|
|
|
@ -24,30 +24,30 @@ class MemorizingLazyConstrComponent:
|
|
|
|
|
def __init__(self, clf: Any, extractor: FeaturesExtractor) -> None:
|
|
|
|
|
self.clf = clf
|
|
|
|
|
self.extractor = extractor
|
|
|
|
|
self.violations_: List[Hashable] = []
|
|
|
|
|
self.constrs_: List[Hashable] = []
|
|
|
|
|
self.n_features_: int = 0
|
|
|
|
|
self.n_targets_: int = 0
|
|
|
|
|
|
|
|
|
|
def fit(self, train_h5: List[str]) -> None:
|
|
|
|
|
logger.info("Reading training data...")
|
|
|
|
|
n_samples = len(train_h5)
|
|
|
|
|
x, y, violations, n_features = [], [], [], None
|
|
|
|
|
violation_to_idx: Dict[Hashable, int] = {}
|
|
|
|
|
x, y, constrs, n_features = [], [], [], None
|
|
|
|
|
constr_to_idx: Dict[Hashable, int] = {}
|
|
|
|
|
for h5_filename in train_h5:
|
|
|
|
|
with H5File(h5_filename, "r") as h5:
|
|
|
|
|
|
|
|
|
|
# Store lazy constraints
|
|
|
|
|
sample_violations_str = h5.get_scalar("mip_constr_violations")
|
|
|
|
|
assert sample_violations_str is not None
|
|
|
|
|
assert isinstance(sample_violations_str, str)
|
|
|
|
|
sample_violations = eval(sample_violations_str)
|
|
|
|
|
assert isinstance(sample_violations, list)
|
|
|
|
|
sample_constrs_str = h5.get_scalar("mip_lazy")
|
|
|
|
|
assert sample_constrs_str is not None
|
|
|
|
|
assert isinstance(sample_constrs_str, str)
|
|
|
|
|
sample_constrs = eval(sample_constrs_str)
|
|
|
|
|
assert isinstance(sample_constrs, list)
|
|
|
|
|
y_sample = []
|
|
|
|
|
for v in sample_violations:
|
|
|
|
|
if v not in violation_to_idx:
|
|
|
|
|
violation_to_idx[v] = len(violation_to_idx)
|
|
|
|
|
violations.append(v)
|
|
|
|
|
y_sample.append(violation_to_idx[v])
|
|
|
|
|
for c in sample_constrs:
|
|
|
|
|
if c not in constr_to_idx:
|
|
|
|
|
constr_to_idx[c] = len(constr_to_idx)
|
|
|
|
|
constrs.append(c)
|
|
|
|
|
y_sample.append(constr_to_idx[c])
|
|
|
|
|
y.append(y_sample)
|
|
|
|
|
|
|
|
|
|
# Extract features
|
|
|
|
@ -62,8 +62,8 @@ class MemorizingLazyConstrComponent:
|
|
|
|
|
logger.info("Constructing matrices...")
|
|
|
|
|
assert n_features is not None
|
|
|
|
|
self.n_features_ = n_features
|
|
|
|
|
self.violations_ = violations
|
|
|
|
|
self.n_targets_ = len(violation_to_idx)
|
|
|
|
|
self.constrs_ = constrs
|
|
|
|
|
self.n_targets_ = len(constr_to_idx)
|
|
|
|
|
x_np = np.vstack(x)
|
|
|
|
|
assert x_np.shape == (n_samples, n_features)
|
|
|
|
|
y_np = MultiLabelBinarizer().fit_transform(y)
|
|
|
|
@ -82,8 +82,8 @@ class MemorizingLazyConstrComponent:
|
|
|
|
|
model: GurobiModel,
|
|
|
|
|
stats: Dict[str, Any],
|
|
|
|
|
) -> None:
|
|
|
|
|
assert self.violations_ is not None
|
|
|
|
|
if model.fix_violations is None:
|
|
|
|
|
assert self.constrs_ is not None
|
|
|
|
|
if model.lazy_enforce is None:
|
|
|
|
|
return
|
|
|
|
|
|
|
|
|
|
# Read features
|
|
|
|
@ -99,7 +99,7 @@ class MemorizingLazyConstrComponent:
|
|
|
|
|
y = y.reshape(-1)
|
|
|
|
|
|
|
|
|
|
# Enforce constraints
|
|
|
|
|
violations = [self.violations_[i] for (i, yi) in enumerate(y) if yi > 0.5]
|
|
|
|
|
violations = [self.constrs_[i] for (i, yi) in enumerate(y) if yi > 0.5]
|
|
|
|
|
logger.info(f"Enforcing {len(violations)} constraints ahead-of-time...")
|
|
|
|
|
model.fix_violations(model, violations, "aot")
|
|
|
|
|
model.lazy_enforce(model, violations, "aot")
|
|
|
|
|
stats["Lazy Constraints: AOT"] = len(violations)
|
|
|
|
|