obitools/obitools3
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 66 Wiki Activity

DNA_seq_indexer API

Browse Source
This commit is contained in:
Celine Mercier
2016-04-12 16:38:47 +02:00
parent 04e3a7b5a9
commit be05c889e2
16 changed files with 200 additions and 97 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
python/obitools3/obidms/_obidms.cfiles
View File

@ -2,6 +2,8 @@
../../../src/bloom.c
../../../src/crc64.h
../../../src/crc64.c
../../../src/dna_seq_indexer.h
../../../src/dna_seq_indexer.c
../../../src/encode.h
../../../src/encode.c
../../../src/MurmurHash2.h

2
python/obitools3/obidms/_obidmscolumn_bool.cfiles
View File

@ -4,6 +4,8 @@
../../../src/bloom.c
../../../src/crc64.h
../../../src/crc64.c
../../../src/dna_seq_indexer.h
../../../src/dna_seq_indexer.c
../../../src/encode.h
../../../src/encode.c
../../../src/MurmurHash2.h

2
python/obitools3/obidms/_obidmscolumn_char.cfiles
View File

@ -4,6 +4,8 @@
../../../src/bloom.c
../../../src/crc64.h
../../../src/crc64.c
../../../src/dna_seq_indexer.h
../../../src/dna_seq_indexer.c
../../../src/encode.h
../../../src/encode.c
../../../src/MurmurHash2.h

2
python/obitools3/obidms/_obidmscolumn_float.cfiles
View File

@ -4,6 +4,8 @@
../../../src/bloom.c
../../../src/crc64.h
../../../src/crc64.c
../../../src/dna_seq_indexer.h
../../../src/dna_seq_indexer.c
../../../src/encode.h
../../../src/encode.c
../../../src/MurmurHash2.h

2
python/obitools3/obidms/_obidmscolumn_int.cfiles
View File

@ -4,6 +4,8 @@
../../../src/bloom.c
../../../src/crc64.h
../../../src/crc64.c
../../../src/dna_seq_indexer.h
../../../src/dna_seq_indexer.c
../../../src/encode.h
../../../src/encode.c
../../../src/MurmurHash2.h

2
python/obitools3/obidms/_obidmscolumn_seq.cfiles
View File

@ -4,6 +4,8 @@
../../../src/bloom.c
../../../src/crc64.h
../../../src/crc64.c
../../../src/dna_seq_indexer.h
../../../src/dna_seq_indexer.c
../../../src/encode.h
../../../src/encode.c
../../../src/MurmurHash2.h

2
python/obitools3/obidms/_obidmscolumn_str.cfiles
View File

@ -4,6 +4,8 @@
../../../src/bloom.c
../../../src/crc64.h
../../../src/crc64.c
../../../src/dna_seq_indexer.h
../../../src/dna_seq_indexer.c
../../../src/encode.h
../../../src/encode.c
../../../src/MurmurHash2.h

2
python/obitools3/obidms/_obiseq.cfiles
View File

@ -2,6 +2,8 @@
../../../src/bloom.c
../../../src/crc64.h
../../../src/crc64.c
../../../src/dna_seq_indexer.h
../../../src/dna_seq_indexer.c
../../../src/encode.h
../../../src/encode.c
../../../src/MurmurHash2.h

2
python/obitools3/obidms/_obitaxo.cfiles
View File

@ -2,6 +2,8 @@
../../../src/bloom.c
../../../src/crc64.h
../../../src/crc64.c
../../../src/dna_seq_indexer.h
../../../src/dna_seq_indexer.c
../../../src/encode.h
../../../src/encode.c
../../../src/MurmurHash2.h

