Defines API for reading a collection of data. More...

#include <DC.h>

Inheritance diagram for VAPoR::DC:

Classes
class	Attribute
	Variable or global metadata. More...

class	BaseVar
	Base class for storing variable metadata. More...

class	CoordVar
	Coordinate variable metadata. More...

class	DataVar
	Data variable metadata. More...

class	Dimension
	Metadata describing a coordinate dimension. More...

Public Types
enum	XType { INVALID = -1, FLOAT, DOUBLE, INT32, INT64, TEXT }
	External storage types for primitive data. More...

Public Types inherited from VetsUtil::MyBase
typedef void(*	ErrMsgCB_T) (const char *msg, int err_code)

typedef void(*	DiagMsgCB_T) (const char *msg)

Public Member Functions
	DC ()

virtual	~DC ()

virtual int	Initialize (const vector< string > &paths)=0

virtual bool	GetDimension (string dimname, DC::Dimension &dimension) const =0

virtual std::vector< string >	GetDimensionNames () const =0

virtual bool	GetCoordVarInfo (string varname, DC::CoordVar &cvar) const =0

virtual bool	GetDataVarInfo (string varname, DC::DataVar &datavar) const =0

virtual bool	GetBaseVarInfo (string varname, DC::BaseVar &var) const =0

virtual std::vector< string >	GetDataVarNames () const =0

virtual std::vector< string >	GetCoordVarNames () const =0

virtual int	GetNumRefLevels (string varname) const =0

virtual int	GetAtt (string varname, string attname, vector< double > &values) const =0

virtual int	GetAtt (string varname, string attname, vector< long > &values) const =0

virtual int	GetAtt (string varname, string attname, string &values) const =0

virtual std::vector< string >	GetAttNames (string varname) const =0

virtual XType	GetAttType (string varname, string attname) const =0

virtual int	GetDimLensAtLevel (string varname, int level, std::vector< size_t > &dims_at_level, std::vector< size_t > &bs_at_level) const =0

virtual int	GetMapProjection (string lonname, string latname, string &projstring) const =0

virtual int	OpenVariableRead (size_t ts, string varname, int level=0, int lod=0)=0

virtual int	CloseVariable ()=0

virtual int	Read (float *data)=0

virtual int	ReadSlice (float *slice)=0

virtual int	ReadRegion (const vector< size_t > &min, const vector< size_t > &max, float *region)=0

virtual int	ReadRegionBlock (const vector< size_t > &min, const vector< size_t > &max, float *region)=0

virtual int	GetVar (string varname, int level, int lod, float *data)=0

virtual int	GetVar (size_t ts, string varname, int level, int lod, float *data)=0

virtual bool	VariableExists (size_t ts, string varname, int reflevel=0, int lod=0) const =0

virtual bool	GetDimension (string dimname, size_t &length, int &axis) const

virtual std::vector< string >	GetDataVarNames (int ndim, bool spatial) const

virtual std::vector< string >	GetCoordVarNames (int ndim, bool spatial) const

virtual bool	IsTimeVarying (string varname) const

virtual bool	IsCompressed (string varname) const

virtual int	GetNumTimeSteps (string varname) const

virtual int	GetCRatios (string varname, vector< size_t > &cratios) const

virtual bool	IsDataVar (string varname) const

virtual bool	IsCoordVar (string varname) const

virtual int	GetMapProjection (string varname, string &projstring) const

Public Member Functions inherited from VetsUtil::MyBase
	MyBase ()

const string &	getClassName () const

Static Public Member Functions
static bool	ParseDimensions (const vector< DC::Dimension > &dimensions, vector< size_t > &sdims, size_t &numts)

Static Public Member Functions inherited from VetsUtil::MyBase
static void	SetErrMsg (const char *format,...)
	Record a formatted error message. More...

static void	SetErrMsg (int errcode, const char *format,...)
	Record a formatted error message and an error code. More...

