Major changes : new cython subclasses to handle columns with multiple

elements per line in a more efficient way + now elements_names are
passed as a list + new function to recover only the header of a column
This commit is contained in:
Celine Mercier
2015-10-14 18:05:34 +02:00
parent 21923e213d
commit 0eaa5aa784
22 changed files with 749 additions and 516 deletions

View File

@ -639,7 +639,7 @@ OBIDMS_column_p obi_create_column(OBIDMS_p dms,
minimum_line_count = get_line_count_per_page(data_type, nb_elements_per_line);
if (nb_lines > MAXIMUM_LINE_COUNT)
{
obidebug(1, "\nCan't create column because of line count greater than the maximum allowed (%lld)", MAXIMUM_LINE_COUNT);
obidebug(1, "\nCan't create column because of line count greater than the maximum allowed (%ld)", MAXIMUM_LINE_COUNT);
return NULL;
}
else if (nb_lines < minimum_line_count)
@ -1341,7 +1341,7 @@ OBIType_t obi_column_get_data_type(OBIDMS_column_p column)
}
OBIType_t obi_column_get_data_type_from_name(OBIDMS_p dms, const char* column_name)
OBIDMS_column_header_p obi_column_get_header_from_name(OBIDMS_p dms, const char* column_name)
{
OBIDMS_column_header_p header;
OBIDMS_column_directory_p column_directory;
@ -1349,7 +1349,6 @@ OBIType_t obi_column_get_data_type_from_name(OBIDMS_p dms, const char* column_na
int column_file_descriptor;
int column_dir_file_descriptor;
size_t header_size;
OBIType_t data_type;
obiversion_t version_number;
// Get the column directory structure associated to the column
@ -1357,7 +1356,7 @@ OBIType_t obi_column_get_data_type_from_name(OBIDMS_p dms, const char* column_na
if (column_directory == NULL)
{
obidebug(1, "\nError opening a column directory structure");
return -1;
return NULL;
}
// Get the file descriptor associated to the column directory
@ -1367,7 +1366,7 @@ OBIType_t obi_column_get_data_type_from_name(OBIDMS_p dms, const char* column_na
obi_set_errno(OBICOLDIR_UNKNOWN_ERROR);
obidebug(1, "\nError getting the file descriptor of a column directory");
obi_close_column_directory(column_directory);
return -1;
return NULL;
}
// Calculate the header size
@ -1378,14 +1377,14 @@ OBIType_t obi_column_get_data_type_from_name(OBIDMS_p dms, const char* column_na
if (version_number < 0)
{
obidebug(1, "\nError getting the latest version number in a column directory");
return -1;
return NULL;
}
// Get the column file name
column_file_name = build_column_file_name(column_name, version_number);
if (column_file_name == NULL)
{
return -1;
return NULL;
}
// Open the column file (READ-ONLY)
@ -1395,7 +1394,7 @@ OBIType_t obi_column_get_data_type_from_name(OBIDMS_p dms, const char* column_na
obidebug(1, "\nError opening a column file");
obi_set_errno(OBICOL_UNKNOWN_ERROR);
free(column_file_name);
return -1;
return NULL;
}
// Fill the header structure
@ -1413,103 +1412,27 @@ OBIType_t obi_column_get_data_type_from_name(OBIDMS_p dms, const char* column_na
obidebug(1, "\nError mmapping the header of a column");
close(column_file_descriptor);
free(column_file_name);
return -1;
return NULL;
}
// TODO Check endianness?
data_type = header->data_type;
free(column_file_name);
close(column_file_descriptor);
munmap(header, header_size);
return data_type;
return header;
}
size_t obi_column_get_line_count_from_name(OBIDMS_p dms, const char* column_name)
int obi_unmap_header(OBIDMS_column_header_p header)
{
OBIDMS_column_header_p header;
OBIDMS_column_directory_p column_directory;
char* column_file_name;
int column_file_descriptor;
int column_dir_file_descriptor;
size_t header_size;
size_t line_count;
obiversion_t version_number;
// Get the column directory structure associated to the column
column_directory = obi_open_column_directory(dms, column_name);
if (column_directory == NULL)
{
obidebug(1, "\nError opening a column directory structure");
return -1;
}
// Get the file descriptor associated to the column directory
column_dir_file_descriptor = dirfd(column_directory->directory);
if (column_dir_file_descriptor < 0)
{
obi_set_errno(OBICOLDIR_UNKNOWN_ERROR);
obidebug(1, "\nError getting the file descriptor of a column directory");
obi_close_column_directory(column_directory);
return -1;
}
// Calculate the header size
header_size = obi_get_platform_header_size();
// Get the latest version number
version_number = obi_get_latest_version_number(column_directory);
if (version_number < 0)
{
obidebug(1, "\nError getting the latest version number in a column directory");
return -1;
}
// Get the column file name
column_file_name = build_column_file_name(column_name, version_number);
if (column_file_name == NULL)
{
return -1;
}
// Open the column file (READ-ONLY)
column_file_descriptor = openat(column_dir_file_descriptor, column_file_name, O_RDONLY);
if (column_file_descriptor < 0)
{
obidebug(1, "\nError opening a column file");
obi_set_errno(OBICOL_UNKNOWN_ERROR);
free(column_file_name);
return -1;
}
// Fill the header structure
header = mmap(NULL,
header_size,
PROT_READ,
MAP_SHARED,
column_file_descriptor,
0
);
if (header == MAP_FAILED)
if (munmap(header, header->header_size) < 0)
{
obi_set_errno(OBICOL_UNKNOWN_ERROR);
obidebug(1, "\nError mmapping the header of a column");
close(column_file_descriptor);
free(column_file_name);
obidebug(1, "\nError munmapping a column header");
return -1;
}
// TODO Check endianness?
line_count = header->line_count;
free(column_file_name);
close(column_file_descriptor);
munmap(header, header_size);
return line_count;
return 0;
}
@ -1567,7 +1490,7 @@ size_t obi_column_get_nb_elements_per_line(OBIDMS_column_p column)
}
char* obi_column_get_formatted_creation_date(OBIDMS_column_p column)
char* obi_column_get_formatted_creation_date(OBIDMS_column_p column) // TODO
{
char* formatted_time;
struct tm* tmp;
@ -1593,7 +1516,7 @@ char* obi_column_get_formatted_creation_date(OBIDMS_column_p column)
}
char* obi_column_get_formatted_creation_date_from_name(OBIDMS_p dms, const char* column_name)
char* obi_column_get_formatted_creation_date_from_name(OBIDMS_p dms, const char* column_name) // TODO
{
OBIDMS_column_header_p header;
OBIDMS_column_directory_p column_directory;
@ -1691,8 +1614,15 @@ char* obi_column_get_formatted_creation_date_from_name(OBIDMS_p dms, const char*
return NULL;
}
// Munmap header
if (munmap(header, header->header_size) < 0)
{
obi_set_errno(OBICOL_UNKNOWN_ERROR);
obidebug(1, "\nError munmapping header after getting the creation date of a column");
}
free(column_file_name);
close(column_file_descriptor);
munmap(header, header_size);
return formatted_time;
}

