| 1 | /*
|
| 2 | * Souffle - A Datalog Compiler
|
| 3 | * Copyright (c) 2016, The Souffle Developers. All rights reserved
|
| 4 | * Licensed under the Universal Permissive License v 1.0 as shown at:
|
| 5 | * - https://opensource.org/licenses/UPL
|
| 6 | * - <souffle root>/licenses/SOUFFLE-UPL.txt
|
| 7 | */
|
| 8 |
|
| 9 | #pragma once
|
| 10 |
|
| 11 | #include "souffle/profile/CellInterface.h"
|
| 12 | #include "souffle/profile/Row.h"
|
| 13 | #include "souffle/profile/Table.h"
|
| 14 | #include <algorithm>
|
| 15 | #include <chrono>
|
| 16 | #include <cmath>
|
| 17 | #include <cstdio>
|
| 18 | #include <fstream>
|
| 19 | #include <iomanip>
|
| 20 | #include <ios>
|
| 21 | #include <memory>
|
| 22 | #include <sstream>
|
| 23 | #include <string>
|
| 24 | #include <vector>
|
| 25 |
|
| 26 | #ifndef _MSC_VER
|
| 27 | #include <unistd.h>
|
| 28 | #endif
|
| 29 |
|
| 30 | #include <sys/stat.h>
|
| 31 |
|
| 32 | namespace souffle {
|
| 33 | namespace profile {
|
| 34 |
|
| 35 | /*
|
| 36 | * A series of functions necessary throughout the code
|
| 37 | * Mostly string manipulation
|
| 38 | */
|
| 39 | namespace Tools {
|
| 40 | static const std::vector<std::string> abbreviations{
|
| 41 | "K", "M", "B", "t", "q", "Q", "s", "S", "o", "n", "d", "U"};
|
| 42 |
|
| 43 | inline std::string formatNum(double amount) {
|
| 44 | std::stringstream ss;
|
| 45 | ss << amount;
|
| 46 | return ss.str();
|
| 47 | }
|
| 48 |
|
| 49 | inline std::string formatNum(int precision, int64_t amount) {
|
| 50 | // assumes number is < 999*10^12
|
| 51 | if (amount == 0) {
|
| 52 | return "0";
|
| 53 | }
|
| 54 |
|
| 55 | if (precision <= 0) {
|
| 56 | return std::to_string(amount);
|
| 57 | }
|
| 58 |
|
| 59 | std::string result;
|
| 60 |
|
| 61 | if (amount < 1000) {
|
| 62 | return std::to_string(amount);
|
| 63 | }
|
| 64 |
|
| 65 | for (std::size_t i = 0; i < abbreviations.size(); ++i) {
|
| 66 | if (amount > std::pow(1000, i + 2)) {
|
| 67 | continue;
|
| 68 | }
|
| 69 |
|
| 70 | double r = amount / std::pow(1000, i + 1);
|
| 71 | result = std::to_string(r);
|
| 72 |
|
| 73 | if (r >= 100) { // 1000 > result >= 100
|
| 74 |
|
| 75 | switch (precision) {
|
| 76 | case 1: result = result.substr(0, 1) + "00"; break;
|
| 77 | case 2: result = result.substr(0, 2) + "0"; break;
|
| 78 | case 3: result = result.substr(0, 3); break;
|
| 79 | default: result = result.substr(0, precision + 1);
|
| 80 | }
|
| 81 | } else if (r >= 10) { // 100 > result >= 10
|
| 82 | switch (precision) {
|
| 83 | case 1: result = result.substr(0, 1) + "0"; break;
|
| 84 | case 2: result = result.substr(0, 2); break;
|
| 85 | default: result = result.substr(0, precision + 1);
|
| 86 | }
|
| 87 | } else { // 10 > result > 0
|
| 88 | switch (precision) {
|
| 89 | case 1: result = result.substr(0, 1); break;
|
| 90 | default: result = result.substr(0, precision + 1);
|
| 91 | }
|
| 92 | }
|
| 93 | result += abbreviations.at(i);
|
| 94 | return result;
|
| 95 | }
|
| 96 | // If we ever have integers too large to handle, fall back to this
|
| 97 | return std::to_string(amount);
|
| 98 | }
|
| 99 |
|
| 100 | inline std::string formatMemory(uint64_t kbytes) {
