Skip to content
Navigation Menu
{{ message }}
-
Notifications
You must be signed in to change notification settings - Fork 223
Expand file tree
/
Copy pathdataset.cpp
More file actions
283 lines (249 loc) · 9.09 KB
/
Copy pathdataset.cpp
File metadata and controls
283 lines (249 loc) · 9.09 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
//
// Created by jiashuai on 17-9-17.
//
#include "thundersvm/dataset.h"
#include <omp.h>
using std::fstream;
using std::stringstream;
DataSet::DataSet() : total_count_(0), n_features_(0) {}
DataSet::DataSet(const DataSet::node2d &instances, int n_features, const vector<float_type> &y) :
instances_(instances), n_features_(n_features), y_(y), total_count_(instances_.size()) {}
inline char *findlastline(char *ptr, char *begin) {
while (ptr != begin && *ptr != '\n') --ptr;
return ptr;
}
void DataSet::load_from_file(string file_name) {
LOG(INFO)<<"loading dataset from file \""<<file_name<<"\"";
y_.clear();
instances_.clear();
total_count_ = 0;
n_features_ = 0;
std::ifstream ifs(file_name, std::ifstream::binary);
CHECK(ifs.is_open()) << "file " << file_name << " not found";
int buffer_size = 16 << 20; //16MB
char *buffer = (char *)malloc(buffer_size);
const int nthread = omp_get_max_threads();
while (ifs) {
char *head = buffer;
ifs.read(buffer, buffer_size);
size_t size = ifs.gcount();
vector<vector<float_type>> y_thread(nthread);
vector<node2d> instances_thread(nthread);
vector<int> local_feature(nthread, 0);
#pragma omp parallel num_threads(nthread)
{
//get working area of this thread
int tid = omp_get_thread_num();
size_t nstep = (size + nthread - 1) / nthread;
size_t sbegin = min(tid * nstep, size);
size_t send = min((tid + 1) * nstep, size);
char *pbegin = findlastline(head + sbegin, head);
char *pend = findlastline(head + send, head);
//move stream start position to the end of last line
if (tid == nthread - 1) ifs.seekg(pend - head - send + 1, std::ios_base::cur);
//read instances line by line
char *lbegin = pbegin;
char *lend = lbegin;
while (lend != pend) {
//get one line
lend = lbegin + 1;
while (lend != pend && *lend != '\n') {
++lend;
}
string line(lbegin, lend);
stringstream ss(line);
//read label of an instance
y_thread[tid].emplace_back();
ss >> y_thread[tid].back();
//read features of an instance
instances_thread[tid].emplace_back();
string tuple;
while (ss >> tuple) {
int i;
float v;
CHECK_EQ(sscanf(tuple.c_str(), "%d:%f", &i, &v), 2) << "read error, using [index]:[value] format";
instances_thread[tid].back().emplace_back(i, v);
if (i > local_feature[tid]) local_feature[tid] = i;
};
//read next instance
lbegin = lend;
}
}
for (int i = 0; i < nthread; i++) {
if (local_feature[i] > n_features_)
n_features_ = local_feature[i];
total_count_ += instances_thread[i].size();
}
for (int i = 0; i < nthread; i++) {
this->y_.insert(y_.end(), y_thread[i].begin(), y_thread[i].end());
this->instances_.insert(instances_.end(), instances_thread[i].begin(), instances_thread[i].end());
}
}
free(buffer);
LOG(INFO)<<"#instances = "<<this->n_instances()<<", #features = "<<this->n_features();
}
void DataSet::load_from_python(float *y, char **x, int len) {
y_.clear();
instances_.clear();
total_count_ = 0;
n_features_ = 0;
for (int i = 0; i < len; i++) {
int ind;
float v;
string line = x[i];
stringstream ss(line);
y_.push_back(y[i]);
instances_.emplace_back();
string tuple;
while (ss >> tuple) {
CHECK_EQ(sscanf(tuple.c_str(), "%d:%f", &ind, &v), 2) << "read error, using [index]:[value] format";
instances_[total_count_].emplace_back(ind, v);
if (ind > n_features_) n_features_ = ind;
};
total_count_++;
}
}
void DataSet::load_from_sparse(int row_size, float* val, int* row_ptr, int* col_ptr, float* label) {
y_.clear();
instances_.clear();
total_count_ = 0;
n_features_ = 0;
for(int i = 0; i < row_size; i++){
int ind;
float v;
if(label != NULL)
y_.push_back(label[i]);
instances_.emplace_back();
for(int i = row_ptr[total_count_]; i < row_ptr[total_count_ + 1]; i++){
