Files
ArffFiles/ArffFiles.hpp

817 lines
32 KiB
C++

#ifndef ARFFFILES_HPP
#define ARFFFILES_HPP
#include <string>
#include <vector>
#include <map>
#include <set>
#include <sstream>
#include <fstream>
#include <cctype> // std::isdigit
#include <algorithm> // std::all_of std::transform
#include <filesystem> // For file size checking
// Summary information structure for ARFF files
struct ArffSummary {
size_t numSamples; // Number of data samples
size_t numFeatures; // Number of feature attributes (excluding class)
size_t numClasses; // Number of different class values
std::string className; // Name of the class attribute
std::string classType; // Type/values of the class attribute
std::vector<std::string> classLabels; // List of unique class values
std::vector<std::pair<std::string, std::string>> featureInfo; // Feature names and types
};
/**
* @brief Header-only C++17 library for parsing ARFF (Attribute-Relation File Format) files
*
* This class provides functionality to load and parse ARFF files, automatically detecting
* numeric vs categorical features and performing factorization of categorical attributes.
*
* @warning THREAD SAFETY: This class is NOT thread-safe!
*
* Thread Safety Considerations:
* - Multiple instances can be used safely in different threads (each instance is independent)
* - A single instance MUST NOT be accessed concurrently from multiple threads
* - All member functions (including getters) modify or access mutable state
* - Static methods (summary, trim, split) are thread-safe as they don't access instance state
*
* Memory Safety:
* - Built-in protection against resource exhaustion with configurable limits
* - File size limit: 100 MB (DEFAULT_MAX_FILE_SIZE)
* - Sample count limit: 1 million samples (DEFAULT_MAX_SAMPLES)
* - Feature count limit: 10,000 features (DEFAULT_MAX_FEATURES)
*
* Usage Patterns:
* - Single-threaded: Create one instance, call load(), then access data via getters
* - Multi-threaded: Create separate instances per thread, or use external synchronization
*
* @example
* // Thread-safe usage pattern:
* void processFile(const std::string& filename) {
* ArffFiles arff; // Each thread has its own instance
* arff.load(filename);
* auto X = arff.getX();
* auto y = arff.getY();
* // Process data...
* }
*
* @example
* // UNSAFE usage pattern:
* ArffFiles globalArff; // Global instance
* // Thread 1: globalArff.load("file1.arff"); // UNSAFE!
* // Thread 2: globalArff.load("file2.arff"); // UNSAFE!
*/
class ArffFiles {
const std::string VERSION = "1.1.0";
// Memory usage limits (configurable via environment variables)
static constexpr size_t DEFAULT_MAX_FILE_SIZE = 100 * 1024 * 1024; // 100 MB
static constexpr size_t DEFAULT_MAX_SAMPLES = 1000000; // 1 million samples
static constexpr size_t DEFAULT_MAX_FEATURES = 10000; // 10k features
public:
ArffFiles() = default;
// Move constructor
ArffFiles(ArffFiles&& other) noexcept
: lines(std::move(other.lines))
, numeric_features(std::move(other.numeric_features))
, attributes(std::move(other.attributes))
, className(std::move(other.className))
, classType(std::move(other.classType))
, states(std::move(other.states))
, X(std::move(other.X))
, y(std::move(other.y))
{
// Other object is left in a valid but unspecified state
}
// Move assignment operator
ArffFiles& operator=(ArffFiles&& other) noexcept
{
if (this != &other) {
lines = std::move(other.lines);
numeric_features = std::move(other.numeric_features);
attributes = std::move(other.attributes);
className = std::move(other.className);
classType = std::move(other.classType);
states = std::move(other.states);
X = std::move(other.X);
y = std::move(other.y);
}
return *this;
}
// Copy constructor (explicitly defaulted)
ArffFiles(const ArffFiles& other) = default;
// Copy assignment operator (explicitly defaulted)
ArffFiles& operator=(const ArffFiles& other) = default;
// Destructor (explicitly defaulted)
~ArffFiles() = default;
void load(const std::string& fileName, bool classLast = true)
{
if (fileName.empty()) {
throw std::invalid_argument("File name cannot be empty");
}
int labelIndex;
loadCommon(fileName);
// Validate we have attributes before accessing them
if (attributes.empty()) {
