379 lines
13 KiB
C++
379 lines
13 KiB
C++
// Copyright 2010-2025 Google LLC
|
|
// Licensed under the Apache License, Version 2.0 (the "License");
|
|
// you may not use this file except in compliance with the License.
|
|
// You may obtain a copy of the License at
|
|
//
|
|
// http://www.apache.org/licenses/LICENSE-2.0
|
|
//
|
|
// Unless required by applicable law or agreed to in writing, software
|
|
// distributed under the License is distributed on an "AS IS" BASIS,
|
|
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
// See the License for the specific language governing permissions and
|
|
// limitations under the License.
|
|
|
|
#ifndef OR_TOOLS_SAT_OPB_READER_H_
|
|
#define OR_TOOLS_SAT_OPB_READER_H_
|
|
|
|
#include <algorithm>
|
|
#include <cstdint>
|
|
#include <limits>
|
|
#include <string>
|
|
#include <utility>
|
|
#include <vector>
|
|
|
|
#include "absl/base/attributes.h"
|
|
#include "absl/container/flat_hash_map.h"
|
|
#include "absl/container/flat_hash_set.h"
|
|
#include "absl/log/check.h"
|
|
#include "absl/log/log.h"
|
|
#include "absl/strings/numbers.h"
|
|
#include "absl/strings/str_split.h"
|
|
#include "absl/types/span.h"
|
|
#include "ortools/base/logging.h"
|
|
#include "ortools/base/stl_util.h"
|
|
#include "ortools/sat/cp_model.pb.h"
|
|
#include "ortools/sat/cp_model_utils.h"
|
|
#include "ortools/util/filelineiter.h"
|
|
|
|
namespace operations_research {
|
|
namespace sat {
|
|
|
|
// This class loads a file in pbo file format into a LinearBooleanProblem.
|
|
// The format is described here:
|
|
// http://www.cril.univ-artois.fr/PB24/format.pdf
|
|
class OpbReader {
|
|
public:
|
|
OpbReader() = default;
|
|
// This type is neither copyable nor movable.
|
|
OpbReader(const OpbReader&) = delete;
|
|
OpbReader& operator=(const OpbReader&) = delete;
|
|
|
|
// Returns the number of variables in the problem.
|
|
int num_variables() const { return num_variables_; }
|
|
|
|
// Returns true if the model is supported. A model is not supported if it
|
|
// contains an integer that does not fit in int64_t.
|
|
bool model_is_supported() const { return model_is_supported_; }
|
|
|
|
// Loads the given opb filename into the given problem.
|
|
// Returns true on success.
|
|
ABSL_MUST_USE_RESULT bool LoadAndValidate(const std::string& filename,
|
|
CpModelProto* model) {
|
|
model->Clear();
|
|
model->set_name(ExtractProblemName(filename));
|
|
|
|
num_variables_ = 0;
|
|
int num_lines = 0;
|
|
model_is_supported_ = true;
|
|
|
|
// Read constraints line by line (1 constraint per line).
|
|
// We process into a temporary structure to support non linear constraints
|
|
// and weighted constraints.
|
|
for (const std::string& line : FileLines(filename)) {
|
|
++num_lines;
|
|
ProcessNewLine(line);
|
|
|
|
// Check if the model is supported. It is not supported if one constant
|
|
// contains an integer that does not fit in an int64_t.
|
|
if (!model_is_supported_) return false;
|
|
}
|
|
if (num_lines == 0) {
|
|
LOG(ERROR) << "File '" << filename << "' is empty or can't be read.";
|
|
return false;
|
|
}
|
|
|
|
LOG(INFO) << "Read " << num_lines << " lines from " << filename;
|
|
LOG(INFO) << "#variables: " << num_variables_;
|
|
LOG(INFO) << "#constraints: " << constraints_.size();
|
|
LOG(INFO) << "#objective: " << objective_.size();
|
|
|
|
const std::string error_message = ValidateModel();
|
|
if (!error_message.empty()) {
|
|
LOG(ERROR) << "Error while trying to parse '" << filename
|
|
<< "': " << error_message;
|
|
return false;
|
|
}
|
|
|
|
BuildModel(model);
|
|
return true;
|
|
}
|
|
|
|
private:
|
|
// A term is coeff * Product(literals).