ind = col_ptr[i];
v = val[i];
instances_[total_count_].emplace_back(ind, v);
if(ind > n_features_) n_features_ = ind;
}
total_count_++;
}
n_features_++;
LOG(INFO)<<"#instances = "<<this->n_instances()<<", #features = "<<this->n_features();
}
void DataSet::load_from_dense(int row_size, int features, float* data, float* label){
y_.clear();
instances_.clear();
total_count_ = 0;
n_features_ = 0;
int off = 0;
for(int i = 0; i < row_size; i++){
int ind;
float v;
if(label != NULL)
y_.push_back(label[i]);
instances_.emplace_back();
for(int j = 0; j < features; j++){
ind = j;
v = data[off];
off++;
instances_[total_count_].emplace_back(ind, v);
}
total_count_++;
}
n_features_ = features;
LOG(INFO)<<"#instances = "<<this->n_instances()<<", #features = "<<this->n_features();
}
const vector<int> &DataSet::count() const {//return the number of instances of each class
return count_;
}
const vector<int> &DataSet::start() const {
return start_;
}
size_t DataSet::n_classes() const {
return start_.size();
}
const vector<int> &DataSet::label() const {
return label_;
}
void DataSet::group_classes(bool classification) {
if (classification) {
start_.clear();
count_.clear();
label_.clear();
perm_.clear();
vector<int> dataLabel(y_.size());//temporary labels of all the instances
//get the class labels; count the number of instances in each class.
for (int i = 0; i < y_.size(); ++i) {
int j;
for (j = 0; j < label_.size(); ++j) {
if (y_[i] == label_[j]) {
count_[j]++;
break;
}
}
dataLabel[i] = j;
//if the label is unseen, add it to label vector.
if (j == label_.size()) {
//real to int conversion is safe, because group_classes only used in classification
label_.push_back(int(y_[i]));
count_.push_back(1);
}
}
//logically put instances of the same class consecutively.
start_.push_back(0);
for (int i = 1; i < count_.size(); ++i) {
start_.push_back(start_[i - 1] + count_[i - 1]);
}
vector<int> start_copy(start_);
perm_ = vector<int>(y_.size());//index of each instance in the original array
for (int i = 0; i < y_.size(); ++i) {
perm_[start_copy[dataLabel[i]]] = i;
start_copy[dataLabel[i]]++;
}
} else {
for (int i = 0; i < instances_.size(); ++i) {
perm_.push_back(i);
}
start_.push_back(0);
count_.push_back(instances_.size());
}
}
size_t DataSet::n_instances() const {//return the total number of instances
return total_count_;
}
size_t DataSet::n_features() const {
return n_features_;
}
const DataSet::node2d &DataSet::instances() const {//return all the instances
return instances_;
}
const DataSet::node2d DataSet::instances(int y_i) const {//return instances of a given class
int si = start_[y_i];
int ci = count_[y_i];
node2d one_class_ins;
for (int i = si; i < si + ci; ++i) {
one_class_ins.push_back(instances_[perm_[i]]);
}
return one_class_ins;
}
const DataSet::node2d DataSet::instances(int y_i, int y_j) const {//return instances of two classes
node2d two_class_ins;
node2d i_ins = instances(y_i);
node2d j_ins = instances(y_j);
two_class_ins.insert(two_class_ins.end(), i_ins.begin(), i_ins.end());
two_class_ins.insert(two_class_ins.end(), j_ins.begin(), j_ins.end());
return two_class_ins;
}
const vector<int> DataSet::original_index() const {//index of each instance in the original array
return perm_;
}
const vector<int> DataSet::original_index(int y_i) const {//index of each instance in the original array for one class
return vector<int>(perm_.begin() + start_[y_i], perm_.begin() + start_[y_i] + count_[y_i]);
}
const vector<int>
DataSet::original_index(int y_i, int y_j) const {//index of each instance in the original array for two class
vector<int> two_class_idx;
vector<int> i_idx = original_index(y_i);
vector<int> j_idx = original_index(y_j);
two_class_idx.insert(two_class_idx.end(), i_idx.begin(), i_idx.end());
two_class_idx.insert(two_class_idx.end(), j_idx.begin(), j_idx.end());
return two_class_idx;
}
const vector<float_type> &DataSet::y() const {
return y_;
}
You can’t perform that action at this time.