throw std::invalid_argument("No attributes found in file");
}
if (classLast) {
className = std::get<0>(attributes.back());
classType = std::get<1>(attributes.back());
attributes.pop_back();
labelIndex = static_cast<int>(attributes.size());
} else {
className = std::get<0>(attributes.front());
classType = std::get<1>(attributes.front());
attributes.erase(attributes.begin());
labelIndex = 0;
}
// Validate class name is not empty
if (className.empty()) {
throw std::invalid_argument("Class attribute name cannot be empty");
}
preprocessDataset(labelIndex);
generateDataset(labelIndex);
}
void load(const std::string& fileName, const std::string& name)
{
if (fileName.empty()) {
throw std::invalid_argument("File name cannot be empty");
}
if (name.empty()) {
throw std::invalid_argument("Class name cannot be empty");
}
int labelIndex;
loadCommon(fileName);
// Validate we have attributes before searching
if (attributes.empty()) {
throw std::invalid_argument("No attributes found in file");
}
bool found = false;
for (size_t i = 0; i < attributes.size(); ++i) {
if (attributes[i].first == name) {
className = std::get<0>(attributes[i]);
classType = std::get<1>(attributes[i]);
attributes.erase(attributes.begin() + i);
labelIndex = static_cast<int>(i);
found = true;
break;
}
}
if (!found) {
throw std::invalid_argument("Class name '" + name + "' not found in attributes");
}
preprocessDataset(labelIndex);
generateDataset(labelIndex);
}
// Static method to get summary information without loading all data (default: class is last)
static ArffSummary summary(const std::string& fileName)
{
return summary(fileName, true);
}
// Static method to get summary information without loading all data
static ArffSummary summary(const std::string& fileName, bool classLast)
{
if (fileName.empty()) {
throw std::invalid_argument("File name cannot be empty");
}
return summarizeFile(fileName, classLast);
}
// Static method to get summary information with specified class attribute (const char* overload)
static ArffSummary summary(const std::string& fileName, const char* className)
{
return summary(fileName, std::string(className));
}
// Static method to get summary information with specified class attribute
static ArffSummary summary(const std::string& fileName, const std::string& className)
{
if (fileName.empty()) {
throw std::invalid_argument("File name cannot be empty");
}
if (className.empty()) {
throw std::invalid_argument("Class name cannot be empty");
}
return summarizeFile(fileName, className);
}
const std::vector<std::string>& getLines() const { return lines; }
size_t getSize() const { return lines.size(); }
std::string getClassName() const { return className; }
std::string getClassType() const { return classType; }
const std::map<std::string, std::vector<std::string>>& getStates() const { return states; }
std::vector<std::string> getLabels() const { return states.at(className); }
static std::string trim(const std::string& source)
{
std::string s(source);
s.erase(0, s.find_first_not_of(" '\n\r\t"));
s.erase(s.find_last_not_of(" '\n\r\t") + 1);
return s;
}
std::vector<std::vector<float>>& getX() { return X; }
const std::vector<std::vector<float>>& getX() const { return X; }
std::vector<int>& getY() { return y; }
const std::vector<int>& getY() const { return y; }
const std::map<std::string, bool>& getNumericAttributes() const { return numeric_features; }
const std::vector<std::pair<std::string, std::string>>& getAttributes() const { return attributes; };
// Move-enabled getters for efficient data transfer
// WARNING: These methods move data OUT of the object, leaving it in an empty but valid state
// Use these when you want to transfer ownership of large data structures for performance
std::vector<std::vector<float>> moveX() noexcept { return std::move(X); }
std::vector<int> moveY() noexcept { return std::move(y); }
std::vector<std::string> moveLines() noexcept { return std::move(lines); }
std::map<std::string, std::vector<std::string>> moveStates() noexcept { return std::move(states); }
std::vector<std::pair<std::string, std::string>> moveAttributes() noexcept { return std::move(attributes); }
std::map<std::string, bool> moveNumericAttributes() noexcept { return std::move(numeric_features); }