static const char *	GetErrMsg ()

static void	SetErrCode (int err_code)
	Record an error code. More...

static int	GetErrCode ()
	Retrieve the current error code. More...

static void	SetErrMsgCB (ErrMsgCB_T cb)

static ErrMsgCB_T	GetErrMsgCB ()

static void	SetErrMsgFilePtr (FILE *fp)

static const FILE *	SetErrMsgFilePtr ()

static void	SetDiagMsg (const char *format,...)
	Record a formatted diagnostic message. More...

static const char *	GetDiagMsg ()

static void	SetDiagMsgCB (DiagMsgCB_T cb)

static DiagMsgCB_T	GetDiagMsgCB ()

static void	SetDiagMsgFilePtr (FILE *fp)

static bool	EnableErrMsg (bool enable)

Additional Inherited Members
Static Public Attributes inherited from VetsUtil::MyBase
static char *	ErrMsg

static int	ErrCode

static int	ErrMsgSize

static FILE *	ErrMsgFilePtr

static ErrMsgCB_T	ErrMsgCB

static char *	DiagMsg

static int	DiagMsgSize

static FILE *	DiagMsgFilePtr

static DiagMsgCB_T	DiagMsgCB

static bool	Enabled

Protected Member Functions inherited from VetsUtil::MyBase
void	SetClassName (const string &name)

Detailed Description

Defines API for reading a collection of data.

Author: John Clyne

Date: January, 2015

The abstract Data Collection (DC) class defines an API for reading metadata and sampled data from a data collection. A data collection is a set of related data, most typically the discrete outputs from a single numerical simulation. The DC class is an abstract virtual class, providing a public API, but performing no actual storage operations. Derived implementations of the DC base class are required to support the API.

Variables in a DC may have 1, 2, or 3 spatial dimensions, and 0 or 1 temporal dimensions.

The DC is structured in the spirit of the "NetCDF Climate and Forecast (CF) Metadata Conventions", version 1.6, 5, December 2011. It supports only a subset of the CF functionality (e.g. there is no support for "Discrete Sampling Geometries"). Moreover, it is more restrictive than the CF in a number of areas. Particular items of note include:

The API supports variables with 1 to 4 dimensions only.

Coordinate variables representing time must be 1D

Each dimension is associated with exactly one axis: X (or longitude) Y (or latitude), Z (height), and T (time)

All data variables have a "coordinate" attribute identifying the coordinate (or auxilliary coordinate) variables associated with each axis

To be consistent with VAPOR, when specified in vector form the ordering of dimension lengths and dimension names is from fastest varying dimension to slowest. For example, if 'dims' is a vector of dimensions, then dims[0] is the fastest varying dimension, dim[1] is the next fastest, and so on. This ordering is the opposite of the ordering used by NetCDF.

This class inherits from VetsUtil::MyBase. Unless otherwise documented any method that returns an integer value is returning status. A negative value indicates failure. Error messages are logged via VetsUtil::MyBase::SetErrMsg(). Methods that return a boolean do not, unless otherwise documented, log an error message upon failure (return of false).

Parameters

level

Grid refinement level for multiresolution variables. Compressed variables in the DC, if they exist, have a multi-resolution representation: the sampling grid for multi-resolution variables is hierarchical, and the dimension lengths of adjacent levels in the hierarchy differ by a factor of two. The level parameter is used to select a particular depth of the hierarchy.

To provide maximum flexibility as well as compatibility with previous versions of the DC the interpretation of level is somewhat complex. Both positive and negative values may be used to specify the refinement level and have different interpretations.

For positive values of level, a value of 0 indicates the coarsest member of the grid hierarchy. A value of 1 indicates the next grid refinement after the coarsest, and so on. Using postive values the finest level in the hierarchy is given by GetNumRefLevels() - 1. Values of level that are greater than GetNumRefLevels() - 1 are treated as if they were equal to GetNumRefLevels() - 1.

