mirror of
https://github.com/proffesor-for-testing/agentic-qe.git
synced 2026-09-19 08:45:47 +08:00
Profa
c9aa5fca9a
## Agent Performance Optimizations - CoverageAnalyzerAgent: O(n²) → O(n) using Map/Set lookups - FlakyTestHunterAgent: Fixed timestamp deserialization from JSON - PerformanceTracker: Fixed Date deserialization from stored data - QualityGateAgent: Null check for context.changes, improved reasoning ## Initialization Robustness - database-init.ts: Defensive directory creation ## New Features - Release verification script: `npm run verify:release` - Automated e2e verification before publishing - Tests: aqe init, hooks, MCP server, learning capture ## Tests - Fixed journey test assertions to match agent behavior - Adjusted CI thresholds for environment variability - All journey tests pass (58 passed, 12 skipped) 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
186 lines
7.9 KiB
Bash
Executable File
186 lines
7.9 KiB
Bash
Executable File
#!/bin/bash
|
|
#
|
|
# Release Verification Script for Agentic QE
|
|
#
|
|
# Performs end-to-end verification of a release before publishing:
|
|
# 1. Creates a fresh test project
|
|
# 2. Runs aqe init -y
|
|
# 3. Verifies all files are copied correctly
|
|
# 4. Tests MCP server starts
|
|
# 5. Tests learning capture flow
|
|
#
|
|
# Usage: npm run verify:release
|
|
# ./scripts/verify-release.sh
|
|
#
|
|
# Exit codes:
|
|
# 0 - All checks passed
|
|
# 1 - One or more checks failed
|
|
|
|
set -e
|
|
|
|
# Colors for output
|
|
RED='\033[0;31m'
|
|
GREEN='\033[0;32m'
|
|
YELLOW='\033[1;33m'
|
|
BLUE='\033[0;34m'
|
|
NC='\033[0m' # No Color
|
|
|
|
# Script directory (for resolving paths)
|
|
SCRIPT_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
|
|
PROJECT_ROOT="$( cd "$SCRIPT_DIR/.." && pwd )"
|
|
TEST_PROJECT="/tmp/aqe-release-verification-$$"
|
|
FAILED=0
|
|
|
|
echo -e "${BLUE}═══════════════════════════════════════════════════════════════${NC}"
|
|
echo -e "${BLUE} Agentic QE Release Verification${NC}"
|
|
echo -e "${BLUE}═══════════════════════════════════════════════════════════════${NC}"
|
|
echo ""
|
|
|
|
# Get version from package.json
|
|
VERSION=$(node -p "require('$PROJECT_ROOT/package.json').version")
|
|
echo -e "Version: ${YELLOW}$VERSION${NC}"
|
|
echo -e "Test project: ${YELLOW}$TEST_PROJECT${NC}"
|
|
echo ""
|
|
|
|
# Cleanup on exit
|
|
cleanup() {
|
|
if [ -d "$TEST_PROJECT" ]; then
|
|
rm -rf "$TEST_PROJECT"
|
|
fi
|
|
}
|
|
trap cleanup EXIT
|
|
|
|
# Helper function for check results
|
|
check_result() {
|
|
local name="$1"
|
|
local result="$2"
|
|
if [ "$result" -eq 0 ]; then
|
|
echo -e " ${GREEN}✓${NC} $name"
|
|
else
|
|
echo -e " ${RED}✗${NC} $name"
|
|
FAILED=1
|
|
fi
|
|
}
|
|
|
|
# ============================================================================
|
|
# Step 1: Build the project
|
|
# ============================================================================
|
|
echo -e "${YELLOW}[1/6] Building project...${NC}"
|
|
cd "$PROJECT_ROOT"
|
|
npm run build > /dev/null 2>&1
|
|
check_result "TypeScript compilation" $?
|
|
|
|
# ============================================================================
|
|
# Step 2: Create test project
|
|
# ============================================================================
|
|
echo -e "${YELLOW}[2/6] Creating test project...${NC}"
|
|
mkdir -p "$TEST_PROJECT"
|
|
cd "$TEST_PROJECT"
|
|
npm init -y > /dev/null 2>&1
|
|
check_result "Test project created" $?
|
|
|
|
# ============================================================================
|
|
# Step 3: Run aqe init -y
|
|
# ============================================================================
|
|
echo -e "${YELLOW}[3/6] Running aqe init -y...${NC}"
|
|
node "$PROJECT_ROOT/dist/cli/index.js" init -y > /tmp/aqe-init-output.txt 2>&1
|
|
INIT_RESULT=$?