View File

@ -334,31 +334,34 @@ OBIType_t obi_column_get_data_type(OBIDMS_column_p column);
/**
* @brief Recovers the data type of an OBIDMS column from the column name.
* @brief Recovers the header of an OBIDMS column from the column name.
*
* @warning The header structure has to be munmapped by the caller.
*
* @param dms A pointer on an OBIDMS.
* @param column_name The name of an OBIDMS column.
*
* @returns The data type of the column.
* @returns A pointer on the mmapped header of the column.
* @retval NULL if an error occurred.
*
* @since July 2015
* @since October 2015
* @author Celine Mercier (celine.mercier@metabarcoding.org)
*/
OBIType_t obi_column_get_data_type_from_name(OBIDMS_p dms, const char* column_name);
OBIDMS_column_header_p obi_column_get_header_from_name(OBIDMS_p dms, const char* column_name);
/**
* @brief Recovers the line count of an OBIDMS column from the column name.
* @brief Munmap a mmapped header as returned by obi_column_get_header_from_name().
*
* @param dms A pointer on an OBIDMS.
* @param column_name The name of an OBIDMS column.
* @param header A pointer on the mmapped header structure.
*
* @returns The line count of the column.
* @retval 0 if the operation was successfully completed.
* @retval -1 if an error occurred.
*
* @since September 2015
* @since October 2015
* @author Celine Mercier (celine.mercier@metabarcoding.org)
*/
size_t obi_column_get_line_count_from_name(OBIDMS_p dms, const char* column_name);
int obi_unmap_header(OBIDMS_column_header_p header);
/**