For negative values of level a value of -1 indicates the variable's native grid resolution (the finest resolution available). A value of -2 indicates the next coarsest member in the hierarchy after the finest, and so on. Using negative values the coarsest available level in the hierarchy is given by negating the value returned by GetNumRefLevels(). Values of level that are less than the negation of GetNumRefLevels() are treated as if they were equal to the negation of the GetNumRefLevels() return value.

lod The level-of-detail parameter, lod, selects the approximation level for a compressed variable. The lod parameter is similar to the level parameter in that it provides control over accuracy of a compressed variable. However, instead of selecting the grid resolution the lod parameter controls the compression factor by indexing into the cratios vector (see below). As with the level parameter, both positive and negative values may be used to index into cratios and different interpretations.

For positive values of lod, a value of 0 indicates the the first element of cratios, a value of 1 indicates the second element, and so on up to the size of the cratios vector (See DC::GetCRatios()).

For negative values of lod a value of -1 indexes the last element of cratios, a value of -2 indexes the second to last element, and so on. Using negative values the first element of cratios - the greatest compression rate - is indexed by negating the size of the cratios vector.

Parameters

cratios	A monotonically decreasing vector of compression ratios. Compressed variables in the DC are stored with a fixed, finite number of compression factors. The `cratios` vector is used to specify the available compression factors (ratios). A compression factor of 1 indicates no compression (1:1). A value of 2 indciates two to one compression (2:1), and so on. The minimum valid value of `cratios` is 1. The maximum value is determined by a number of factors and can be obtained using the CompressionInfo() method.
bs	An ordered list of block dimensions that specifies the block decomposition of the variable. The rank of `bs` may be less than that of a variable's array dimensions, in which case only the n fastest varying variable dimensions will be blocked, where n is the rank of `bs`. The ordering of the dimensions in `bs` is from fastest to slowest. A block is the basic unit of compression in the DC: variables are decomposed into blocks, and individual blocks are compressed independently.
wname	Name of wavelet used for transforming compressed variables between wavelet and physical space. Valid values are "bior1.1", "bior1.3", "bior1.5", "bior2.2", "bior2.4", "bior2.6", "bior2.8", "bior3.1", "bior3.3", "bior3.5", "bior3.7", "bior3.9", "bior4.4"

Definition at line 150 of file DC.h.

Member Enumeration Documentation

enum VAPoR::DC::XType

External storage types for primitive data.

Enumerator
INVALID
FLOAT
DOUBLE
INT32
INT64
TEXT

Definition at line 155 of file DC.h.

Constructor & Destructor Documentation

VAPoR::DC::DC ( )

inline

Class constuctor

Definition at line 786 of file DC.h.

virtual VAPoR::DC::~DC ( )

inlinevirtual

Definition at line 787 of file DC.h.

Member Function Documentation

virtual int VAPoR::DC::CloseVariable ( )

pure virtual

Close the currently opened variable

Close the handle for variable opened with OpenVariableRead()

See also: OpenVariableRead()

Implemented in VAPoR::VDC, VAPoR::VDCNetCDF, and VAPoR::DCWRF.

virtual int VAPoR::DC::GetAtt	(	string	varname,
		string	attname,
		vector< double > &	values
	)		const

pure virtual

Read an attribute

This method reads an attribute from the DC. The attribute can either be "global", if varname is the empty string, or bound to a variable if varname indentifies a variable in the DC.

Parameters

[in]	varname	The name of the variable the attribute is bound to, or the empty string if the attribute is global
[in]	attname	The attributes name
[out]	type	The primitive data type storage format. This is the type that will be used to store the attribute on disk
[out]	values	A vector to contain the returned floating point attribute values

Return values

status A negative int is returned on failure

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual int VAPoR::DC::GetAtt	(	string	varname,
		string	attname,
		vector< long > &	values
	)		const

pure virtual

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual int VAPoR::DC::GetAtt	(	string	varname,
		string	attname,
		string &	values
	)		const

pure virtual

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual std::vector<string> VAPoR::DC::GetAttNames ( string varname ) const

pure virtual

Return a list of available attribute's names

Returns a vector of all attribute names for the variable, varname. If varname is the empty string the names of all of the global attributes are returned. If varname is not defined an empty vector is returned.