|
|
check_result "aqe init -y completed" $INIT_RESULT
|
|
|
|
# Verify directories created
|
|
[ -d ".agentic-qe" ] && check_result "Directory: .agentic-qe" 0 || check_result "Directory: .agentic-qe" 1
|
|
[ -d ".claude" ] && check_result "Directory: .claude" 0 || check_result "Directory: .claude" 1
|
|
[ -d "scripts/hooks" ] && check_result "Directory: scripts/hooks" 0 || check_result "Directory: scripts/hooks" 1
|
|
|
|
# Verify hook scripts copied
|
|
[ -f "scripts/hooks/capture-task-learning.js" ] && check_result "Hook: capture-task-learning.js" 0 || check_result "Hook: capture-task-learning.js" 1
|
|
[ -f "scripts/hooks/emit-task-spawn.sh" ] && check_result "Hook: emit-task-spawn.sh" 0 || check_result "Hook: emit-task-spawn.sh" 1
|
|
[ -f "scripts/hooks/emit-task-complete.sh" ] && check_result "Hook: emit-task-complete.sh" 0 || check_result "Hook: emit-task-complete.sh" 1
|
|
|
|
# Verify Claude configuration
|
|
[ -f ".claude/settings.json" ] && check_result "File: .claude/settings.json" 0 || check_result "File: .claude/settings.json" 1
|
|
[ -d ".claude/agents" ] && check_result "Directory: .claude/agents" 0 || check_result "Directory: .claude/agents" 1
|
|
[ -d ".claude/skills" ] && check_result "Directory: .claude/skills" 0 || check_result "Directory: .claude/skills" 1
|
|
[ -d ".claude/commands" ] && check_result "Directory: .claude/commands" 0 || check_result "Directory: .claude/commands" 1
|
|
|
|
# Verify database created
|
|
[ -f ".agentic-qe/memory.db" ] && check_result "Database: memory.db" 0 || check_result "Database: memory.db" 1
|
|
|
|
# ============================================================================
|
|
# Step 4: Verify MCP server starts
|
|
# ============================================================================
|
|
echo -e "${YELLOW}[4/6] Testing MCP server...${NC}"
|
|
timeout 5 "$PROJECT_ROOT/bin/aqe-mcp" > /tmp/aqe-mcp-output.txt 2>&1 || true
|
|
if grep -q "Agentic QE MCP Server started successfully" /tmp/aqe-mcp-output.txt; then
|
|
check_result "MCP server starts" 0
|
|
|
|
# Count tools
|
|
TOOL_COUNT=$(grep -o "mcp__agentic_qe__" /tmp/aqe-mcp-output.txt | wc -l)
|
|
if [ "$TOOL_COUNT" -ge 80 ]; then
|
|
check_result "MCP tools available (≥80)" 0
|
|
else
|
|
check_result "MCP tools available (got $TOOL_COUNT, expected ≥80)" 1
|
|
fi
|
|
else
|
|
check_result "MCP server starts" 1
|
|
fi
|
|
|
|
# ============================================================================
|
|
# Step 5: Test learning capture
|
|
# ============================================================================
|
|
echo -e "${YELLOW}[5/6] Testing learning capture...${NC}"
|
|
|
|
# Simulate a Task completion with PostToolUse hook input
|
|
HOOK_INPUT='{"tool_name":"Task","tool_input":{"prompt":"Generate unit tests","subagent_type":"qe-test-generator","description":"Test gen"},"tool_response":{"status":"completed","content":[{"text":"Generated 5 unit tests successfully:\n1. test_create\n2. test_update\n3. test_delete\n4. test_read\n5. test_validate"}],"totalDurationMs":2500,"totalTokens":1500,"totalToolUseCount":3,"agentId":"test-gen-001"}}'
|
|
|
|
# Run the hook with NODE_PATH set to find better-sqlite3
|
|
HOOK_OUTPUT=$(echo "$HOOK_INPUT" | NODE_PATH="$PROJECT_ROOT/node_modules" node scripts/hooks/capture-task-learning.js 2>&1)
|
|
if echo "$HOOK_OUTPUT" | grep -q "Learning captured"; then
|
|
check_result "Hook captures learning" 0
|
|
else
|
|
check_result "Hook captures learning (output: $HOOK_OUTPUT)" 1
|
|
fi
|
|
|
|
# Verify learning status shows data
|
|
LEARN_OUTPUT=$(node "$PROJECT_ROOT/dist/cli/index.js" learn status 2>&1)
|
|
if echo "$LEARN_OUTPUT" | grep -q "Total Experiences: 1"; then
|
|
check_result "Learning status shows experience" 0
|
|
else
|
|
check_result "Learning status shows experience" 1
|
|
fi
|
|
|
|
# ============================================================================
|
|
# Step 6: Verify settings.json hooks configuration
|
|
# ============================================================================
|
|
echo -e "${YELLOW}[6/6] Verifying hooks configuration...${NC}"
|
|
|
|
# Check PostToolUse hook for Task
|
|
if grep -q '"matcher": "Task"' .claude/settings.json && grep -q 'capture-task-learning.js' .claude/settings.json; then
|
|
check_result "PostToolUse hook configured for Task" 0
|
|
else
|
|
check_result "PostToolUse hook configured for Task" 1
|
|
fi
|
|
|
|
# Check PreToolUse hook for memory patterns
|
|
if grep -q '"matcher": "Write|Edit|MultiEdit"' .claude/settings.json; then
|
|
check_result "PreToolUse hook configured for edits" 0
|
|
else
|
|
check_result "PreToolUse hook configured for edits" 1
|
|
fi
|
|
|
|
# ============================================================================
|
|
# Summary
|
|
# ============================================================================
|
|
echo ""
|
|
echo -e "${BLUE}═══════════════════════════════════════════════════════════════${NC}"
|
|
if [ "$FAILED" -eq 0 ]; then
|
|
echo -e "${GREEN} ✓ All release verification checks passed!${NC}"
|
|
echo -e "${GREEN} Version $VERSION is ready for release.${NC}"
|
|
else
|
|
echo -e "${RED} ✗ Some checks failed. Please fix before releasing.${NC}"
|
|
fi
|
|
echo -e "${BLUE}═══════════════════════════════════════════════════════════════${NC}"
|
|
echo ""
|
|
|
|
exit $FAILED
|