View File

@ -72,27 +72,10 @@ obibool_t obi_column_get_obibool_with_elt_idx(OBIDMS_column_p column, size_t lin
int obi_column_set_obibool_with_elt_name(OBIDMS_column_p column, size_t line_nb, const char* element_name, obibool_t value)
{
size_t element_idx;
if (!strcmp(element_name, "\0")) // element name is empty
{
if (obi_column_get_nb_elements_per_line(column) == 1) // check that there is only one element per line
element_idx = 0;
else // there is more than one element per line
{
obi_set_errno(OBICOL_UNKNOWN_ERROR);
obidebug(1, "\nAn element name must be specified");
return -1;
}
}
else
{
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return -1;
}
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return -1;
obi_column_set_obibool_with_elt_idx(column, line_nb, element_idx, value);
return 0;
}
@ -101,24 +84,9 @@ obibool_t obi_column_get_obibool_with_elt_name(OBIDMS_column_p column, size_t li
{
size_t element_idx;
if (!strcmp(element_name, "\0")) // element name is empty
{
if (obi_column_get_nb_elements_per_line(column) == 1) // check that there is only one element per line
element_idx = 0;
else // there is more than one element per line
{
obi_set_errno(OBICOL_UNKNOWN_ERROR);
obidebug(1, "\nAn element name must be specified");
return OBIBool_NA;
}
}
else
{
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return OBIBool_NA;
}
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return OBIBool_NA;
return obi_column_get_obibool_with_elt_idx(column, line_nb, element_idx);
}

View File

@ -65,7 +65,6 @@ obichar_t obi_column_get_obichar_with_elt_idx(OBIDMS_column_p column, size_t lin
obidebug(1, "\nError trying to get a value that is beyond the current number of lines used");
return OBIChar_NA;
}
return *(((obichar_t*) (column->data)) + (line_nb * ((column->header)->nb_elements_per_line)) + element_idx);
}
@ -73,27 +72,10 @@ obichar_t obi_column_get_obichar_with_elt_idx(OBIDMS_column_p column, size_t lin
int obi_column_set_obichar_with_elt_name(OBIDMS_column_p column, size_t line_nb, const char* element_name, obichar_t value)
{
size_t element_idx;
if (!strcmp(element_name, "\0")) // element name is empty
{
if (obi_column_get_nb_elements_per_line(column) == 1) // check that there is only one element per line
element_idx = 0;
else // there is more than one element per line
{
obi_set_errno(OBICOL_UNKNOWN_ERROR);
obidebug(1, "\nAn element name must be specified");
return -1;
}
}
else
{
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return -1;
}
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return -1;
obi_column_set_obichar_with_elt_idx(column, line_nb, element_idx, value);
return 0;
}
@ -102,24 +84,9 @@ obichar_t obi_column_get_obichar_with_elt_name(OBIDMS_column_p column, size_t li
{
size_t element_idx;
if (!strcmp(element_name, "\0")) // element name is empty
{
if (obi_column_get_nb_elements_per_line(column) == 1) // check that there is only one element per line
element_idx = 0;
else // there is more than one element per line
{
obi_set_errno(OBICOL_UNKNOWN_ERROR);
obidebug(1, "\nAn element name must be specified");
return OBIChar_NA;
}
}
else
{
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return OBIChar_NA;
}
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return OBIChar_NA;
return obi_column_get_obichar_with_elt_idx(column, line_nb, element_idx);
}

View File

@ -72,27 +72,10 @@ obifloat_t obi_column_get_obifloat_with_elt_idx(OBIDMS_column_p column, size_t l
int obi_column_set_obifloat_with_elt_name(OBIDMS_column_p column, size_t line_nb, const char* element_name, obifloat_t value)
{
size_t element_idx;
if (!strcmp(element_name, "\0")) // element name is empty
{
if (obi_column_get_nb_elements_per_line(column) == 1) // check that there is only one element per line
element_idx = 0;
else // there is more than one element per line
{
obi_set_errno(OBICOL_UNKNOWN_ERROR);
obidebug(1, "\nAn element name must be specified");
return -1;
}
}
else
{
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return -1;
}
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return -1;
obi_column_set_obifloat_with_elt_idx(column, line_nb, element_idx, value);
return 0;
}
@ -101,24 +84,9 @@ obifloat_t obi_column_get_obifloat_with_elt_name(OBIDMS_column_p column, size_t
{
size_t element_idx;
if (!strcmp(element_name, "\0")) // element name is empty
{
if (obi_column_get_nb_elements_per_line(column) == 1) // check that there is only one element per line
element_idx = 0;
else // there is more than one element per line
{
obi_set_errno(OBICOL_UNKNOWN_ERROR);
obidebug(1, "\nAn element name must be specified");
return OBIFloat_NA;
}
}
else
{
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return OBIFloat_NA;
}
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return OBIFloat_NA;
return obi_column_get_obifloat_with_elt_idx(column, line_nb, element_idx);
}

