|
| 1 | +/* |
| 2 | +Copyright 2024 Huawei Technologies Co., Ltd. |
| 3 | +
|
| 4 | +Licensed under the Apache License, Version 2.0 (the "License"); |
| 5 | +you may not use this file except in compliance with the License. |
| 6 | +You may obtain a copy of the License at |
| 7 | +
|
| 8 | + http://www.apache.org/licenses/LICENSE-2.0 |
| 9 | +
|
| 10 | +Unless required by applicable law or agreed to in writing, software |
| 11 | +distributed under the License is distributed on an "AS IS" BASIS, |
| 12 | +WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 13 | +See the License for the specific language governing permissions and |
| 14 | +limitations under the License. |
| 15 | +
|
| 16 | +@author Toni Boehnlein, Christos Matzoros, Benjamin Lozes, Pal Andras Papp, Raphael S. Steiner |
| 17 | +*/ |
| 18 | + |
| 19 | +#pragma once |
| 20 | + |
| 21 | +#include <filesystem> |
| 22 | +#include <fstream> |
| 23 | +#include <iostream> |
| 24 | +#include <limits> |
| 25 | +#include <sstream> |
| 26 | +#include <string> |
| 27 | +#include <vector> |
| 28 | + |
| 29 | +#include "osp/auxiliary/io/filepath_checker.hpp" |
| 30 | +#include "osp/partitioning/model/hypergraph.hpp" |
| 31 | + |
| 32 | +namespace osp { |
| 33 | +namespace file_reader { |
| 34 | + |
| 35 | +// reads a Hypergraph from a file in hMetis format |
| 36 | +template <typename IndexType, typename WorkwType, typename MemwType, typename CommwType> |
| 37 | +bool ReadHypergraphMetisFormat(std::ifstream &infile, |
| 38 | + Hypergraph<IndexType, WorkwType, MemwType, CommwType> &hgraph) { |
| 39 | + std::string line; |
| 40 | + |
| 41 | + // Skip comments or empty lines (robustly) |
| 42 | + while (std::getline(infile, line)) { |
| 43 | + if (line.empty() || line[0] == '%') { |
| 44 | + continue; |
| 45 | + } |
| 46 | + |
| 47 | + // Null byte check |
| 48 | + if (line.find('\0') != std::string::npos) { |
| 49 | + std::cerr << "Error: Null byte detected in header line.\n"; |
| 50 | + return false; |
| 51 | + } |
| 52 | + |
| 53 | + if (line.size() > MAX_LINE_LENGTH) { |
| 54 | + std::cerr << "Error: Line too long, possible malformed or malicious file.\n"; |
| 55 | + return false; |
| 56 | + } |
| 57 | + break; // We found the actual header line |
| 58 | + } |
| 59 | + |
| 60 | + if (infile.eof()) { |
| 61 | + std::cerr << "Error: Unexpected end of file while reading header.\n"; |
| 62 | + return false; |
| 63 | + } |
| 64 | + |
| 65 | + IndexType numVertices, numHyperedges, weightCode = 0; |
| 66 | + |
| 67 | + std::istringstream headerStream(line); |
| 68 | + if (!(headerStream >> numHyperedges >> numVertices) || numVertices <= 0 || numHyperedges <= 0) { |
| 69 | + std::cerr << "Error: Invalid header.\n"; |
| 70 | + return false; |
| 71 | + } |
| 72 | + |
| 73 | + headerStream >> weightCode; |
| 74 | + bool hasVertexWeights = false, hasHyperedgeWeights = false; |
| 75 | + |
| 76 | + switch(weightCode) { |
| 77 | + case 0: |
| 78 | + break; |
| 79 | + case 1: |
| 80 | + hasHyperedgeWeights = true; |
| 81 | + break; |
| 82 | + case 10: |
| 83 | + hasVertexWeights = true; |
| 84 | + break; |
| 85 | + case 11: |
| 86 | + hasHyperedgeWeights = true; |
| 87 | + hasVertexWeights = true; |
| 88 | + break; |
| 89 | + default: |
| 90 | + std::cerr << "Error: Invalid weight code in header (must be 0, 1, 10 or 11).\n"; |
| 91 | + return false; |
| 92 | + } |
| 93 | + |
| 94 | + hgraph.Reset(numVertices, 0); |
| 95 | + |
| 96 | + IndexType edgesRead = 0; |
| 97 | + while (edgesRead < numHyperedges && std::getline(infile, line)) { |
| 98 | + if (line.empty() || line[0] == '%') { |
| 99 | + continue; |
| 100 | + } |
| 101 | + if (line.size() > MAX_LINE_LENGTH) { |
| 102 | + std::cerr << "Error: Line too long.\n"; |
| 103 | + return false; |
| 104 | + } |
| 105 | + |
| 106 | + std::istringstream edgeStream(line); |
| 107 | + |
| 108 | + CommwType weight = 1; |
| 109 | + if (hasHyperedgeWeights && !(edgeStream >> weight)) { |
| 110 | + std::cerr << "Error: Malformed hyperedge row (weight).\n"; |