Parameters

[in] varname The name of the variable to query, or the empty string if the names of global attributes are desired.

Return values

attnames A vector of returned attribute names

See also: GetAtt()

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual XType VAPoR::DC::GetAttType	(	string	varname,
		string	attname
	)		const

pure virtual

Return the external data type for an attribute

Returns the external storage type of the named variable attribute.

Parameters

[in]	varname	The name of the variable to query, or the empty string if the names of global attributes are desired.
[in]	name	Name of the attribute.

Return values

If	an attribute named by `name` does not exist, a negative value is returned.

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual bool VAPoR::DC::GetBaseVarInfo	(	string	varname,
		DC::BaseVar &	var
	)		const

pure virtual

Return metadata about a data or coordinate variable

If the variable varname is defined as either a data or coordinate variable its metadata will be returned in var.

Return values

bool	If the named variable cannot be found false is returned and the values of `var` are undefined.

See also: GetDataVarInfo(), GetCoordVarInfo()

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual bool VAPoR::DC::GetCoordVarInfo	(	string	varname,
		DC::CoordVar &	cvar
	)		const

pure virtual

Return a coordinate variable's definition

Return a reference to a DC::CoordVar object describing the coordinate variable named by varname

Parameters

[in]	varname	A string specifying the name of the coordinate variable.
[out]	coordvar	A CoordVar object containing the definition of the named variable.

Return values

bool	False is returned if the named coordinate variable does not exist, and the contents of `cvar` will be undefined.

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual std::vector<string> VAPoR::DC::GetCoordVarNames ( ) const

pure virtual

Return a list of names for all of the defined coordinate variables.

Returns a list of names for all coordinate variables defined

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual std::vector<string> VAPoR::DC::GetCoordVarNames	(	int	ndim,
		bool	spatial
	)		const

virtual

Return a list of coordinate variables with a given dimension rank

Returns a list of all coordinate variables defined having a dimension rank of ndim. If spatial is true, only the spatial dimension rank of the variable is compared against ndim

Parameters

[in]	ndim	Rank of dimensions for comparision
[in]	spatial	Only compare spatial dimensions against `ndim`

virtual int VAPoR::DC::GetCRatios	(	string	varname,
		vector< size_t > &	cratios
	)		const

virtual

Return the compression ratio vector for the indicated variable

Return the compression ratio vector for the indicated variable. The vector returned contains an ordered list of available compression ratios for the variable named by variable. If the variable is not compressed, the cratios parameter will contain a single element, one.

Parameters

[in]	varname	Data or coordinate variable name.
[out]	cratios	Ordered vector of compression ratios

Return values

status A negative int is returned on failure

virtual bool VAPoR::DC::GetDataVarInfo	(	string	varname,
		DC::DataVar &	datavar
	)		const

pure virtual

Return a data variable's definition

Return a reference to a DC::DataVar object describing the data variable named by varname

Parameters

[in]	varname	A string specifying the name of the variable.
[out]	datavar	A DataVar object containing the definition of the named Data variable.

Return values

bool	If the named data variable cannot be found false is returned and the values of `datavar` are undefined.

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual std::vector<string> VAPoR::DC::GetDataVarNames ( ) const

pure virtual

Return a list of names for all of the defined data variables.

