File size: 7,406 Bytes
d3cadd5 0edbd7b d3cadd5 0edbd7b d3cadd5 0edbd7b |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 |
#!/usr/bin/env python3
"""Manual demo for the smart model-mapping helpers."""
from __future__ import annotations
import sys
from pathlib import Path
sys.path.insert(0, str(Path(__file__).resolve().parent.parent))
from kiro_proxy.config import map_model_name, detect_model_tier, get_best_model_by_tier
def test_tier_detection():
"""测试等级检测功能"""
print("测试模型等级检测:")
test_cases = [
# Opus 等级 (最强)
("claude-4-opus", "opus"),
("gpt-o1-preview", "opus"),
("gemini-1.5-pro", "opus"),
("claude-3-opus-20240229", "opus"),
("some-premium-model", "opus"),
# Sonnet 等级 (平衡)
("claude-3.5-sonnet", "sonnet"),
("gpt-4o", "sonnet"),
("gemini-2.0-flash", "sonnet"),
("claude-4-standard", "sonnet"),
# Haiku 等级 (快速)
("claude-3-haiku", "haiku"),
("gpt-4o-mini", "haiku"),
("gpt-3.5-turbo", "haiku"),
("claude-haiku-fast", "haiku"),
# 未知模型
("unknown-model-xyz", "sonnet"), # 默认中等
("", "sonnet"), # 空值默认
]
for model, expected in test_cases:
result = detect_model_tier(model)
status = "OK" if result == expected else "FAIL"
print(f" {status} {model:<25} -> {result:<6} (期望: {expected})")
def test_dynamic_mapping():
"""测试动态模型映射（等级对等 + 智能降级）"""
print("\n测试动态模型映射（等级对等策略）:")
# 模拟不同的可用模型场景
scenarios = [
{
"name": "全部可用",
"available": {"claude-sonnet-4.5", "claude-sonnet-4", "claude-haiku-4.5", "auto"}
},
{
"name": "缺少4.5版本",
"available": {"claude-sonnet-4", "claude-haiku-4.5", "auto"}
},
{
"name": "仅有Haiku",
"available": {"claude-haiku-4.5", "auto"}
},
{
"name": "仅有Sonnet-4",
"available": {"claude-sonnet-4", "auto"}
}
]
test_models = [
("claude-4-opus", "opus"), # 应该优先选择 sonnet-4.5
("gpt-4o", "sonnet"), # 应该优先选择 sonnet-4.5
("gpt-4o-mini", "haiku"), # 应该优先选择 haiku-4.5
("unknown-future-model", "sonnet") # 未知模型，默认 sonnet-4.5
]
for scenario in scenarios:
print(f"\n 场景: {scenario['name']}")
print(f" 可用模型: {scenario['available']}")
for model, expected_tier in test_models:
result = map_model_name(model, scenario['available'])
tier = detect_model_tier(model)
print(f" {model:<20} ({tier:<6}) -> {result}")
def test_tier_mapping_logic():
"""测试等级对等映射逻辑"""
print("\n测试等级对等映射逻辑:")
# 全部可用时的期望映射
full_available = {"claude-sonnet-4.5", "claude-sonnet-4", "claude-haiku-4.5", "auto"}
test_cases = [
# 格式: (输入模型, 期望等级, 期望输出模型)
("claude-4-opus", "opus", "claude-sonnet-4.5"), # Opus -> 最强
("gpt-4o", "sonnet", "claude-sonnet-4.5"), # Sonnet -> 高性能
("gpt-4o-mini", "haiku", "claude-haiku-4.5"), # Haiku -> 快速
("o1-preview", "opus", "claude-sonnet-4.5"), # O1 -> 最强
("claude-3.5-sonnet", "sonnet", "claude-sonnet-4.5"), # Sonnet -> 高性能
("gpt-3.5-turbo", "haiku", "claude-haiku-4.5"), # 3.5 -> 快速
]
for model, expected_tier, expected_output in test_cases:
tier = detect_model_tier(model)
result = map_model_name(model, full_available)
tier_ok = "OK" if tier == expected_tier else "FAIL"
output_ok = "OK" if result == expected_output else "FAIL"
print(f" {tier_ok}/{output_ok} {model:<20} -> {tier:<6} -> {result}")
if tier != expected_tier:
print(f" 等级检测错误: 期望 {expected_tier}, 实际 {tier}")
if result != expected_output:
print(f" 映射错误: 期望 {expected_output}, 实际 {result}")
def test_degradation_paths():
"""测试降级路径"""
print("\n测试降级路径:")
degradation_scenarios = [
{
"name": "Opus降级测试",
"model": "claude-4-opus",
"scenarios": [
({"claude-sonnet-4.5", "auto"}, "claude-sonnet-4.5"), # 首选可用
({"claude-sonnet-4", "auto"}, "claude-sonnet-4"), # 降级到次强
({"claude-haiku-4.5", "auto"}, "claude-haiku-4.5"), # 降级到快速
({"auto"}, "auto"), # 最终回退
]
},
{
"name": "Haiku降级测试",
"model": "gpt-4o-mini",
"scenarios": [
({"claude-haiku-4.5", "auto"}, "claude-haiku-4.5"), # 首选可用
({"claude-sonnet-4", "auto"}, "claude-sonnet-4"), # 降级到标准
({"claude-sonnet-4.5", "auto"}, "claude-sonnet-4.5"), # 降级到高性能
({"auto"}, "auto"), # 最终回退
]
}
]
for test_group in degradation_scenarios:
print(f"\n {test_group['name']}:")
model = test_group['model']
tier = detect_model_tier(model)
for available, expected in test_group['scenarios']:
result = map_model_name(model, available)
status = "OK" if result == expected else "FAIL"
print(f" {status} 可用:{available} -> {result} (期望:{expected})")
def test_backward_compatibility():
"""测试向后兼容性"""
print("\n测试向后兼容性:")
# 原有的精确映射应该仍然工作
legacy_tests = [
("gpt-4o", "claude-sonnet-4"),
("claude-3-5-sonnet-20241022", "claude-sonnet-4"),
("o1-preview", "claude-sonnet-4.5"),
("gemini-1.5-pro", "claude-sonnet-4.5"),
]
for model, expected in legacy_tests:
result = map_model_name(model)
status = "OK" if result == expected else "FAIL"
print(f" {status} {model:<25} -> {result:<20} (期望: {expected})")
def test_edge_cases():
"""测试边界情况"""
print("\n测试边界情况:")
edge_cases = [
("", "auto"), # 空字符串
(None, "auto"), # None值 (需要修改函数处理)
("CLAUDE-4-OPUS", "claude-sonnet-4.5"), # 大写
("gpt-4o-MINI-turbo", "claude-haiku-4.5"), # 混合大小写
("claude_sonnet_4", "claude-sonnet-4"), # 下划线
]
for model, expected in edge_cases:
try:
result = map_model_name(model or "")
tier = detect_model_tier(model or "")
status = "OK" if result == expected else "FAIL"
print(f" {status} {str(model):<25} ({tier}) -> {result}")
except Exception as e:
print(f" ERROR {str(model):<25} -> 错误: {e}")
if __name__ == "__main__":
print("KiroProxy 智能模型映射测试（等级对等策略）\n")
test_tier_detection()
test_tier_mapping_logic()
test_degradation_paths()
test_dynamic_mapping()
test_backward_compatibility()
test_edge_cases()
print("\n测试完成!")
|