std::vector<std::string> split(const std::string& text, char delimiter)
{
std::vector<std::string> result;
std::stringstream ss(text);
std::string token;
while (std::getline(ss, token, delimiter)) {
result.push_back(trim(token));
}
return result;
}
std::string version() const { return VERSION; }
private:
// Helper function to validate file path for security
static void validateFilePath(const std::string& fileName) {
if (fileName.empty()) {
throw std::invalid_argument("File path cannot be empty");
}
// Check for path traversal attempts
if (fileName.find("..") != std::string::npos) {
throw std::invalid_argument("Path traversal detected in file path: " + fileName);
}
// Check for absolute paths starting with / (Unix) or drive letters (Windows)
if (fileName[0] == '/' || (fileName.length() >= 3 && fileName[1] == ':')) {
// Allow absolute paths but log a warning - this is for user awareness
// In production, you might want to restrict this based on your security requirements
}
// Check for suspicious characters that could be used in path manipulation
const std::string suspiciousChars = "\x00\x01\x02\x03\x04\x05\x06\x07\x08\x09\x0b\x0c\x0e\x0f\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f";
for (char c : suspiciousChars) {
if (fileName.find(c) != std::string::npos) {
throw std::invalid_argument("Invalid character detected in file path");
}
}
// Check for excessively long paths (potential buffer overflow attempts)
constexpr size_t MAX_PATH_LENGTH = 4096; // Common filesystem limit
if (fileName.length() > MAX_PATH_LENGTH) {
throw std::invalid_argument("File path too long (exceeds " + std::to_string(MAX_PATH_LENGTH) + " characters)");
}
// Additional validation using filesystem operations when available
try {
// Check if the file exists and validate its canonical path
if (std::filesystem::exists(fileName)) {
std::filesystem::path normalizedPath = std::filesystem::canonical(fileName);
std::string normalizedStr = normalizedPath.string();
// Check if normalized path still contains traversal attempts
if (normalizedStr.find("..") != std::string::npos) {
throw std::invalid_argument("Path traversal detected after normalization: " + normalizedStr);
}
}
} catch (const std::filesystem::filesystem_error& e) {
// If filesystem operations fail, we can still proceed with basic validation
// This ensures compatibility with systems where filesystem might not be fully available
}
}
// Helper function to validate resource usage limits
static void validateResourceLimits(const std::string& fileName, size_t sampleCount = 0, size_t featureCount = 0) {
// Check file size limit
try {
if (std::filesystem::exists(fileName)) {
auto fileSize = std::filesystem::file_size(fileName);
if (fileSize > DEFAULT_MAX_FILE_SIZE) {
throw std::invalid_argument("File size (" + std::to_string(fileSize) + " bytes) exceeds maximum allowed size (" + std::to_string(DEFAULT_MAX_FILE_SIZE) + " bytes)");
}
}
} catch (const std::filesystem::filesystem_error&) {
// If filesystem operations fail, continue without size checking
// This ensures compatibility with systems where filesystem might not be available
}
// Check sample count limit
if (sampleCount > DEFAULT_MAX_SAMPLES) {
throw std::invalid_argument("Number of samples (" + std::to_string(sampleCount) + ") exceeds maximum allowed (" + std::to_string(DEFAULT_MAX_SAMPLES) + ")");
}
// Check feature count limit
if (featureCount > DEFAULT_MAX_FEATURES) {
throw std::invalid_argument("Number of features (" + std::to_string(featureCount) + ") exceeds maximum allowed (" + std::to_string(DEFAULT_MAX_FEATURES) + ")");
}
}
protected:
std::vector<std::string> lines;
std::map<std::string, bool> numeric_features;
std::vector<std::pair<std::string, std::string>> attributes;
std::string className;
std::string classType;
std::vector<std::vector<float>> X; // X[feature][sample] - feature-major layout
std::vector<int> y;
std::map<std::string, std::vector<std::string>> states;
private:
void preprocessDataset(int labelIndex)
{
//
// Learn the numeric features
//
numeric_features.clear();
for (const auto& attribute : attributes) {
auto feature = attribute.first;
if (feature == className)
continue;