|
| 101 | if (kbytes < 1024UL * 2UL) {
|
| 102 | return std::to_string(kbytes) + "kB";
|
| 103 | } else if (kbytes < 1024UL * 1024UL * 2UL) {
|
| 104 | return std::to_string(kbytes / 1024UL) + "MB";
|
| 105 | } else if (kbytes < 1024UL * 1024UL * 1024UL * 2UL) {
|
| 106 | return std::to_string(kbytes / (1024UL * 1024UL)) + "GB";
|
| 107 | }
|
| 108 | return std::to_string(kbytes / (1024UL * 1024UL * 1024UL)) + "TB";
|
| 109 | }
|
| 110 |
|
| 111 | inline std::string formatTime(std::chrono::microseconds number) {
|
| 112 | uint64_t sec = number.count() / 1000000;
|
| 113 | if (sec >= 100) {
|
| 114 | uint64_t min = static_cast<uint64_t>(std::floor(sec / 60));
|
| 115 | if (min >= 100) {
|
| 116 | uint64_t hours = static_cast<uint64_t>(std::floor(min / 60));
|
| 117 | if (hours >= 100) {
|
| 118 | uint64_t days = static_cast<uint64_t>(std::floor(hours / 24));
|
| 119 | return std::to_string(days) + "D";
|
| 120 | }
|
| 121 | return std::to_string(hours) + "h";
|
| 122 | }
|
| 123 | if (min < 10) {
|
| 124 | // temp should always be 1 digit long
|
| 125 | uint64_t temp = static_cast<uint64_t>(std::floor((sec - (min * 60.0)) * 10.0 / 6.0));
|
| 126 | return std::to_string(min) + "." + std::to_string(temp).substr(0, 1) + "m";
|
| 127 | }
|
| 128 | return std::to_string(min) + "m";
|
| 129 | } else if (sec >= 10) {
|
| 130 | return std::to_string(sec) + "s";
|
| 131 | } else if (number.count() >= 1000000) {
|
| 132 | std::string temp = std::to_string(number.count() / 100);
|
| 133 | return temp.substr(0, 1) + "." + temp.substr(1, 2) + "s";
|
| 134 | } else if (number.count() >= 100000) {
|
| 135 | std::string temp = std::to_string(number.count() / 1000);
|
| 136 | return "." + temp.substr(0, 3) + "s";
|
| 137 | } else if (number.count() >= 10000) {
|
| 138 | std::string temp = std::to_string(number.count() / 1000);
|
| 139 | return ".0" + temp.substr(0, 2) + "s";
|
| 140 | } else if (number.count() >= 1000) {
|
| 141 | std::string temp = std::to_string(number.count() / 1000);
|
| 142 | return ".00" + temp.substr(0, 1) + "s";
|
| 143 | }
|
| 144 |
|
| 145 | return ".000s";
|
| 146 | }
|
| 147 |
|
| 148 | inline std::vector<std::vector<std::string>> formatTable(Table table, int precision) {
|
| 149 | std::vector<std::vector<std::string>> result;
|
| 150 | for (auto& row : table.getRows()) {
|
| 151 | std::vector<std::string> result_row;
|
| 152 | for (auto& cell : row->getCells()) {
|
| 153 | if (cell != nullptr) {
|
| 154 | result_row.push_back(cell->toString(precision));
|
| 155 | } else {
|
| 156 | result_row.push_back("-");
|
| 157 | }
|
| 158 | }
|
| 159 | result.push_back(result_row);
|
| 160 | }
|
| 161 | return result;
|
| 162 | }
|
| 163 |
|
| 164 | /** @brief split on the delimiter */
|
| 165 | inline std::vector<std::string> split(std::string toSplit, std::string delimiter) {
|
| 166 | std::vector<std::string> elements;
|
| 167 | std::string::size_type lastPos = 0;
|
| 168 | auto pos = toSplit.find(delimiter, lastPos);
|
| 169 |
|
| 170 | while (pos != std::string::npos) {
|
| 171 | if (pos > 0) {
|
| 172 | std::string newElement = toSplit.substr(lastPos, pos - lastPos);
|
