NVBIO
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Macros Groups Pages
reads.h
Go to the documentation of this file.
1 /*
2  * nvbio
3  * Copyright (c) 2011-2014, NVIDIA CORPORATION. All rights reserved.
4  *
5  * Redistribution and use in source and binary forms, with or without
6  * modification, are permitted provided that the following conditions are met:
7  * * Redistributions of source code must retain the above copyright
8  * notice, this list of conditions and the following disclaimer.
9  * * Redistributions in binary form must reproduce the above copyright
10  * notice, this list of conditions and the following disclaimer in the
11  * documentation and/or other materials provided with the distribution.
12  * * Neither the name of the NVIDIA CORPORATION nor the
13  * names of its contributors may be used to endorse or promote products
14  * derived from this software without specific prior written permission.
15  *
16  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
17  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
18  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
19  * DISCLAIMED. IN NO EVENT SHALL NVIDIA CORPORATION BE LIABLE FOR ANY
20  * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
21  * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
22  * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
23  * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
24  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
25  * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
26  */
27 
28 #pragma once
29 
30 #include <nvbio/basic/strided_iterator.h>
31 #include <nvbio/basic/packedstream.h>
32 #include <nvbio/basic/vector_view.h>
33 #include <nvbio/strings/string_set.h>
34 #include <stdio.h>
35 #include <stdlib.h>
36 #include <vector>
37 
38 namespace nvbio {
39 namespace io {
40 
58 
61 
73 
74 // describes the quality encoding for a given read file
75 enum QualityEncoding
76 {
77  // phred quality
78  Phred = 0,
79  // phred quality + 33
80  Phred33 = 1,
81  // phred quality + 64
82  Phred64 = 2,
83  Solexa = 3,
84 };
85 
86 // a set of flags describing the types of supported read strands
87 enum ReadEncoding
88 {
89  FORWARD = 0x0001,
90  REVERSE = 0x0002,
91  FORWARD_COMPLEMENT = 0x0004,
92  REVERSE_COMPLEMENT = 0x0008,
93 };
94 
95 // how mates of a paired-end read are encoded
96 // F = forward, R = reverse
97 enum PairedEndPolicy
98 {
99  PE_POLICY_FF = 0,
100  PE_POLICY_FR = 1,
101  PE_POLICY_RF = 2,
102  PE_POLICY_RR = 3,
103 };
104 
116 template <
117  typename IndexIterator,
118  typename ReadStorageIterator,
119  typename QualStorageIterator,
120  typename NameStorageIterator>
121 struct ReadDataView
122 {
123  typedef IndexIterator index_iterator;
124  typedef typename to_const<index_iterator>::type const_index_iterator;
125 
126  typedef ReadStorageIterator read_storage_iterator;
127  typedef typename to_const<read_storage_iterator>::type const_read_storage_iterator;
128 
129  typedef QualStorageIterator qual_storage_iterator;
130  typedef typename to_const<qual_storage_iterator>::type const_qual_storage_iterator;
131 
132  typedef NameStorageIterator name_storage_iterator;
133  typedef typename to_const<name_storage_iterator>::type const_name_storage_iterator;
134 
135  // symbol size for reads
136  static const uint32 READ_BITS = 4;
137  // big endian?
138  static const bool HI_BITS = false; // deprecated
139  // big endian?
140  static const bool READ_BIG_ENDIAN = false;
141  // symbols per word
142  static const uint32 READ_SYMBOLS_PER_WORD = (4*sizeof(uint32))/READ_BITS;
143 
144  typedef PackedStream<
145  read_storage_iterator,uint8,READ_BITS,READ_BIG_ENDIAN> read_stream_type;
146  typedef PackedStream<
147  const_read_storage_iterator,uint8,READ_BITS,READ_BIG_ENDIAN> const_read_stream_type;
148 
149  typedef typename read_stream_type::iterator read_stream_iterator;
150  typedef typename const_read_stream_type::iterator const_read_stream_iterator;
151 
152  typedef vector_view<read_stream_iterator> read_string;
153  typedef vector_view<const_read_stream_iterator> const_read_string;
154 
155  typedef ConcatenatedStringSet<
156  read_stream_iterator,
157  index_iterator> read_string_set_type;
158 
159  typedef ConcatenatedStringSet<
160  const_read_stream_iterator,
161  const_index_iterator> const_read_string_set_type;
162 
163  typedef ConcatenatedStringSet<
164  qual_storage_iterator,
165  index_iterator> qual_string_set_type;
166 
167  typedef ConcatenatedStringSet<
168  const_qual_storage_iterator,
169  const_index_iterator> const_qual_string_set_type;
170 
171  typedef ConcatenatedStringSet<
172  name_storage_iterator,
173  index_iterator> name_string_set_type;
174 
175  typedef ConcatenatedStringSet<
176  const_name_storage_iterator,
177  const_index_iterator> const_name_string_set_type;
178 
181  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE
182  ReadDataView()
183  : m_n_reads(0),
184  m_name_stream_len(0),
185  m_read_stream_len(0),
186  m_read_stream_words(0),
187  m_min_read_len(uint32(-1)),
188  m_max_read_len(0),
189  m_avg_read_len(0)
190  {};
191 
194  template <
195  typename InIndexIterator,
196  typename InReadIterator,
197  typename InQualIterator,
198  typename InNameIterator>
199  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE
200  ReadDataView(const ReadDataView<InIndexIterator,InReadIterator,InQualIterator,InNameIterator>& in)
201  : m_n_reads (in.m_n_reads),
202  m_name_stream (NameStorageIterator(in.m_name_stream)),
203  m_name_stream_len (in.m_name_stream_len),
204  m_name_index (IndexIterator(in.m_name_index)),
205  m_read_stream (ReadStorageIterator(in.m_read_stream)),
206  m_read_stream_len (in.m_read_stream_len),
207  m_read_stream_words (in.m_read_stream_words),
208  m_read_index (IndexIterator(in.m_read_index)),
209  m_qual_stream (QualStorageIterator(in.m_qual_stream)),
210  m_min_read_len (in.m_min_read_len),
211  m_max_read_len (in.m_max_read_len),
212  m_avg_read_len (in.m_avg_read_len)
213  {}
214 
215  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE index_iterator name_index() { return m_name_index; }
216  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE index_iterator read_index() { return m_read_index; }
217  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE name_storage_iterator name_stream() { return m_name_stream; }
218  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE read_storage_iterator read_stream_storage() { return m_read_stream; }
219  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE read_stream_type read_stream() { return read_stream_type( m_read_stream ); }
220  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE qual_storage_iterator qual_stream() { return m_qual_stream; }
221 
222  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_index_iterator const_name_index() const { return m_name_index; }
223  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_index_iterator const_read_index() const { return m_read_index; }
224  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_name_storage_iterator const_name_stream() const { return m_name_stream; }
225  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_read_storage_iterator const_read_stream_storage() const { return m_read_stream; }
226  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_read_stream_type const_read_stream() const { return const_read_stream_type( m_read_stream ); }
227  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_qual_storage_iterator const_qual_stream() const { return m_qual_stream; }
228 
229  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_index_iterator name_index() const { return const_name_index(); }
230  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_index_iterator read_index() const { return const_read_index(); }
231  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_name_storage_iterator name_stream() const { return const_name_stream(); }
232  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_read_storage_iterator read_stream_storage() const { return const_read_stream_storage(); }
233  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_read_stream_type read_stream() const { return const_read_stream(); }
234  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_qual_storage_iterator qual_stream() const { return const_qual_stream(); }
235 
236  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE uint32 size() const { return m_n_reads; }
237  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE uint32 bps() const { return m_read_stream_len; }
238  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE uint32 words() const { return m_read_stream_words; }
239  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE uint32 name_stream_len() const { return m_name_stream_len; }
240  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE uint32 max_read_len() const { return m_max_read_len; }
241  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE uint32 min_read_len() const { return m_min_read_len; }
242  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE uint32 avg_read_len() const { return m_avg_read_len; }
243  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE uint2 get_range(const uint32 i) const { return make_uint2(m_read_index[i],m_read_index[i+1]); }
244 
247  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE read_string_set_type read_string_set()
248  {
249  return read_string_set_type(
250  size(),
251  read_stream().begin(),
252  read_index() );
253  }
254 
257  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_read_string_set_type read_string_set() const
258  {
259  return const_read_string_set_type(
260  size(),
261  read_stream().begin(),
262  read_index() );
263  }
264 
267  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_read_string_set_type const_read_string_set() const
268  {
269  return const_read_string_set_type(
270  size(),
271  read_stream().begin(),
272  read_index() );
273  }
274 
277  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE read_string get_read(const uint32 i)
278  {
279  const uint2 read_range = get_range( i );
280  return read_string( read_range.y - read_range.x, read_stream().begin() + read_range.x );
281  }
282 
285  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE read_string get_read(const uint32 i) const
286  {
287  const uint2 read_range = get_range( i );
288  return const_read_string( read_range.y - read_range.x, read_stream().begin() + read_range.x );
289  }
290 
293  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE qual_string_set_type qual_string_set()