auto values = attribute.second;
std::transform(values.begin(), values.end(), values.begin(), ::toupper);
// Enhanced attribute type detection
bool isNumeric = values == "REAL" || values == "INTEGER" || values == "NUMERIC";
bool isDate = values.find("DATE") != std::string::npos;
bool isString = values == "STRING";
// For now, treat DATE and STRING as categorical (non-numeric)
// This provides basic compatibility while maintaining existing functionality
numeric_features[feature] = isNumeric;
}
}
std::vector<int> factorize(const std::string feature, const std::vector<std::string>& labels_t)
{
std::vector<int> yy;
states.at(feature).clear();
yy.reserve(labels_t.size());
std::map<std::string, int> labelMap;
int i = 0;
for (const std::string& label : labels_t) {
if (labelMap.find(label) == labelMap.end()) {
labelMap[label] = i++;
bool allDigits = std::all_of(label.begin(), label.end(), ::isdigit);
if (allDigits)
states[feature].push_back("Class " + label);
else
states[feature].push_back(label);
}
yy.push_back(labelMap[label]);
}
return yy;
}
void generateDataset(int labelIndex)
{
const size_t numSamples = lines.size();
const size_t numFeatures = attributes.size();
// Validate inputs
if (numSamples == 0) {
throw std::invalid_argument("No data samples found in file");
}
if (numFeatures == 0) {
throw std::invalid_argument("No feature attributes found");
}
if (labelIndex < 0) {
throw std::invalid_argument("Invalid label index: cannot be negative");
}
// Pre-allocate with feature-major layout: X[feature][sample]
X.assign(numFeatures, std::vector<float>(numSamples));
// Cache feature types for fast lookup during data processing
std::vector<bool> isNumericFeature(numFeatures);
for (size_t i = 0; i < numFeatures; ++i) {
isNumericFeature[i] = numeric_features.at(attributes[i].first);
}
// Temporary storage for categorical data per feature (only for non-numeric features)
std::vector<std::vector<std::string>> categoricalData(numFeatures);
for (size_t i = 0; i < numFeatures; ++i) {
if (!isNumericFeature[i]) {
categoricalData[i].reserve(numSamples);
}
}
std::vector<std::string> yy;
yy.reserve(numSamples);
// Parse each sample
for (size_t sampleIdx = 0; sampleIdx < numSamples; ++sampleIdx) {
const auto tokens = split(lines[sampleIdx], ',');
// Validate token count matches expected number (features + class)
const size_t expectedTokens = numFeatures + 1;
if (tokens.size() != expectedTokens) {
throw std::invalid_argument("Sample " + std::to_string(sampleIdx) + " has " + std::to_string(tokens.size()) + " tokens, expected " + std::to_string(expectedTokens));
}
int pos = 0;
int featureIdx = 0;
for (const auto& token : tokens) {
if (pos++ == labelIndex) {
if (token.empty()) {
throw std::invalid_argument("Empty class label at sample " + std::to_string(sampleIdx));
}
yy.push_back(token);
} else {
if (featureIdx >= static_cast<int>(numFeatures)) {
throw std::invalid_argument("Too many feature values at sample " + std::to_string(sampleIdx));
}
if (isNumericFeature[featureIdx]) {
// Parse numeric value with exception handling
try {
X[featureIdx][sampleIdx] = std::stof(token);
}
catch (const std::exception& e) {
const auto& featureName = attributes[featureIdx].first;
throw std::invalid_argument("Invalid numeric value '" + token + "' at sample " + std::to_string(sampleIdx) + ", feature " + featureName);
}
} else {
// Store categorical value temporarily
if (token.empty()) {
const auto& featureName = attributes[featureIdx].first;
throw std::invalid_argument("Empty categorical value at sample " + std::to_string(sampleIdx) + ", feature " + featureName);
}
categoricalData[featureIdx].push_back(token);
}
featureIdx++;
}
}
}
// Convert categorical features to numeric
for (size_t featureIdx = 0; featureIdx < numFeatures; ++featureIdx) {
if (!isNumericFeature[featureIdx]) {
const auto& featureName = attributes[featureIdx].first;
auto encodedValues = factorize(featureName, categoricalData[featureIdx]);
// Copy encoded values to X[feature][sample]
for (size_t sampleIdx = 0; sampleIdx < numSamples; ++sampleIdx) {
X[featureIdx][sampleIdx] = static_cast<float>(encodedValues[sampleIdx]);
}
}
}
y = factorize(className, yy);
}
void loadCommon(std::string fileName)
{
// Clear previous data
