-
Notifications
You must be signed in to change notification settings - Fork 1
/
cosmo-benchmark.cpp
333 lines (293 loc) · 12.4 KB
/
cosmo-benchmark.cpp
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
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
#include <iostream>
#include <fstream>
#include <algorithm>
#include <chrono>
#include <libgen.h> // basename
#include <sys/mman.h> // mlockall
#include "tclap/CmdLine.h"
#include <sdsl/bit_vectors.hpp>
#include <sdsl/wavelet_trees.hpp>
#include <sdsl/wt_algorithm.hpp>
#include <boost/random.hpp>
#include <boost/generator_iterator.hpp>
#include <boost/iterator/function_input_iterator.hpp>
#include <boost/filesystem/path.hpp>
#include <boost/filesystem.hpp>
//#include <boost_iterator/zip_iterator.hpp>
#include "io.hpp"
#include "debruijn_graph.hpp"
#include "debruijn_graph_shifted.hpp"
#include "debruijn_hypergraph.hpp"
#include "algorithm.hpp"
#include "wt_algorithm.hpp"
using namespace std;
using namespace sdsl;
string graph_extension = ".dbg";
string contig_extension = ".fasta";
struct parameters_t {
std::string input_filename = "";
std::string output_prefix = "";
};
void parse_arguments(int argc, char **argv, parameters_t & params);
void parse_arguments(int argc, char **argv, parameters_t & params)
{
TCLAP::CmdLine cmd("Cosmo Copyright (c) Alex Bowe (alexbowe.com) 2014", ' ', VERSION);
TCLAP::UnlabeledValueArg<std::string> input_filename_arg("input",
".dbg file (output from cosmo-build).", true, "", "input_file", cmd);
string output_short_form = "output_prefix";
TCLAP::ValueArg<std::string> output_prefix_arg("o", "output_prefix",
"Output prefix. Contigs will be written to [" + output_short_form + "]" + contig_extension + ". " +
"Default prefix: basename(input_file).", false, "", output_short_form, cmd);
cmd.parse( argc, argv );
// -d flag for decompression to original kmer biz
params.input_filename = input_filename_arg.getValue();
params.output_prefix = output_prefix_arg.getValue();
}
int main(int argc, char* argv[]) {
parameters_t p;
parse_arguments(argc, argv, p);
auto base_path = boost::filesystem::path(p.input_filename).parent_path().string();
auto base_name = base_path + "/" + boost::filesystem::path(p.input_filename).stem().string();
COSMO_LOG(debug) << base_name;
// TO LOAD:
debruijn_graph_shifted<> g;
load_from_file(g, p.input_filename);
auto dbg_size = size_in_mega_bytes(g);
cerr << "k : " << g.k << endl;
cerr << "num_nodes() : " << g.num_nodes() << endl;
cerr << "num_edges() : " << g.num_edges() << endl;
cerr << "W size : " << size_in_mega_bytes(g.m_edges) << " MB" << endl;
cerr << "L size : " << size_in_mega_bytes(g.m_node_flags) << " MB" << endl;
cerr << "DBG size : " << dbg_size << " MB" << endl;
cerr << "Bits per edge : " << bits_per_element(g) << " Bits" << endl;
#ifdef VAR_ORDER
wt_int<rrr_vector<63>> lcs;
load_from_file(lcs, base_name + ".lcs");
auto lcs_size = size_in_mega_bytes(lcs);
auto total_size = dbg_size + lcs_size;
cerr << "LCS size : " << lcs_size << " MB" << endl;
cerr << "LCS bits/edge : " << bits_per_element(lcs) << " Bits" << endl;
cerr << "Total size : " << total_size << " MB" << endl;
typedef debruijn_hypergraph<> dbh;
typedef dbh::node_type node_type;
dbh h(g, lcs);
#endif
/*
cerr << "symbol ends:" << endl;
for (size_t i = 0; i<5; ++i) {
cerr << i << ": " << g.m_symbol_ends[i] << endl;
}
*/
int num_queries = 2e4;
size_t min_k = 8; // this could be 0, but it affects longer too much
size_t max_k = g.k-2; // K is the edge length, and K-1 node lenght.