|
|
// Note that it is okay to have duplicate literals here, we will just merge
|
|
// them. Having a literal and its negation will always result in a product of
|
|
// zero.
|
|
struct PbTerm {
|
|
int64_t coeff;
|
|
std::vector<int> literals; // CpModelProto literals
|
|
};
|
|
|
|
enum PbConstraintType {
|
|
UNDEFINED_OPERATION,
|
|
GE_OPERATION,
|
|
EQ_OPERATION,
|
|
};
|
|
|
|
struct PbConstraint {
|
|
std::vector<PbTerm> terms;
|
|
PbConstraintType type = UNDEFINED_OPERATION;
|
|
int64_t rhs = std::numeric_limits<int64_t>::min();
|
|
int64_t soft_cost = std::numeric_limits<int64_t>::max();
|
|
};
|
|
|
|
// Since the problem name is not stored in the opb format, we infer it from
|
|
// the file name.
|
|
static std::string ExtractProblemName(const std::string& filename) {
|
|
const int found = filename.find_last_of('/');
|
|
const std::string problem_name =
|
|
found != std::string::npos ? filename.substr(found + 1) : filename;
|
|
return problem_name;
|
|
}
|
|
|
|
void ProcessNewLine(const std::string& line) {
|
|
const std::vector<std::string> words =
|
|
absl::StrSplit(line, absl::ByAnyChar(" ;"), absl::SkipEmpty());
|
|
if (words.empty() || words[0].empty() || words[0][0] == '*') {
|
|
// TODO(user): Parse comments.
|
|
return;
|
|
}
|
|
|
|
// We ignore the number of soft constraints.
|
|
if (words[0] == "soft:") return;
|
|
|
|
if (words[0] == "min:") {
|
|
for (int i = 1; i < words.size(); ++i) {
|
|
const std::string& word = words[i];
|
|
if (word.empty() || word[0] == ';') continue;
|
|
if (word[0] == 'x') {
|
|
const int index = ParseIndex(word.substr(1));
|
|
num_variables_ = std::max(num_variables_, index);
|
|
objective_.back().literals.push_back(
|
|
PbLiteralToCpModelLiteral(index));
|
|
} else if (word[0] == '~' && word[1] == 'x') {
|
|
const int index = ParseIndex(word.substr(2));
|
|
num_variables_ = std::max(num_variables_, index);
|
|
objective_.back().literals.push_back(
|
|
NegatedRef(PbLiteralToCpModelLiteral(index)));
|
|
} else {
|
|
// Note that coefficient always appear before the variable/variables.
|
|
PbTerm term;
|
|
if (!ParseInt64Into(word, &term.coeff)) return;
|
|
objective_.emplace_back(std::move(term));
|
|
}
|
|
}
|
|
|
|
// Normalize objective literals.
|
|
for (PbTerm& term : objective_) {
|
|
if (term.literals.size() <= 1) continue;
|
|
gtl::STLSortAndRemoveDuplicates(&term.literals);
|
|
CHECK_GT(term.literals.size(), 1);
|
|
}
|
|
|
|
return;
|
|
}
|
|
|
|
PbConstraint constraint;
|
|
for (int i = 0; i < words.size(); ++i) {
|
|
const std::string& word = words[i];
|
|
CHECK(!word.empty());
|
|
if (word[0] == '[') { // Soft constraint.