lines.clear();
attributes.clear();
states.clear();
numeric_features.clear();
// Validate file path for security
validateFilePath(fileName);
// Validate file size before processing
validateResourceLimits(fileName);
std::ifstream file(fileName);
if (!file.is_open()) {
throw std::invalid_argument("Unable to open file: " + fileName);
}
std::string line;
std::string keyword;
std::string attribute;
std::string type;
std::string type_w;
while (getline(file, line)) {
if (line.empty() || line[0] == '%' || line == "\r" || line == " ") {
continue;
}
// Skip sparse data format for now (lines starting with '{')
// Future enhancement: implement full sparse data support
if (!line.empty() && line[0] == '{') {
continue;
}
if (line.find("@attribute") != std::string::npos || line.find("@ATTRIBUTE") != std::string::npos) {
std::stringstream ss(line);
ss >> keyword >> attribute;
// Validate attribute name
if (attribute.empty()) {
throw std::invalid_argument("Empty attribute name in line: " + line);
}
// Check for duplicate attribute names
for (const auto& existing : attributes) {
if (existing.first == attribute) {
throw std::invalid_argument("Duplicate attribute name: " + attribute);
}
}
// Efficiently build type string
std::ostringstream typeStream;
while (ss >> type_w) {
if (typeStream.tellp() > 0) typeStream << " ";
typeStream << type_w;
}
type = typeStream.str();
// Validate type is not empty
if (type.empty()) {
throw std::invalid_argument("Empty attribute type for attribute: " + attribute);
}
attributes.emplace_back(trim(attribute), trim(type));
continue;
}
if (line[0] == '@') {
continue;
}
// More sophisticated missing value detection
// Skip lines with '?' not inside quoted strings
if (containsMissingValue(line)) {
continue;
}
lines.push_back(line);
}
// Final validation
if (attributes.empty()) {
throw std::invalid_argument("No attributes found in file");
}
if (lines.empty()) {
throw std::invalid_argument("No data samples found in file");
}
// Validate loaded data dimensions against limits
validateResourceLimits(fileName, lines.size(), attributes.size());
// Initialize states for all attributes
for (const auto& attribute : attributes) {
states[attribute.first] = std::vector<std::string>();
}
}
// Helper function for better missing value detection
bool containsMissingValue(const std::string& line)
{
bool inQuotes = false;
char quoteChar = '\0';
for (size_t i = 0; i < line.length(); ++i) {
char c = line[i];
if (!inQuotes && (c == '\'' || c == '\"')) {
inQuotes = true;
quoteChar = c;
} else if (inQuotes && c == quoteChar) {
inQuotes = false;
quoteChar = '\0';
} else if (!inQuotes && c == '?') {
// Found unquoted '?' - this is a missing value
return true;
}
}
return false;
}
// Static version of missing value detection for summary methods
static bool containsMissingValueStatic(const std::string& line)
{
bool inQuotes = false;
char quoteChar = '\0';
for (size_t i = 0; i < line.length(); ++i) {
char c = line[i];
if (!inQuotes && (c == '\'' || c == '\"')) {
inQuotes = true;
quoteChar = c;
} else if (inQuotes && c == quoteChar) {
inQuotes = false;
quoteChar = '\0';
} else if (!inQuotes && c == '?') {
// Found unquoted '?' - this is a missing value
return true;
}
}
return false;
}
// Common helper function to parse ARFF file attributes and count samples
static int parseArffFile(const std::string& fileName,
std::vector<std::pair<std::string, std::string>>& attributes,
std::set<std::string>& uniqueClasses,
size_t& sampleCount,
int classIndex = -1,
const std::string& classNameToFind = "") {
// Validate file path for security
validateFilePath(fileName);
std::ifstream file(fileName);
if (!file.is_open()) {
throw std::invalid_argument("Unable to open file: " + fileName);
}
std::string line;
attributes.clear();
uniqueClasses.clear();
sampleCount = 0;
// Parse header
while (getline(file, line)) {
if (line.empty() || line[0] == '%' || line == "\r" || line == " ") {
continue;
}
// Skip sparse data format for now (lines starting with '{')
if (!line.empty() && line[0] == '{') {
continue;
}
if (line.find("@attribute") != std::string::npos || line.find("@ATTRIBUTE") != std::string::npos) {
std::stringstream ss(line);
std::string keyword, attribute, type_w;
