feat: add migrations, test scripts, and utility tools
- Add database migrations (006-028) for face recognition, identity, file_uuid - Add test scripts for ASR, face, search, processing - Add portal frontend (Tauri) - Add config, benchmark, and monitoring utilities - Add model checkpoints and pretrained model references
This commit is contained in:
Warren
@@ -0,0 +1,248 @@
|
||||
#!/opt/homebrew/bin/python3.11
|
||||
"""
|
||||
微基准测试 - 测试合约合规处理器的初始化开销
|
||||
Micro Benchmark - Test initialization overhead of contract-compliant processors
|
||||
"""
|
||||
|
||||
import sys
|
||||
import json
|
||||
import os
|
||||
import time
|
||||
import subprocess
|
||||
import statistics
|
||||
from datetime import datetime
|
||||
from typing import Dict, List, Any
|
||||
|
||||
# Test configuration
|
||||
NUM_RUNS = 10 # More runs for statistical significance
|
||||
|
||||
# Processors to test
|
||||
PROCESSORS = {
|
||||
"asr": {
|
||||
"legacy": "scripts/asr_processor.py",
|
||||
"contract": "scripts/asr_processor_contract_v2.py",
|
||||
},
|
||||
"ocr": {
|
||||
"legacy": "scripts/ocr_processor.py",
|
||||
"contract": "scripts/ocr_processor_contract_v1.py",
|
||||
},
|
||||
}
|
||||
|
||||
|
||||
def measure_import_time(script_path: str) -> float:
|
||||
"""测量处理器导入时间"""
|
||||
|
||||
test_code = f"""
|
||||
import sys
|
||||
import time
|
||||
|
||||
start_time = time.time()
|
||||
try:
|
||||
# Import the module
|
||||
sys.path.insert(0, 'scripts')
|
||||
import {os.path.basename(script_path).replace(".py", "")} as processor
|
||||
elapsed = time.time() - start_time
|
||||
print(f"IMPORT_TIME:{{elapsed}}")
|
||||
except Exception as e:
|
||||
print(f"IMPORT_ERROR:{{e}}")
|
||||
sys.exit(1)
|
||||
"""
|
||||
|
||||
try:
|
||||
result = subprocess.run(
|
||||
[sys.executable, "-c", test_code],
|
||||
capture_output=True,
|
||||
text=True,
|
||||
timeout=10,
|
||||
cwd=os.path.dirname(os.path.abspath(__file__)),
|
||||
)
|
||||
|
||||
for line in result.stdout.split("\n"):
|
||||
if line.startswith("IMPORT_TIME:"):
|
||||
return float(line.split(":")[1])
|
||||
|
||||
return float("inf") # Failed to measure
|
||||
|
||||
except Exception as e:
|
||||
print(f"测量导入时间失败: {e}")
|
||||
return float("inf")
|
||||
|
||||
|
||||
def measure_health_check_time(script_path: str) -> float:
|
||||
"""测量健康检查执行时间"""
|
||||
|
||||
cmd = [sys.executable, script_path, "--check-health", "dummy.mp4", "dummy.json"]
|
||||
|
||||
try:
|
||||
start_time = time.time()
|
||||
result = subprocess.run(
|
||||
cmd,
|
||||
capture_output=True,
|
||||
text=True,
|
||||
timeout=30,
|
||||
cwd=os.path.dirname(os.path.abspath(__file__)),
|
||||
)
|
||||
elapsed = time.time() - start_time
|
||||
|
||||
if result.returncode == 0:
|
||||
return elapsed
|
||||
else:
|
||||
return float("inf")
|
||||
|
||||
except Exception as e:
|
||||
print(f"测量健康检查时间失败: {e}")
|
||||
return float("inf")
|
||||
|
||||
|
||||
def run_micro_benchmark():
|
||||
"""运行微基准测试"""
|
||||
|
||||
print("=" * 80)
|
||||
print("微基准测试 - 合约合规处理器开销分析")
|
||||
print(f"时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
|
||||
print("=" * 80)
|
||||
print()
|
||||
|
||||
results = {}
|
||||
|
||||
# Test each processor
|
||||
for processor_type in PROCESSORS:
|
||||
print(f"\n测试 {processor_type.upper()} 处理器...")
|
||||
print("-" * 40)
|
||||
|
||||
processor_results = {
|
||||
"legacy": {"import_times": [], "health_check_times": [], "summary": {}},
|
||||
"contract": {"import_times": [], "health_check_times": [], "summary": {}},
|
||||
}
|
||||
|
||||
# Test both versions
|
||||
for version in ["legacy", "contract"]:
|
||||
print(f"\n版本: {version}")
|
||||
|
||||
script_path = PROCESSORS[processor_type][version]
|
||||
|
||||
# Measure import time (multiple runs)
|
||||
print(" 测量导入时间...")
|
||||
import_times = []
|
||||
for run in range(NUM_RUNS):
|
||||
import_time = measure_import_time(script_path)
|
||||
if import_time < float("inf"):
|
||||
import_times.append(import_time)
|
||||
print(f" 运行 #{run}: {import_time * 1000:.1f} ms")
|
||||
else:
|
||||
print(f" 运行 #{run}: 失败")
|
||||
|
||||
# Measure health check time (multiple runs)
|
||||
print(" 测量健康检查时间...")
