lemonade SDK
	run: \|
	echo "=== Installing Lemonade SDK ==="
	uv pip install lemonade-sdk --python .venv/bin/python

	echo "=== Verifying Lemonade Installation ==="
	lemonade-server-dev -h

	- name: Verify GAIA CLI installation
	run: \|
	echo "=== Testing GAIA CLI Installation ==="

	# Test version command
	echo "Testing version command:"
	gaia --version

	# Test help command
	echo "Testing help command:"
	gaia --help

	# Verify gaia command is available
	which gaia
	echo "GAIA CLI installed successfully!"

	- name: Start Lemonade Server and Test Core Commands
	env:
	# Authenticate HF model pulls — unauthenticated requests get 429-throttled
	# on shared CI runners, which kills Lemonade startup (matches the Windows job).
	HUGGINGFACE_ACCESS_TOKEN: ${{ secrets.HUGGINGFACE_ACCESS_TOKEN }}
	HF_TOKEN: ${{ secrets.HUGGINGFACE_ACCESS_TOKEN }}
	run: \|
	echo "=== Starting Lemonade Server ==="

	# Start Lemonade server in background with a lightweight model
	echo "Starting Lemonade server with Qwen model..."
	echo "Note: First run may take 5-10 minutes to download the model (~3GB)"

	# Start server and capture both stdout and stderr
	lemonade-server-dev run Qwen3-0.6B-GGUF > lemonade.log 2>&1 &
	LEMONADE_PID=$!
	echo "Lemonade server started with PID: $LEMONADE_PID"

	# Function to check if process is still running
	check_process_alive() {
	if ! kill -0 $LEMONADE_PID 2>/dev/null; then
	echo "❌ Lemonade server process has died. Log output:"
	cat lemonade.log
	return 1
	fi
	return 0
	}

	# Wait for server to be ready (5 min hard cap, poll every 10s)
	echo "Waiting for Lemonade server to start..."
	max_wait=300 # 5 minutes
	elapsed=0

	while [ $elapsed -lt $max_wait ]; do
	if ! check_process_alive; then
	exit 1
	fi

	if curl -sf -m 5 http://localhost:8000/api/v1/health >/dev/null 2>&1; then
	echo "✅ Lemonade server is responding! (${elapsed}s)"
	break
	fi

	sleep 10
	elapsed=$((elapsed + 10))

	if [ $((elapsed % 60)) -eq 0 ]; then
	echo "Still waiting... (${elapsed}s elapsed)"
	tail -3 lemonade.log 2>/dev/null \|\| true
	fi
	done

	if [ $elapsed -ge $max_wait ]; then
	echo "❌ Lemonade server failed to start within ${max_wait}s"
	cat lemonade.log
	kill $LEMONADE_PID 2>/dev/null \|\| true
	exit 1
	fi

	# Pull the model now that server is running
	echo "=== Pulling Qwen3-0.6B-GGUF model ==="
	lemonade-server-dev pull Qwen3-0.6B-GGUF

	# List available models for debugging
	echo "=== Listing Available Models ==="
	curl -s http://localhost:8000/api/v1/models \| jq '.' \|\| echo "Could not list models"

	# Python lemonade-server-dev runs on port 8000; tell GAIA CLI where to connect
	export LEMONADE_BASE_URL=http://localhost:8000/api/v1

	echo "=== Testing Core GAIA CLI Commands with Lemonade ==="

	# Test chat command with Qwen model (should now work with Lemonade)
	echo "Testing chat command with Qwen3-0.6B-GGUF model:"
	timeout 30s gaia chat --model "Qwen3-0.6B-GGUF" -q "Hello, this is a test message. Please respond briefly." \|\| echo "Chat command completed"

	# Test prompt command with Qwen model
	echo "Testing prompt command with Qwen3-0.6B-GGUF model:"
	timeout 30s gaia prompt --model "Qwen3-0.6B-GGUF" "What is 2+2?" \|\| echo "Prompt command completed"

	# Test llm command with Qwen model
	echo "Testing llm command with Qwen3-0.6B-GGUF model:"
	timeout 30s gaia llm --model "Qwen3-0.6B-GGUF" "Say hello" \|\| echo "LLM command completed"

	echo "✅ Core commands tested successfully!"

	# Now test summarizer while server is still running
	echo ""
	echo "=== Testing Summarizer Integration ==="
	echo "Testing complete CLI workflow: gaia summarize command"

	# Set the test model for summarizer tests
	export GAIA_TEST_MODEL="Qwen3-0.6B-GGUF"

	# Run the summarizer integration tests
	python -m pytest tests/test_summarizer.py -vs --tb=short \|\| TEST_EXIT=$?

	if [ "${TEST_EXIT:-0}" -eq 0 ]; then
	echo "✅ Summarizer CLI integration tests passed successfully!"
	else
	echo "❌ Summarizer CLI integration tests failed with exit code: ${TEST_EXIT}"
	echo "Note: Error details displayed above"
	fi

	echo ""
	echo "=== Testing RAG Functionality ==="
	echo "Testing RAG (Retrieval-Augmented Generation) system"