|
|
if (!ParseInt64Into(word.substr(1, word.size() - 2),
|
|
&constraint.soft_cost)) {
|
|
return;
|
|
}
|
|
} else if (word == ">=") {
|
|
CHECK_LT(i + 1, words.size());
|
|
constraint.type = GE_OPERATION;
|
|
if (!ParseInt64Into(words[i + 1], &constraint.rhs)) return;
|
|
break;
|
|
} else if (word == "=") {
|
|
CHECK_LT(i + 1, words.size());
|
|
constraint.type = EQ_OPERATION;
|
|
if (!ParseInt64Into(words[i + 1], &constraint.rhs)) return;
|
|
break;
|
|
} else if (word[0] == 'x') {
|
|
const int index = ParseIndex(word.substr(1));
|
|
num_variables_ = std::max(num_variables_, index);
|
|
constraint.terms.back().literals.push_back(
|
|
PbLiteralToCpModelLiteral(index));
|
|
} else if (word[0] == '~' && word[1] == 'x') {
|
|
const int index = ParseIndex(word.substr(2));
|
|
num_variables_ = std::max(num_variables_, index);
|
|
constraint.terms.back().literals.push_back(
|
|
NegatedRef(PbLiteralToCpModelLiteral(index)));
|
|
} else {
|
|
// Note that coefficient always appear before the variable/variables.
|
|
PbTerm term;
|
|
if (!ParseInt64Into(word, &term.coeff)) return;
|
|
constraint.terms.emplace_back(std::move(term));
|
|
}
|
|
}
|
|
|
|
// Normalize literals.
|
|
for (PbTerm& term : constraint.terms) {
|
|
if (term.literals.size() <= 1) continue;
|
|
gtl::STLSortAndRemoveDuplicates(&term.literals);
|
|
CHECK_GT(term.literals.size(), 1);
|
|
}
|
|
|
|
constraints_.push_back(std::move(constraint));
|
|
}
|
|
|
|
std::string ValidateModel() {
|
|
// Normalize and validate constraints.
|
|
for (const PbConstraint& constraint : constraints_) {
|
|
if (constraint.rhs == std::numeric_limits<int64_t>::min()) {
|
|
return "constraint error: undefined rhs";
|
|
}
|
|
|
|
if (constraint.type == UNDEFINED_OPERATION) {
|
|
return "constraint error: undefined operation";
|
|
}
|
|
|
|
for (const PbTerm& term : constraint.terms) {
|
|
if (term.coeff == 0) {
|
|
return "constraint error: coefficient cannot be zero";
|
|
}
|
|
if (term.literals.empty()) return "constraint error: empty literals";
|
|
if (term.literals.size() == 1) {
|
|
if (!RefIsPositive(term.literals[0])) {
|
|
return "constraint error: linear terms must use positive literals";
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
// Normalize and validate objective.
|
|
if (objective_.empty()) return ""; // No objective.
|
|
|
|
for (const PbTerm& term : objective_) {
|
|
if (term.coeff == 0) return "objective error: coefficient cannot be zero";
|
|
if (term.literals.empty()) return "objective error: empty literals";
|
|
if (term.literals.size() == 1) {
|
|
if (!RefIsPositive(term.literals[0])) {
|
|
return "objective error: linear terms must use positive literals";
|
|
}
|
|
return "";
|
|
}
|
|
}
|
|
|
|
return "";
|
|
}
|
|
|
|
static int PbLiteralToCpModelLiteral(int pb_literal) {
|
|
return pb_literal > 0 ? pb_literal - 1 : -pb_literal;
|
|
}
|
|
|
|
bool ParseInt64Into(const std::string& word, int64_t* value) {
|
|
if (!absl::SimpleAtoi(word, value)) {
|
|
VLOG(1) << "Failed to parse int64_t: " << word;
|
|
model_is_supported_ = false;
|
|
return false;
|
|
}
|
|
return true;
|
|
}
|
|
|
|
static int ParseIndex(const std::string& word) {
|
|
int index;
|
|
CHECK(absl::SimpleAtoi(word, &index));
|
|
return index;
|
|
}
|
|
|
|
int GetVariable(const PbTerm& term, CpModelProto* model) {
|
|
CHECK(!term.literals.empty());
|
|
if (term.literals.size() == 1) {
|
|
CHECK(RefIsPositive(term.literals[0]));
|
|
return term.literals[0];
|
|
}
|
|
|
|
const auto it = product_to_var_.find(term.literals);
|
|
if (it != product_to_var_.end()) {
|
|
return it->second;
|
|
}
|
|
|
|
const int var_index = model->variables_size();
|
|
IntegerVariableProto* var_proto = model->add_variables();
|
|
var_proto->add_domain(0);
|
|
var_proto->add_domain(1);
|
|
|
|
product_to_var_[term.literals] = var_index;
|
|
|
|
// Link the new variable to the terms.
