def test_conditional_space(self):
# 测试conditional
cs = HDL2CS()({
"model(choice)": {
"linearsvc": {
"max_iter": {"_type": "int_quniform", "_value": [300, 3000, 100], "_default": 600},
"penalty": {"_type": "choice", "_value": ["l1", "l2"], "_default": "l2"},
"dual": {"_type": "choice", "_value": [True, False], "_default": False},
"loss": {"_type": "choice", "_value": ["hinge", "squared_hinge"], "_default": "squared_hinge"},
"C": {"_type": "loguniform", "_value": [0.01, 10000], "_default": 1.0},
"multi_class": "ovr",
"random_state": 42,
"__forbidden": [
{"penalty": "l1", "loss": "hinge"},
{"penalty": "l2", "dual": False, "loss": "hinge"},
{"penalty": "l1", "dual": False},
{"penalty": "l1", "dual": True, "loss": "squared_hinge"},
]
},
"svc": {
"C": {"_type": "loguniform", "_value": [0.01, 10000], "_default": 1.0},
"kernel": {"_type": "choice", "_value": ["rbf", "poly", "sigmoid"], "_default": "rbf"},
"degree": {"_type": "int_uniform", "_value": [2, 5], "_default": 3},
"gamma": {"_type": "loguniform", "_value": [1e-05, 8], "_default": 0.1},
"coef0": {"_type": "quniform", "_value": [-1, 1], "_default": 0},
"shrinking": {"_type": "choice", "_value": [True, False], "_default": True},
"class_weight": None,
"probability": True,
"decision_function_shape": "ovr",
"__activate": {
"kernel": {
"rbf": ["gamma"],
"sigmoid": ["gamma", "coef0"],
"poly": ["degree", "gamma", "coef0"]
}
},
"random_state": 42
},
}
})
print(cs)
configs=cs.sample_configuration(200)
res=[layering_config(config) for config in configs]
print(len(res))
print("OK")
def test_multi_choice(self):
cs = HDL2CS()({
"A(choice)": {
"A1(choice)": {
"B1": {
"B1_h": {
"_type": "ordinal",
"_value": ["1", "2", "3"]
},
},
"B2": {
"B2_h": {
"_type": "choice",
"_value": ["1", "2", "3"]
},
},
},
"A2": {
"A2_h1": {
"_type": "choice",
"_value": ["1", "2", "3"]
},
"A2_h2": {
"_type": "uniform",
"_value": [-999, 666]
}
},
"A3": {
"A3_h1": {
"_type": "choice",
"_value": ["1", "2", "3"]
},
"A3_h2": {
"_type": "uniform",
"_value": [-999, 666]
}
},
}
})
print(cs)
def test(self):
cs = HDL2CS().recursion({
"A(choice)": {
"A1": {
"A1_h1": {
"_type": "choice",
"_value": ["1", "2", "3"]
},
"A1_h2": {
"_type": "uniform",
"_value": [-999, 666]
}
},
"A2": {
"A2_h1": {
"_type": "choice",
"_value": ["1", "2", "3"]
},
"A2_h2": {
"_type": "uniform",
"_value": [-999, 666]
}
},
"A3": {
"A3_h1": {
"_type": "choice",
"_value": ["1", "2", "3"]
},
"A3_h2": {
"_type": "uniform",
"_value": [-999, 666]
}
},
}
})
configs = initial_design_2(cs, 1, 1)
arr = get_array_from_configs(configs)
print(pd.Series(arr[:, 0]).nunique() == 3)
# experiment = "GB1"
experiment = "PhoQ"
# experiment = "RNA"
if experiment == "GB1":
df = pd.read_csv("GB1.csv")
df2 = df.copy()
df2.index = df2.pop('Variants')
y = df.pop("Fitness")
for i in range(4):
df[f'X{i}'] = df['Variants'].str[i]
df.pop('Variants')
choices = sorted(Counter(df['X0']).keys())
hdl = {f"X{i}": {"_type": "choice", "_value": choices} for i in range(4)}
config_space = HDL2CS().recursion(hdl)
def evaluation(config: Configuration):
return -df2.loc["".join([config.get(f"X{i}")
for i in range(4)]), 'Fitness']
elif experiment == "PhoQ":
df = pd.read_csv("PhoQ.csv")
df2 = df.copy()
df2.index = df2.pop('Variants')
y = df.pop("Fitness")
for i in range(4):
df[f'X{i}'] = df['Variants'].str[i]
df.pop('Variants')
choices = sorted(Counter(df['X0']).keys())
hdl = {f"X{i}": {"_type": "choice", "_value": choices} for i in range(4)}