|
||||
health_check_times = []
|
||||
for run in range(NUM_RUNS):
|
||||
health_check_time = measure_health_check_time(script_path)
|
||||
if health_check_time < float("inf"):
|
||||
health_check_times.append(health_check_time)
|
||||
print(f" 运行 #{run}: {health_check_time * 1000:.1f} ms")
|
||||
else:
|
||||
print(f" 运行 #{run}: 失败")
|
||||
|
||||
# Store results
|
||||
processor_results[version]["import_times"] = import_times
|
||||
processor_results[version]["health_check_times"] = health_check_times
|
||||
|
||||
# Calculate statistics
|
||||
if import_times:
|
||||
processor_results[version]["summary"]["import"] = {
|
||||
"runs": len(import_times),
|
||||
"min_ms": min(import_times) * 1000,
|
||||
"max_ms": max(import_times) * 1000,
|
||||
"avg_ms": statistics.mean(import_times) * 1000,
|
||||
"median_ms": statistics.median(import_times) * 1000,
|
||||
"std_dev_ms": statistics.stdev(import_times) * 1000
|
||||
if len(import_times) > 1
|
||||
else 0,
|
||||
}
|
||||
|
||||
if health_check_times:
|
||||
processor_results[version]["summary"]["health_check"] = {
|
||||
"runs": len(health_check_times),
|
||||
"min_ms": min(health_check_times) * 1000,
|
||||
"max_ms": max(health_check_times) * 1000,
|
||||
"avg_ms": statistics.mean(health_check_times) * 1000,
|
||||
"median_ms": statistics.median(health_check_times) * 1000,
|
||||
"std_dev_ms": statistics.stdev(health_check_times) * 1000
|
||||
if len(health_check_times) > 1
|
||||
else 0,
|
||||
}
|
||||
|
||||
results[processor_type] = processor_results
|
||||
|
||||
# Calculate overhead
|
||||
if processor_results["legacy"]["summary"].get("import") and processor_results[
|
||||
"contract"
|
||||
]["summary"].get("import"):
|
||||
legacy_import = processor_results["legacy"]["summary"]["import"]["avg_ms"]
|
||||
contract_import = processor_results["contract"]["summary"]["import"][
|
||||
"avg_ms"
|
||||
]
|
||||
|
||||
import_overhead = ((contract_import - legacy_import) / legacy_import) * 100
|
||||
|
||||
print(f"\n导入开销分析:")
|
||||
print(f" 传统版本: {legacy_import:.1f} ms")
|
||||
print(f" 合约版本: {contract_import:.1f} ms")
|
||||
print(f" 开销: {import_overhead:.1f}%")
|
||||
|
||||
if import_overhead <= 5:
|
||||
print(f" ✅ 通过: 导入开销 ≤ 5%")
|
||||
else:
|
||||
print(f" ❌ 失败: 导入开销 > 5%")
|
||||
|
||||
if processor_results["legacy"]["summary"].get(
|
||||
"health_check"
|
||||
) and processor_results["contract"]["summary"].get("health_check"):
|
||||
legacy_hc = processor_results["legacy"]["summary"]["health_check"]["avg_ms"]
|
||||
contract_hc = processor_results["contract"]["summary"]["health_check"][
|
||||
"avg_ms"
|
||||
]
|
||||
|
||||
hc_overhead = ((contract_hc - legacy_hc) / legacy_hc) * 100
|
||||
|
||||
print(f"\n健康检查开销分析:")
|
||||
print(f" 传统版本: {legacy_hc:.1f} ms")
|
||||
print(f" 合约版本: {contract_hc:.1f} ms")
|
||||
print(f" 开销: {hc_overhead:.1f}%")
|
||||
|
||||
if hc_overhead <= 5:
|
||||
print(f" ✅ 通过: 健康检查开销 ≤ 5%")
|
||||
else:
|
||||
print(f" ❌ 失败: 健康检查开销 > 5%")
|
||||
|
||||
# Generate final report
|
||||
print("\n" + "=" * 80)
|
||||
print("微基准测试完成报告")
|
||||
print("=" * 80)
|
||||
|
||||
# Save detailed results
|
||||
report_file = f"/tmp/micro_benchmark_report_{int(time.time())}.json"
|
||||
with open(report_file, "w") as f:
|
||||
json.dump(
|
||||
{
|
||||
"timestamp": datetime.now().isoformat(),
|
||||
"test_config": {
|
||||
"num_runs": NUM_RUNS,
|
||||
"processors_tested": list(PROCESSORS.keys()),
|
||||
},
|
||||
"results": results,
|
||||
},
|
||||
f,
|
||||
indent=2,
|
||||
ensure_ascii=False,
|
||||
)
|
||||
|
||||
print(f"\n详细报告保存到: {report_file}")
|
||||
print("=" * 80)
|
||||
|
||||
return True
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
success = run_micro_benchmark()
|
||||
sys.exit(0 if success else 1)
|
||||
Reference in New Issue
Block a user