106
src/dna_seq_indexer.c Normal file
View File

@ -0,0 +1,106 @@
/****************************************************************************
* DNA sequence indexing functions *
****************************************************************************/
/**
* @file dna_seq_indexer.c
* @author Celine Mercier
* @date April 12th 2016
* @brief Functions handling the indexing and retrieval of DNA sequences.
*/
#include <stdlib.h>
#include <stdio.h>
#include <math.h>
#include "obiblob.h"
#include "obiblob_indexer.h"
#include "obidms.h"
#define DEBUG_LEVEL 0 // TODO has to be defined somewhere else (cython compil flag?)
Obi_blob_p obi_seq_to_blob(const char* seq)
{
Obi_blob_p value_b;
int32_t length_encoded_seq; // length of the encoded sequence in bytes
int32_t seq_length;
byte_t* encoded_seq;
seq_length = strlen(seq);
// Check if just ATGC and encode accordingly
if (only_ATGC(seq))
{
// Compute the length (in bytes) of the encoded sequence
length_encoded_seq = ceil((double) seq_length / (double) 4.0);
// Encode
encoded_seq = encode_seq_on_2_bits(seq, seq_length);
if (encoded_seq == NULL)
return NULL;
value_b = obi_blob(encoded_seq, ELEMENT_SIZE_SEQ_2, length_encoded_seq, seq_length);
}
else
{
// Compute the length (in bytes) of the encoded sequence
length_encoded_seq = ceil((double) seq_length / (double) 2.0);
// Encode
encoded_seq = encode_seq_on_4_bits(seq, seq_length);
if (encoded_seq == NULL)
return NULL;
value_b = obi_blob(encoded_seq, ELEMENT_SIZE_SEQ_4, length_encoded_seq, seq_length);
}
free(encoded_seq);
return value_b;
}
char* obi_blob_to_seq(Obi_blob_p value_b)
{
// Decode
if (value_b->element_size == 2)
return decode_seq_on_2_bits(value_b->value, value_b->length_decoded_value);
else
return decode_seq_on_4_bits(value_b->value, value_b->length_decoded_value);
}
index_t obi_index_dna_seq(Obi_indexer_p indexer, const char* value)
{
Obi_blob_p value_b;
index_t idx;
// Encode value
value_b = obi_seq_to_blob(value);
if (value_b == NULL)
return -1;
// Add in the indexer
idx = obi_indexer_add(indexer, value_b);
free(value_b);
return idx;
}
char* obi_retrieve_dna_seq(Obi_indexer_p indexer, index_t idx)
{
Obi_blob_p value_b;
char* seq;
// Get encoded value
value_b = obi_indexer_get(indexer, idx);
// Decode sequence
seq = obi_blob_to_seq(value_b);
free(value_b);
return seq;
}

63
src/dna_seq_indexer.h Normal file
View File

@ -0,0 +1,63 @@
/****************************************************************************
* DNA sequence indexer header file *
****************************************************************************/
/**
* @file dna_seq_indexer.h
* @author Celine Mercier
* @date April 12th 2016
* @brief Header file for the functions handling the indexing of DNA sequences.
*/
#ifndef DNA_SEQ_INDEXER_H_
#define DNA_SEQ_INDEXER_H_
#include <stdlib.h>
#include <stdio.h>
#include "obidms.h"
#include "obitypes.h"
#include "obiblob.h"
#include "obiblob_indexer.h"
/**
* @brief Converts a DNA sequence to a blob.
*
* @warning The blob must be freed by the caller.
*
* @param value The DNA sequence to convert.
*
* @returns A pointer to the blob created.
* @retval NULL if an error occurred.
*
* @since November 2015
* @author Celine Mercier (celine.mercier@metabarcoding.org)
*/
Obi_blob_p obi_seq_to_blob(const char* seq);
/**
* @brief Converts a blob to a DNA sequence.
*
* @param value_b The blob to convert.
*
* @returns A pointer to the DNA sequence contained in the blob.
* @retval NULL if an error occurred.
*
* @since November 2015
* @author Celine Mercier (celine.mercier@metabarcoding.org)
*/
char* obi_blob_to_seq(Obi_blob_p value_b);
// TODO doc
index_t obi_index_dna_seq(Obi_indexer_p indexer, const char* value);
char* obi_retrieve_dna_seq(Obi_indexer_p indexer, index_t idx);
#endif /* DNA_SEQ_INDEXER_H_ */

4
src/encode.c
View File

@ -54,7 +54,7 @@ bool only_ATGC(const char* seq)
}
byte_t* encode_seq_on_2_bits(char* seq, int32_t length)
byte_t* encode_seq_on_2_bits(const char* seq, int32_t length)
{
byte_t* seq_b;
uint8_t modulo;
@ -163,7 +163,7 @@ char* decode_seq_on_2_bits(byte_t* seq_b, int32_t length_seq)
}
byte_t* encode_seq_on_4_bits(char* seq, int32_t length)
byte_t* encode_seq_on_4_bits(const char* seq, int32_t length)
{
byte_t* seq_b;
uint8_t modulo;

4
src/encode.h
View File

@ -96,7 +96,7 @@ bool only_ATGC(const char* seq);
* @since November 2015
* @author Celine Mercier (celine.mercier@metabarcoding.org)
*/
byte_t* encode_seq_on_2_bits(char* seq, int32_t length);
byte_t* encode_seq_on_2_bits(const char* seq, int32_t length);
/**
@ -147,7 +147,7 @@ char* decode_seq_on_2_bits(byte_t* seq_b, int32_t length_seq);
* @since November 2015
* @author Celine Mercier (celine.mercier@metabarcoding.org)
*/
byte_t* encode_seq_on_4_bits(char* seq, int32_t length);
byte_t* encode_seq_on_4_bits(const char* seq, int32_t length);
/**