|
|
// var_index => and(literals).
|
|
ConstraintProto* var_to_literals = model->add_constraints();
|
|
var_to_literals->add_enforcement_literal(var_index);
|
|
|
|
// and(literals) => var_index.
|
|
ConstraintProto* literals_to_var = model->add_constraints();
|
|
literals_to_var->mutable_bool_and()->add_literals(var_index);
|
|
|
|
for (const int proto_literal : term.literals) {
|
|
var_to_literals->mutable_bool_and()->add_literals(proto_literal);
|
|
literals_to_var->add_enforcement_literal(proto_literal);
|
|
}
|
|
|
|
return var_index;
|
|
}
|
|
|
|
void BuildModel(CpModelProto* model) {
|
|
// We know how many variables we have, so we can add them all.
|
|
for (int i = 0; i < num_variables_; ++i) {
|
|
IntegerVariableProto* var = model->add_variables();
|
|
var->add_domain(0);
|
|
var->add_domain(1);
|
|
}
|
|
|
|
for (const PbConstraint& constraint : constraints_) {
|
|
ConstraintProto* ct = model->add_constraints();
|
|
LinearConstraintProto* lin = ct->mutable_linear();
|
|
for (const PbTerm& term : constraint.terms) {
|
|
lin->add_vars(GetVariable(term, model));
|
|
lin->add_coeffs(term.coeff);
|
|
}
|
|
if (constraint.type == GE_OPERATION) {
|
|
lin->add_domain(constraint.rhs);
|
|
lin->add_domain(std::numeric_limits<int64_t>::max());
|
|
} else if (constraint.type == EQ_OPERATION) {
|
|
lin->add_domain(constraint.rhs);
|
|
lin->add_domain(constraint.rhs);
|
|
} else {
|
|
LOG(FATAL) << "Unsupported operation: " << constraint.type;
|
|
}
|
|
|
|
if (constraint.soft_cost != std::numeric_limits<int64_t>::max()) {
|
|
const int violation_var_index = model->variables_size();
|
|
IntegerVariableProto* violation_var = model->add_variables();
|
|
violation_var->add_domain(0);
|
|
violation_var->add_domain(1);
|
|
|
|
// Update the objective.
|
|
model->mutable_objective()->add_vars(violation_var_index);
|
|
model->mutable_objective()->add_coeffs(constraint.soft_cost);
|
|
|
|
// Add the enforcement literal to ct.
|
|
ct->add_enforcement_literal(NegatedRef(violation_var_index));
|
|
}
|
|
}
|
|
|
|
if (!objective_.empty()) {
|
|
CpObjectiveProto* obj = model->mutable_objective();
|
|
for (const PbTerm& term : objective_) {
|
|
obj->add_vars(GetVariable(term, model));
|
|
obj->add_coeffs(term.coeff);
|
|
}
|
|
}
|
|
}
|
|
|
|
int num_variables_;
|
|
std::vector<PbTerm> objective_;
|
|
std::vector<PbConstraint> constraints_;
|
|
absl::flat_hash_map<absl::Span<const int>, int> product_to_var_;
|
|
bool model_is_supported_ = true;
|
|
};
|
|
|
|
} // namespace sat
|
|
} // namespace operations_research
|
|
|
|
#endif // OR_TOOLS_SAT_OPB_READER_H_
|