ss >> keyword >> attribute;
if (attribute.empty()) {
throw std::invalid_argument("Empty attribute name in line: " + line);
}
// Build type string
std::ostringstream typeStream;
while (ss >> type_w) {
if (typeStream.tellp() > 0) typeStream << " ";
typeStream << type_w;
}
std::string type = typeStream.str();
if (type.empty()) {
throw std::invalid_argument("Empty attribute type for attribute: " + attribute);
}
attributes.emplace_back(trim(attribute), trim(type));
continue;
}
if (line[0] == '@') {
continue;
}
// Start of data section
break;
}
if (attributes.empty()) {
throw std::invalid_argument("No attributes found in file");
}
// Find class index if class name is specified
int actualClassIndex = classIndex;
if (!classNameToFind.empty()) {
actualClassIndex = -1;
for (size_t i = 0; i < attributes.size(); ++i) {
if (attributes[i].first == classNameToFind) {
actualClassIndex = static_cast<int>(i);
break;
}
}
if (actualClassIndex == -1) {
throw std::invalid_argument("Class name '" + classNameToFind + "' not found in attributes");
}
}
// Count samples and collect unique class values
do {
if (!line.empty() && line[0] != '@' && line[0] != '%' && line[0] != '{' && !containsMissingValueStatic(line)) {
auto tokens = splitStatic(line, ',');
if (!tokens.empty()) {
std::string classValue;
if (actualClassIndex == -1) {
// Use last token (default behavior)
classValue = trim(tokens.back());
} else if (actualClassIndex == 0) {
// Use first token
classValue = trim(tokens.front());
} else if (actualClassIndex > 0 && static_cast<size_t>(actualClassIndex) < tokens.size()) {
// Use specific index
classValue = trim(tokens[actualClassIndex]);
}
if (!classValue.empty()) {
uniqueClasses.insert(classValue);
sampleCount++;
}
}
}
}
while (getline(file, line));
return actualClassIndex;
}
// Helper function for summary with classLast parameter
static ArffSummary summarizeFile(const std::string& fileName, bool classLast)
{
ArffSummary summary;
std::vector<std::pair<std::string, std::string>> attributes;
std::set<std::string> uniqueClasses;
size_t sampleCount = 0;
// Use common parsing function
parseArffFile(fileName, attributes, uniqueClasses, sampleCount, classLast ? -1 : 0);
// Determine class attribute
if (classLast) {
summary.className = attributes.back().first;
summary.classType = attributes.back().second;
attributes.pop_back();
} else {
summary.className = attributes.front().first;
summary.classType = attributes.front().second;
attributes.erase(attributes.begin());
}
summary.numFeatures = attributes.size();
// Copy feature information
for (const auto& attr : attributes) {
summary.featureInfo.emplace_back(attr.first, attr.second);
}
summary.numSamples = sampleCount;
summary.numClasses = uniqueClasses.size();
summary.classLabels.assign(uniqueClasses.begin(), uniqueClasses.end());
return summary;
}
// Helper function for summary with className parameter
static ArffSummary summarizeFile(const std::string& fileName, const std::string& className)
{
ArffSummary summary;
std::vector<std::pair<std::string, std::string>> attributes;
std::set<std::string> uniqueClasses;
size_t sampleCount = 0;
int classIndex = -1;
// Use common parsing function to find class by name
classIndex = parseArffFile(fileName, attributes, uniqueClasses, sampleCount, -1, className);
// Set class information from the found attribute
summary.className = attributes[classIndex].first;
summary.classType = attributes[classIndex].second;
// Remove class attribute from features
attributes.erase(attributes.begin() + classIndex);
summary.numFeatures = attributes.size();
// Copy feature information
for (const auto& attr : attributes) {
summary.featureInfo.emplace_back(attr.first, attr.second);
}
summary.numSamples = sampleCount;
summary.numClasses = uniqueClasses.size();
summary.classLabels.assign(uniqueClasses.begin(), uniqueClasses.end());
return summary;
}
// Static helper function for split (needed by summarizeFile)
static std::vector<std::string> splitStatic(const std::string& text, char delimiter)
{
std::vector<std::string> result;
std::stringstream ss(text);
std::string token;
while (std::getline(ss, token, delimiter)) {
result.push_back(trim(token));
}
return result;
}
};
#endif