47
src/obiblob.c
View File

@ -82,52 +82,5 @@ const char* obi_blob_to_str(Obi_blob_p value_b)
}
Obi_blob_p obi_seq_to_blob(char* seq)
{
Obi_blob_p value_b;
int32_t length_encoded_seq; // length of the encoded sequence in bytes
int32_t seq_length;
byte_t* encoded_seq;
seq_length = strlen(seq);
// Check if just ATGC and encode accordingly
if (only_ATGC(seq))
{
// Compute the length (in bytes) of the encoded sequence
length_encoded_seq = ceil((double) seq_length / (double) 4.0);
// Encode
encoded_seq = encode_seq_on_2_bits(seq, seq_length);
if (encoded_seq == NULL)
return NULL;
value_b = obi_blob(encoded_seq, ELEMENT_SIZE_SEQ_2, length_encoded_seq, seq_length);
}
else
{
// Compute the length (in bytes) of the encoded sequence
length_encoded_seq = ceil((double) seq_length / (double) 2.0);
// Encode
encoded_seq = encode_seq_on_4_bits(seq, seq_length);
if (encoded_seq == NULL)
return NULL;
value_b = obi_blob(encoded_seq, ELEMENT_SIZE_SEQ_4, length_encoded_seq, seq_length);
}
free(encoded_seq);
return value_b;
}
const char* obi_blob_to_seq(Obi_blob_p value_b)
{
// Decode
if (value_b->element_size == 2)
return decode_seq_on_2_bits(value_b->value, value_b->length_decoded_value);
else
return decode_seq_on_4_bits(value_b->value, value_b->length_decoded_value);
}
// TODO same for int

35
src/obiblob.h
View File

@ -45,6 +45,11 @@ typedef struct Obi_blob {
} Obi_blob_t, *Obi_blob_p;
// TODO doc
Obi_blob_p obi_blob(byte_t* encoded_value, uint8_t element_size, int32_t length_encoded_value, int32_t length_decoded_value);
/**
* @brief Converts a character string to a blob.
*
@ -74,35 +79,5 @@ Obi_blob_p obi_str_to_blob(char* value);
const char* obi_blob_to_str(Obi_blob_p value_b);
/**
* @brief Converts a DNA sequence to a blob with a header.
*
* @warning The blob must be freed by the caller.
*
* @param value The DNA sequence to convert.
*
* @returns A pointer to the blob created.
* @retval NULL if an error occurred.
*
* @since November 2015
* @author Celine Mercier (celine.mercier@metabarcoding.org)
*/
Obi_blob_p obi_seq_to_blob(char* seq);
/**
* @brief Converts a blob to a DNA sequence.
*
* @param value_b The blob to convert.
*
* @returns A pointer to the DNA sequence contained in the blob.
* @retval NULL if an error occurred.
*
* @since November 2015
* @author Celine Mercier (celine.mercier@metabarcoding.org)
*/
const char* obi_blob_to_seq(Obi_blob_p value_b); // TODO move to encode source files
#endif /* OBIBLOB_H_ */

18
src/obidmscolumn_seq.c
View File

@ -18,7 +18,7 @@
#include "obitypes.h"
#include "obierrno.h"
#include "obidebug.h"
#include "obiblob_indexer.h"
#include "dna_seq_indexer.h"
#define DEBUG_LEVEL 0 // TODO has to be defined somewhere else (cython compil flag?)
@ -33,7 +33,6 @@
int obi_column_set_obiseq_with_elt_idx(OBIDMS_column_p column, index_t line_nb, index_t element_idx, char* value)
{
Obi_blob_p value_b;
index_t idx;
// Check that the line number is not greater than the maximum allowed
@ -56,21 +55,13 @@ int obi_column_set_obiseq_with_elt_idx(OBIDMS_column_p column, index_t line_nb,
if ((line_nb+1) > (column->header)->lines_used)
(column->header)->lines_used = line_nb+1;
// Encode the value on a byte array with a header // TODO make function
value_b = obi_seq_to_blob(value);
if (value_b == NULL)
return -1;
// Add in the indexer
idx = obi_indexer_add(column->indexer, value_b);
idx = obi_index_dna_seq(column->indexer, value);
if (idx == -1)
return -1;
// Add the value's index in the column
*(((index_t*) (column->data)) + (line_nb * ((column->header)->nb_elements_per_line)) + element_idx) = idx;
free(value_b);
return 0;
}
@ -115,7 +106,6 @@ int obi_column_set_obiseq_with_elt_idx_in_view(Obiview_p view, OBIDMS_column_p c
const char* obi_column_get_obiseq_with_elt_idx(OBIDMS_column_p column, index_t line_nb, index_t element_idx)
{
index_t idx;
Obi_blob_p value_b;
if ((line_nb+1) > ((column->header)->line_count))
{
@ -130,9 +120,7 @@ const char* obi_column_get_obiseq_with_elt_idx(OBIDMS_column_p column, index_t l
if (idx == OBIIdx_NA)
return OBISeq_NA;
value_b = obi_indexer_get(column->indexer, idx);
return obi_blob_to_seq(value_b);
return obi_retrieve_dna_seq(column->indexer, idx);
}