294  {
295  return qual_string_set_type(
296  size(),
297  qual_stream(),
298  read_index() );
299  }
300 
303  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_qual_string_set_type qual_string_set() const
304  {
305  return const_qual_string_set_type(
306  size(),
307  qual_stream(),
308  read_index() );
309  }
310 
313  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_qual_string_set_type const_qual_string_set() const
314  {
315  return const_qual_string_set_type(
316  size(),
317  qual_stream(),
318  read_index() );
319  }
320 
323  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE name_string_set_type name_string_set()
324  {
325  return name_string_set_type(
326  size(),
327  name_stream(),
328  name_index() );
329  }
330 
333  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_name_string_set_type name_string_set() const
334  {
335  return const_name_string_set_type(
336  size(),
337  name_stream(),
338  name_index() );
339  }
340 
343  NVBIO_HOST_DEVICE NVBIO_FORCEINLINE const_name_string_set_type const_name_string_set() const
344  {
345  return const_name_string_set_type(
346  size(),
347  name_stream(),
348  name_index() );
349  }
350 
351 public:
352  // number of reads in this struct
353  uint32 m_n_reads;
354 
355  // a pointer to a buffer containing the names of all the reads in this batch
356  name_storage_iterator m_name_stream;
357  // the length (in bytes) of the name_stream buffer
358  uint32 m_name_stream_len;
359  // an array of uint32 with the byte indices of the starting locations of each name in name_stream
360  index_iterator m_name_index;
361 
362  // a pointer to a buffer containing the read data
363  // note that this could point at either host or device memory
364  read_storage_iterator m_read_stream;
365  // the length of read_stream in base pairs
366  uint32 m_read_stream_len;
367  // the number of words in read_stream
368  uint32 m_read_stream_words;
369  // an array of uint32 with the indices of the starting locations of each read in read_stream (in base pairs)
370  index_iterator m_read_index;
371 
372  // a pointer to a buffer containing quality data
373  // (the indices in m_read_index are also valid for this buffer)
374  qual_storage_iterator m_qual_stream;
375 
376  // statistics on the reads: minimum size, maximum size, average size
377  uint32 m_min_read_len;
378  uint32 m_max_read_len;
379  uint32 m_avg_read_len;
380 };
381 
385 typedef ReadDataView<uint32*,uint32*,char*,char*> ReadDataCore;
386 
391 struct ReadData : public ReadDataCore
392 {
393  typedef ReadDataView<uint32*,uint32*,char*,char*> plain_view_type;
394  typedef ReadDataView<const uint32*,const uint32*,const char*,const char*> const_plain_view_type;
395 
398  ReadData() : ReadDataCore()
399  {
400  m_name_stream = NULL;
401  m_name_index = NULL;
402  m_read_stream = NULL;
403  m_read_index = NULL;
404  m_qual_stream = NULL;
405  }
406 
409  virtual ~ReadData() {}
410 };
411 
415 struct ReadDataRAM : public ReadData
416 {
419  enum StrandOp
420  {
421  NO_OP = 0x0000,
422  REVERSE_OP = 0x0001,
423  COMPLEMENT_OP = 0x0002,
424  REVERSE_COMPLEMENT_OP = 0x0003,
425  };
426 
427  ReadDataRAM();
428 
431  void reserve(const uint32 n_reads, const uint32 n_bps);
432 
443  void push_back(uint32 read_len,
444  const char* name,
445  const uint8* base_pairs,
446  const uint8* quality,
447  const QualityEncoding quality_encoding,
448  const uint32 truncate_read_len,
449  const StrandOp conversion_flags);
450 
453  void end_batch(void);
454 
455  std::vector<uint32> m_read_vec;
456  std::vector<uint32> m_read_index_vec;
457  std::vector<char> m_qual_vec;
458  std::vector<char> m_name_vec;
459  std::vector<uint32> m_name_index_vec;
460 };
461 
465 struct ReadDataDevice : public ReadData
466 {
467  enum {
468  READS = 0x01,
469  QUALS = 0x02,
470  };
471 
474  ReadDataDevice(const ReadData& host_data, const uint32 flags = READS);
475 
478  ~ReadDataDevice();
479 
480  uint64 allocated() const { return m_allocated; }
481 
482 private:
483  uint64 m_allocated;
484 };
485 
486 typedef ReadDataRAM ReadDataHost;
487 typedef ReadDataDevice ReadDataCUDA;
488 
493 struct ReadDataStream
494 {
495  ReadDataStream(uint32 truncate_read_len = uint32(-1))
496  : m_truncate_read_len(truncate_read_len)
497  {
498  };
499 
502  virtual ~ReadDataStream() {}
503 
506  virtual ReadData* next(const uint32 batch_size, const uint32 batch_bps = uint32(-1)) = 0;
507 
510  virtual bool is_ok() = 0;
511 
512  // maximum length of a read; longer reads are truncated to this size
513  uint32 m_truncate_read_len;
514 };
515 
516 
529 ReadDataStream *open_read_file(const char * read_file_name,
530  const QualityEncoding qualities,
531  const uint32 max_reads = uint32(-1),
532  const uint32 max_read_len = uint32(-1),
533  const ReadEncoding flags = REVERSE);
534 
537 
538 } // namespace io
539 
542 inline
543 io::ReadData::plain_view_type plain_view(io::ReadData& read_data)
544 {
545  return io::ReadData::plain_view_type( read_data );
546 }
547 
550 inline
551 io::ReadData::const_plain_view_type plain_view(const io::ReadData& read_data)
552 {
553  return io::ReadData::const_plain_view_type( read_data );
554 }
555 
556 } // namespace nvbio
Generated on Wed Feb 25 2015 08:33:00 for NVBIO by   doxygen 1.8.4