// Time will be identical if we measure K-1
cerr << "Generating " << num_queries << " random queries" << endl;
// set up RNGs
typedef boost::mt19937 rng_type;
rng_type rng(time(0));
boost::uniform_int<size_t> node_distribution(0, g.num_nodes()-1); // Randomly choose nodes from the graph
boost::uniform_int<size_t> k_distribution(min_k, max_k);
boost::uniform_int<size_t> symbol_distribution(1, 4); // A C G T in our WT (ignoring $=0)
boost::variate_generator<rng_type, boost::uniform_int<size_t>> random_node(rng, node_distribution);
boost::variate_generator<rng_type, boost::uniform_int<size_t>> random_k(rng, k_distribution);
boost::variate_generator<rng_type, boost::uniform_int<size_t>> random_symbol(rng, symbol_distribution);
vector<size_t> query_nodes(boost::make_function_input_iterator(random_node,0),
boost::make_function_input_iterator(random_node,num_queries));
vector<size_t> query_syms(boost::make_function_input_iterator(random_symbol,0),
boost::make_function_input_iterator(random_symbol,num_queries));
#ifdef VAR_ORDER
vector<node_type> query_varnodes;
for (auto u:query_nodes) {
query_varnodes.push_back(h.get_node(u));
}
// Randomly change the order of random nodes
// TODO: WHY is this in this range?
// -2 because we need to remove the last character
auto random_low_bound_k = [&](size_t low) {
assert(low >= min_k);
return boost::uniform_int<size_t>(low, g.k-1)(rng);
};
auto random_high_bound_k = [&](size_t high) {
assert(high <= g.k-1);
return boost::uniform_int<size_t>(min_k, high)(rng);
};
// Make random nodes that have an order that we can increase a certain amount
const vector<size_t> query_order_deltas{1,2,4,8}; // Try all ks?
size_t num_deltas = query_order_deltas.size();
vector<vector<node_type>> shorter_query_varnodes(query_order_deltas.size());
vector<vector<node_type>> longer_query_varnodes(query_order_deltas.size());
for (size_t i=0; i<query_order_deltas.size();++i) {
shorter_query_varnodes[i].reserve(num_queries);//= vector<node_type>(num_queries);
longer_query_varnodes[i].reserve(num_queries);//= vector<node_type>(num_queries);
for (size_t node_idx=0; node_idx < (size_t)num_queries; ++node_idx) {
auto delta = query_order_deltas[i];
auto node = query_varnodes[node_idx];
// Nodes that are of at least a certain k so that we can shorten them later
// Note that we dont adhere to the max_k since these operations
// arent possible on the standard dbg
auto shorter_k = random_low_bound_k(min_k+delta);
assert(min_k+delta <= shorter_k && shorter_k < g.k);
auto longer_k = random_high_bound_k(g.k-1-delta);
node_type new_shorter_node;
if (shorter_k == g.k-1) new_shorter_node = node;
else new_shorter_node = h.shorter(node, shorter_k);
shorter_query_varnodes[i].push_back(new_shorter_node);
longer_query_varnodes[i].push_back(h.shorter(node, longer_k));
}
}
// Convert to variable order nodes
// Need: random lows for longer of each delta
// random highs for shorter of each delta
// random nodes for maxlen, maxlen*, backward, forward, etc
vector<size_t> query_ks(boost::make_function_input_iterator(random_k,0),
boost::make_function_input_iterator(random_k,num_queries));
vector<size_t> maxlen_syms(boost::make_function_input_iterator(random_symbol,0),
boost::make_function_input_iterator(random_symbol,num_queries));
// set shorter for each one that has a shorter k
for (int i=0;i<num_queries;i++) {
if (query_ks[i] == g.k-1) continue;
query_varnodes[i] = h.shorter(query_varnodes[i], query_ks[i]);
}
#else
vector<debruijn_graph_shifted<>::node_type> query_rangenodes;
//transform(query_nodes.begin(), query_nodes.end(), query_varnodes.begin(),[&](size_t v){ return h.get_node(v); });
for (auto u:query_nodes) {
query_rangenodes.push_back(g.get_node(u));
}
#endif
mlockall(MCL_CURRENT);
typedef chrono::nanoseconds unit;
string unit_s = " ns";
#ifndef VAR_ORDER // standard dbg
// backward
auto t1 = chrono::high_resolution_clock::now();
for (auto v : query_rangenodes) {
g.all_preds(v);
}
auto t2 = chrono::high_resolution_clock::now();
auto dur = chrono::duration_cast<unit>(t2-t1).count();
//cerr << "backward total : " << dur << " ns" <<endl;
cerr << "backward : " << (double)dur/num_queries << unit_s <<endl;
// forward
t1 = chrono::high_resolution_clock::now();
for (size_t i=0;i<(size_t)num_queries;i++) {