	# Run the RAG tests
	python -m pytest tests/test_rag.py -vs --tb=short \|\| RAG_TEST_EXIT=$?

	if [ "${RAG_TEST_EXIT:-0}" -eq 0 ]; then
	echo "✅ RAG tests passed successfully!"
	else
	echo "❌ RAG tests failed with exit code: ${RAG_TEST_EXIT}"
	echo "Note: Error details displayed above"
	fi

	echo ""
	echo "=== Testing Lemonade Client Integration ==="
	echo "Testing LemonadeClient API with running server"

	# Run the lemonade client integration tests (skip hybrid NPU test - no NPU on Linux)
	# LEMONADE_PORT=8000: lemonade-server-dev always binds to 8000 (no --port flag)
	LEMONADE_PORT=8000 GAIA_TEST_MODEL="Qwen3-0.6B-GGUF" python -m pytest tests/test_lemonade_client.py -vs --tb=short -k "Integration and not hybrid" \|\| LEMONADE_TEST_EXIT=$?

	if [ "${LEMONADE_TEST_EXIT:-0}" -eq 0 ]; then
	echo "✅ Lemonade client integration tests passed successfully!"
	else
	echo "❌ Lemonade client integration tests failed with exit code: ${LEMONADE_TEST_EXIT}"
	echo "Note: Error details displayed above"
	fi

	# Clean up: Stop the Lemonade server
	echo "Stopping Lemonade server..."
	kill $LEMONADE_PID 2>/dev/null \|\| true
	sleep 5

	# Report failures — summarizer/RAG tests may fail when the agent's
	# default model isn't available (only Qwen3-0.6B-GGUF is pulled in CI).
	# Surface as a GHA warning so failures are visible without blocking the job.
	FIRST_FAILURE=0
	for code in "${TEST_EXIT:-0}" "${RAG_TEST_EXIT:-0}" "${LEMONADE_TEST_EXIT:-0}"; do
	if [ "$code" -ne 0 ] && [ "$FIRST_FAILURE" -eq 0 ]; then
	FIRST_FAILURE=$code
	fi
	done
	if [ "$FIRST_FAILURE" -ne 0 ]; then
	echo "::warning::Integration tests had failures (exit code $FIRST_FAILURE) — see logs above"
	fi

	- name: Test evaluation and utility commands
	run: \|
	echo "=== Testing Evaluation and Utility Commands ==="

	# Test eval command help
	echo "Testing eval command help:"
	gaia eval --help

	# Test eval agent subcommand help
	echo "Testing eval agent subcommand help:"
	gaia eval agent --help

	# Test test command help
	echo "Testing test command help:"
	gaia test --help

	# Test youtube command help
	echo "Testing youtube command help:"
	gaia youtube --help

	# Test kill command help
	echo "Testing kill command help:"
	gaia kill --help

	- name: Test platform compatibility
	run: \|
	echo "=== Testing Platform Compatibility ==="

	# Test that Linux-specific process commands work
	echo "Testing kill command with invalid port (should handle gracefully):"
	if gaia kill --port 99999 2>&1 \| grep -q "No process found"; then
	echo "✅ Kill command handled Linux environment gracefully"
	else
	echo "⚠️ Kill command may not be fully Linux-compatible"
	fi

	- name: Test import statements
	run: \|
	echo "=== Testing Python Import Compatibility ==="

	# Test that all modules can be imported on Linux
	python -c "
	try:
	from gaia.cli import main
	from gaia.version import version
	from gaia.logger import get_logger
	from gaia.llm import LLMClient
	print('✅ All core imports successful on Linux')
	except ImportError as e:
	print(f'❌ Import error: {e}')
	exit(1)
	"

	- name: Summary
	run: \|
	echo "=== Linux Full Integration Test Summary ==="
	echo "✅ GAIA CLI installation works on Linux"
	echo "✅ Python venv environment setup successful"
	echo "✅ Lemonade server installation and startup successful"
	echo "✅ Core CLI commands (chat, prompt, llm) working with Lemonade"
	echo "✅ Evaluation and utility commands working"
	echo "✅ Cross-platform process management working"
	echo "✅ Core Python modules import successfully"
	echo ""
	echo "🎉 Linux now has full GAIA CLI functionality!"
	echo "📋 Tested components:"
	echo " ✅ Lemonade server with Qwen3-0.6B-GGUF model"
	echo " ✅ Core CLI commands using Qwen model (chat, prompt, llm)"
	echo " ✅ Evaluation commands (eval, eval agent)"
	echo " ✅ Summarizer CLI integration (gaia summarize command)"
	echo " ✅ RAG (Retrieval-Augmented Generation) functionality"
	echo " ✅ Lemonade client integration (API tests)"
	echo " ✅ Cross-platform compatibility"
	echo " ✅ Process management and cleanup"
	echo ""
	echo "🚀 Next enhancements:"
	echo " 🔧 Test audio/TTS functionality on Linux"
	echo " 🔧 Add Linux NPU driver support (if available)"
	echo " 🔧 Optimize model loading and performance"

	- name: Upload logs on failure
	if: always()
	uses: actions/upload-artifact@v7
	with:
	name: gaia-cli-linux-logs
	path: \|
	lemonade.log
	if-no-files-found: ignore

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

refactor(agents): migrate code to hub (#1397, #1102) #2799

Workflow file

refactor(agents): migrate code to hub (#1397, #1102) #2799

Uh oh!

Workflow file for this run