| 111 | + return false; |
| 112 | + } |
| 113 | + |
| 114 | + std::vector<IndexType> verticesInHyperedge; |
| 115 | + IndexType vertex; |
| 116 | + while (edgeStream >> vertex) { |
| 117 | + --vertex; // Convert to 0-based |
| 118 | + if (vertex < 0 || vertex >= numVertices) { |
| 119 | + std::cerr << "Error: Malformed hyperedge row (vertex entry).\n"; |
| 120 | + return false; |
| 121 | + } |
| 122 | + verticesInHyperedge.push_back(vertex); |
| 123 | + } |
| 124 | + |
| 125 | + if (verticesInHyperedge.empty()) { |
| 126 | + std::cerr << "Error: Empty hyperedge in file.\n"; |
| 127 | + return false; |
| 128 | + } |
| 129 | + |
| 130 | + std::sort(verticesInHyperedge.begin(), verticesInHyperedge.end()); |
| 131 | + for (unsigned index = 0; index < verticesInHyperedge.size() - 1; ++index) { |
| 132 | + if (verticesInHyperedge[index] == verticesInHyperedge[index + 1]) { |
| 133 | + std::cerr << "Error: Malformed hyperedge row (same vertex appears multiple times).\n"; |
| 134 | + return false; |
| 135 | + } |
| 136 | + } |
| 137 | + |
| 138 | + hgraph.AddHyperedge(verticesInHyperedge, weight); |
| 139 | + ++edgesRead; |
| 140 | + } |
| 141 | + |
| 142 | + if (edgesRead != numHyperedges) { |
| 143 | + std::cerr << "Error: Incomplete hyperedge entries.\n"; |
| 144 | + return false; |
| 145 | + } |
| 146 | + |
| 147 | + if (hasVertexWeights) { |
| 148 | + IndexType vertexWeightsRead = 0; |
| 149 | + while (vertexWeightsRead < numVertices && std::getline(infile, line)) { |
| 150 | + if (line.empty() || line[0] == '%') { |
| 151 | + continue; |
| 152 | + } |
| 153 | + if (line.size() > MAX_LINE_LENGTH) { |
| 154 | + std::cerr << "Error: Line too long.\n"; |
| 155 | + return false; |
| 156 | + } |
| 157 | + |
| 158 | + std::istringstream weightStream(line); |
| 159 | + WorkwType weight = 1; |
| 160 | + if (!(weightStream >> weight)) { |
| 161 | + std::cerr << "Error: Malformed vertex weight row.\n"; |
| 162 | + return false; |
| 163 | + } |
| 164 | + |
| 165 | + hgraph.SetVertexWorkWeight(vertexWeightsRead, weight); |
| 166 | + |
| 167 | + ++vertexWeightsRead; |
| 168 | + } |
| 169 | + |
| 170 | + if (vertexWeightsRead != numVertices) { |
| 171 | + std::cerr << "Error: Incomplete vertex weight lines.\n"; |
| 172 | + return false; |
| 173 | + } |
| 174 | + |
| 175 | + } |
| 176 | + |
| 177 | + while (std::getline(infile, line)) { |
| 178 | + if (!line.empty() && line[0] != '%') { |
| 179 | + std::cerr << "Error: Extra data after hypergraph content.\n"; |
| 180 | + return false; |
| 181 | + } |
| 182 | + } |
| 183 | + |
| 184 | + return true; |
| 185 | +} |
| 186 | + |
| 187 | +template <typename IndexType, typename WorkwType, typename MemwType, typename CommwType> |
| 188 | +bool ReadHypergraphMetisFormat(const std::string &filename, |
| 189 | + Hypergraph<IndexType, WorkwType, MemwType, CommwType> &hgraph) { |
| 190 | + // Ensure the file is .hmetis format |
| 191 | + if (std::filesystem::path(filename).extension() != ".hmetis") { |
| 192 | + std::cerr << "Error: Only .hmetis files are accepted.\n"; |
| 193 | + return false; |
| 194 | + } |
| 195 | + |
| 196 | + if (!IsPathSafe(filename)) { |
| 197 | + std::cerr << "Error: Unsafe file path (potential traversal attack).\n"; |
| 198 | + return false; |
| 199 | + } |
| 200 | + |
| 201 | + if (std::filesystem::is_symlink(filename)) { |
| 202 | + std::cerr << "Error: Symbolic links are not allowed.\n"; |
| 203 | + return false; |
| 204 | + } |
| 205 | + |
| 206 | + if (!std::filesystem::is_regular_file(filename)) { |
| 207 | + std::cerr << "Error: Input is not a regular file.\n"; |
| 208 | + return false; |
| 209 | + } |
| 210 | + |
| 211 | + std::ifstream infile(filename); |
| 212 | + if (!infile.is_open()) { |
| 213 | + std::cerr << "Error: Failed to open file.\n"; |
| 214 | + return false; |
| 215 | + } |
| 216 | + |
| 217 | + return ReadHypergraphMetisFormat(infile, hgraph); |
| 218 | +} |
| 219 | + |
| 220 | +} // namespace file_reader |
| 221 | + |
| 222 | +} // namespace osp |
0 commit comments