View File

@ -10,6 +10,18 @@
*/
/****************************************************************************
* OBIDMS_column_idx functions *
****************************************************************************/
/**
* @file obidsmcolumn_idx.c
* @author Celine Mercier
* @date August 10th 2015
* @brief Functions handling OBIColumns containing data with the OBIType OBI_BOOL.
*/
#include <stdlib.h>
#include <stdio.h>
@ -72,27 +84,10 @@ obiidx_t obi_column_get_obiidx_with_elt_idx(OBIDMS_column_p column, size_t line_
int obi_column_set_obiidx_with_elt_name(OBIDMS_column_p column, size_t line_nb, const char* element_name, obiidx_t value)
{
size_t element_idx;
if (!strcmp(element_name, "\0")) // element name is empty
{
if (obi_column_get_nb_elements_per_line(column) == 1) // check that there is only one element per line
element_idx = 0;
else // there is more than one element per line
{
obi_set_errno(OBICOL_UNKNOWN_ERROR);
obidebug(1, "\nAn element name must be specified");
return -1;
}
}
else
{
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return -1;
}
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return -1;
obi_column_set_obiidx_with_elt_idx(column, line_nb, element_idx, value);
return 0;
}
@ -101,24 +96,9 @@ obiidx_t obi_column_get_obiidx_with_elt_name(OBIDMS_column_p column, size_t line
{
size_t element_idx;
if (!strcmp(element_name, "\0")) // element name is empty
{
if (obi_column_get_nb_elements_per_line(column) == 1) // check that there is only one element per line
element_idx = 0;
else // there is more than one element per line
{
obi_set_errno(OBICOL_UNKNOWN_ERROR);
obidebug(1, "\nAn element name must be specified");
return OBIIdx_NA;
}
}
else
{
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return OBIIdx_NA;
}
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return OBIIdx_NA;
return obi_column_get_obiidx_with_elt_idx(column, line_nb, element_idx);
}

View File

@ -72,27 +72,10 @@ obiint_t obi_column_get_obiint_with_elt_idx(OBIDMS_column_p column, size_t line_
int obi_column_set_obiint_with_elt_name(OBIDMS_column_p column, size_t line_nb, const char* element_name, obiint_t value)
{
size_t element_idx;
if (!strcmp(element_name, "\0")) // element name is empty
{
if (obi_column_get_nb_elements_per_line(column) == 1) // check that there is only one element per line
element_idx = 0;
else // there is more than one element per line
{
obi_set_errno(OBICOL_UNKNOWN_ERROR);
obidebug(1, "\nAn element name must be specified");
return -1;
}
}
else
{
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return -1;
}
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return -1;
obi_column_set_obiint_with_elt_idx(column, line_nb, element_idx, value);
return 0;
}
@ -101,24 +84,9 @@ obiint_t obi_column_get_obiint_with_elt_name(OBIDMS_column_p column, size_t line
{
size_t element_idx;
if (!strcmp(element_name, "\0")) // element name is empty
{
if (obi_column_get_nb_elements_per_line(column) == 1) // check that there is only one element per line
element_idx = 0;
else // there is more than one element per line
{
obi_set_errno(OBICOL_UNKNOWN_ERROR);
obidebug(1, "\nAn element name must be specified");
return OBIInt_NA;
}
}
else
{
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return OBIInt_NA;
}
element_idx = obi_column_get_element_index_from_name(column, element_name);
if (element_idx == SIZE_MAX) //TODO
return OBIInt_NA;
return obi_column_get_obiint_with_elt_idx(column, line_nb, element_idx);
}