auto v = query_rangenodes[i];
auto x = query_syms[i];
g.interval_node_outgoing(v, x);
}
t2 = chrono::high_resolution_clock::now();
dur = chrono::duration_cast<unit>(t2-t1).count();
//cerr << "forward total : " << dur << " ns" <<endl;
cerr << "forward : " << (double)dur/num_queries << unit_s <<endl;
// last_char
t1 = chrono::high_resolution_clock::now();
for (auto v : query_rangenodes) { g.lastchar(v); }
t2 = chrono::high_resolution_clock::now();
dur = chrono::duration_cast<unit>(t2-t1).count();
//cerr << "lastchar total : " << dur << " ns" <<endl;
cerr << "lastchar : " << (double)dur/num_queries << unit_s <<endl;
#else
auto t1 = chrono::high_resolution_clock::now();
// backward
for (auto v : query_varnodes) {
h.backward(v);
}
auto t2 = chrono::high_resolution_clock::now();
auto dur = chrono::duration_cast<unit>(t2-t1).count();
//cerr << "backward total : " << dur << " ns" <<endl;
cerr << "backward : " << (double)dur/num_queries << unit_s <<endl;
// forward
t1 = chrono::high_resolution_clock::now();
for (size_t i=0;i<(size_t)num_queries;i++) {
auto v = query_varnodes[i];
auto x = query_syms[i];
//cerr << "("<< get<0>(v) << ", " << get<1>(v) << ", "<< get<2>(v) << ") " << (int)x << endl;
h.outgoing(v, x);
}
t2 = chrono::high_resolution_clock::now();
dur = chrono::duration_cast<unit>(t2-t1).count();
//cerr << "forward total : " << dur << " ns" <<endl;
cerr << "forward : " << (double)dur/num_queries << unit_s <<endl;
// last_char
t1 = chrono::high_resolution_clock::now();
for (auto v : query_varnodes) { h.lastchar(v); }
t2 = chrono::high_resolution_clock::now();
dur = chrono::duration_cast<unit>(t2-t1).count();
//cerr << "lastchar total : " << dur << " ns" <<endl;
cerr << "lastchar : " << (double)dur/num_queries << unit_s <<endl;
// shorter
for (size_t i=0; i<num_deltas; ++i) {
size_t delta = query_order_deltas[i];
t1 = chrono::high_resolution_clock::now();
for (size_t idx=0;idx<(size_t)num_queries;idx++) {
auto v = shorter_query_varnodes[i][idx];
size_t k = get<2>(v);
size_t new_k = k-delta;
assert(new_k >= min_k);
h.shorter(v, new_k);
}
t2 = chrono::high_resolution_clock::now();
dur = chrono::duration_cast<unit>(t2-t1).count();
cerr << "shorter_" << delta << " : " << (double)dur/(num_queries) << unit_s <<endl;
}
// longer
for (size_t i=0; i<num_deltas; ++i) {
size_t delta = query_order_deltas[i];
t1 = chrono::high_resolution_clock::now();
size_t total = 0;
for (size_t idx=0;idx<(size_t)num_queries;idx++) {
auto v = longer_query_varnodes[i][idx];
size_t k = get<2>(v);
size_t new_k = k+delta;
assert(new_k >= min_k);
assert(new_k < g.k);
auto result = h.longer(v, new_k);
total += result.size();
}
t2 = chrono::high_resolution_clock::now();
dur = chrono::duration_cast<unit>(t2-t1).count();
cerr << "longer_" << delta << " : " << (double)dur/(num_queries) << unit_s <<endl;
cerr << "per node : " << (double)dur/(total) << unit_s << " (" << total << ")" << endl;
}
/*
size_t skipped = 0;
for (size_t k : {1,2,4,8}) {
t1 = chrono::high_resolution_clock::now();
for (size_t i=0;i<(size_t)num_queries;i++) {
auto v = query_varnodes[i];
auto k = higher_ks[i];
if (get<2>(v) > k) {
skipped++;
continue;
}
//cout << get<0>(v) << ", " << get<1>(v) << ", " << get<2>(v) << endl;
h.longer(v, k);
}
t2 = chrono::high_resolution_clock::now();
dur = chrono::duration_cast<unit>(t2-t1).count();
//cerr << "longer(v,+"<<k<<") total : " << dur << " ns" <<endl;
cerr << "longer"<<" mean : " << (double)dur/(num_queries-skipped) << unit_s <<endl;
//}
*/
// maxlen with symbol
t1 = chrono::high_resolution_clock::now();
for (size_t i=0;i<(size_t)num_queries;i++) {
auto v = query_varnodes[i];
auto x = maxlen_syms[i];
h.maxlen(v, x);
}
t2 = chrono::high_resolution_clock::now();
dur = chrono::duration_cast<unit>(t2-t1).count();
//cerr << "maxlen(v,c) total : " << dur << " ns" <<endl;
cerr << "maxlen : " << (double)dur/num_queries << unit_s <<endl;
// Regular maxlen
t1 = chrono::high_resolution_clock::now();
for (auto v : query_varnodes) { h.maxlen(v); }
t2 = chrono::high_resolution_clock::now();
dur = chrono::duration_cast<unit>(t2-t1).count();
//cerr << "maxlen(v,*) total : " << dur << " ns" <<endl;
cerr << "maxlen* : " << (double)dur/num_queries << unit_s <<endl;
#endif
munlockall();
}