| 173 | elements.push_back(newElement);
|
| 174 | }
|
| 175 | lastPos = pos + delimiter.size();
|
| 176 | pos = toSplit.find(delimiter, lastPos);
|
| 177 | }
|
| 178 | if (lastPos < toSplit.size()) {
|
| 179 | elements.push_back(toSplit.substr(lastPos));
|
| 180 | }
|
| 181 |
|
| 182 | return elements;
|
| 183 | }
|
| 184 |
|
| 185 | inline std::string trimWhitespace(std::string str) {
|
| 186 | std::string whitespace = " \t";
|
| 187 | std::size_t first = str.find_first_not_of(whitespace);
|
| 188 | if (first != std::string::npos) {
|
| 189 | str.erase(0, first);
|
| 190 | std::size_t last = str.find_last_not_of(whitespace);
|
| 191 | str.erase(last + 1);
|
| 192 | } else {
|
| 193 | str.clear();
|
| 194 | }
|
| 195 |
|
| 196 | return str;
|
| 197 | }
|
| 198 |
|
| 199 | inline bool file_exists(const std::string& name) {
|
| 200 | struct stat buffer = {};
|
| 201 | if (stat(name.c_str(), &buffer) == 0) {
|
| 202 | if ((buffer.st_mode & S_IFMT) != 0) {
|
| 203 | return true;
|
| 204 | }
|
| 205 | }
|
| 206 | return false;
|
| 207 | }
|
| 208 | /** @brief Remove \n and \t characters, \n and \t sequence of two chars, and wrapping quotes */
|
| 209 | inline std::string cleanString(std::string val) {
|
| 210 | if (val.size() < 2) {
|
| 211 | return val;
|
| 212 | }
|
| 213 |
|
| 214 | std::size_t start_pos = 0;
|
| 215 | while ((start_pos = val.find('\\', start_pos)) != std::string::npos) {
|
| 216 | val.erase(start_pos, 1);
|
| 217 | if (start_pos < val.size()) {
|
| 218 | if (val[start_pos] == 'n' || val[start_pos] == 't') {
|
| 219 | val.replace(start_pos, 1, " ");
|
| 220 | }
|
| 221 | }
|
| 222 | }
|
| 223 |
|
| 224 | if (val.at(0) == '"' && val.at(val.size() - 1) == '"') {
|
| 225 | val = val.substr(1, val.size() - 2);
|
| 226 | }
|
| 227 |
|
| 228 | std::replace(val.begin(), val.end(), '\n', ' ');
|
| 229 | std::replace(val.begin(), val.end(), '\t', ' ');
|
| 230 |
|
| 231 | return val;
|
| 232 | }
|
| 233 |
|
| 234 | /** @brief escape escapes and quotes, and remove surrounding quotes */
|
| 235 | inline std::string cleanJsonOut(std::string value) {
|
| 236 | if (value.size() >= 2) {
|
| 237 | if (value.at(0) == '"' && value.at(value.size() - 1) == '"') {
|
| 238 | value = value.substr(1, value.size() - 2);
|
| 239 | }
|
| 240 | }
|
| 241 |
|
| 242 | std::size_t start_pos = 0;
|
| 243 | while ((start_pos = value.find('\\', start_pos)) != std::string::npos) {
|
| 244 | value.replace(start_pos, 1, "\\\\");
|
| 245 | start_pos += 2;
|
| 246 | }
|
| 247 | start_pos = 0;
|
| 248 | while ((start_pos = value.find('"', start_pos)) != std::string::npos) {
|
| 249 | value.replace(start_pos, 1, "\\\"");
|
| 250 | start_pos += 2;
|
| 251 | }
|
| 252 | return value;
|
| 253 | }
|
| 254 |
|
| 255 | /** @brief Convert doubles to NaN or scientific notation */
|
| 256 | inline std::string cleanJsonOut(double val) {
|
| 257 | if (std::isnan(val)) {
|
| 258 | return "NaN";
|
| 259 | }
|
| 260 | std::ostringstream ss;
|
| 261 | ss << std::scientific << std::setprecision(6) << val;
|
| 262 | return ss.str();
|
| 263 | }
|
| 264 | } // namespace Tools
|
| 265 |
|
| 266 | } // namespace profile
|
| 267 | } // namespace souffle
|