Returns a list of names for all data variables defined

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual std::vector<string> VAPoR::DC::GetDataVarNames	(	int	ndim,
		bool	spatial
	)		const

virtual

Return a list of data variables with a given dimension rank

Returns a list of all data variables defined having a dimension rank of ndim. If spatial is true, only the spatial dimension rank of the variable is compared against ndim

Parameters

[in]	ndim	Rank of dimensions for comparision
[in]	spatial	Only compare spatial dimensions against `ndim`

virtual bool VAPoR::DC::GetDimension	(	string	dimname,
		DC::Dimension &	dimension
	)		const

pure virtual

Return a dimensions's definition

This method returns the definition of the dimension named by dimname as a reference to a DC::Dimension object. If dimname is not defined as a dimension then the name of dimension will be the empty string()

Parameters

[in]	dimname	A string specifying the name of the dimension.
[out]	dimension	The returned Dimension object reference

Return values

bool	If the named dimension can not be found false is returned.

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual bool VAPoR::DC::GetDimension	(	string	dimname,
		size_t &	length,
		int &	axis
	)		const

virtual

Return a dimensions's definition

This method returns the length and axis for the dimension named by dimname. If dimname is not defined as a dimension length and axis will both be set to zero, and false returned.

Parameters

[in]	dimname	A string specifying the name of the dimension.
[out]	length	The dimension length, which must be greater than zero.
[out]	axis	The axis associated with the dimension.

Return values

bool	If the named dimension can not be found false is returned.

See also: DC::Dimension

virtual std::vector<string> VAPoR::DC::GetDimensionNames ( ) const

pure virtual

Return names of all defined dimensions

This method returns the list of names of all of the dimensions defined in the DC.

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual int VAPoR::DC::GetDimLensAtLevel	(	string	varname,
		int	level,
		std::vector< size_t > &	dims_at_level,
		std::vector< size_t > &	bs_at_level
	)		const

pure virtual

Return a variable's dimension lengths at a specified refinement level

Compressed variables have a multi-resolution grid representation. This method returns the variable's ordered spatial and temporal dimension lengths, and block dimensions at the multiresolution refinement level specified by level.

If the variable named by varname is not compressed the variable's native dimensions are returned.

Parameters

[in]	varname	Data or coordinate variable name.
[in]	level	Specifies a member of a multi-resolution variable's grid hierarchy as described above.
[out]	dims_at_level	An ordered vector containing the variable's dimensions at the specified refinement level
[out]	bs_at_level	An ordered vector containing the variable's block dimensions at the specified refinement level

Return values

status Zero is returned upon success, otherwise -1.

See also: VAPoR::DC, DC::BaseVar::GetBS(), DC::BaseVar::GetDimensions()

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual int VAPoR::DC::GetMapProjection	(	string	lonname,
		string	latname,
		string &	projstring
	)		const

pure virtual

Return a Proj4 map projection string.

For georeference data sets that have map projections this method returns a properly formatted Proj4 projection string for mapping from geographic to cartographic coordinates. If no such projection exists an empty string is returned.

Parameters

[in]	lonname	Name of longitude coordinate variable
[in]	latname	Name of latitude coordinate variable
[out]	projstring	An empty string if a Proj4 map projection is not available for the named coordinate pair, otherwise a properly formatted Proj4 projection string is returned.

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual int VAPoR::DC::GetMapProjection	(	string	varname,
		string &	projstring
	)		const

virtual

Return a Proj4 map projection string for the named variable.

For georeference data sets that have map projections this method returns a properly formatted Proj4 projection string for mapping from geographic to cartographic coordinates. If no such projection exists for the named data variable an empty string is returned.

Parameters

[in]	varname	Name of a data variable
[out]	projstring	An empty string if a Proj4 map projection is not available for the named coordinate pair, otherwise a properly formatted Proj4 projection string is returned.

virtual int VAPoR::DC::GetNumRefLevels ( string varname ) const

pure virtual

Return the number of refinement levels for the indicated variable

Compressed variables have a multi-resolution grid representation. This method returns the number of levels in the hiearchy. A value of one indicates that only the native resolution is available. A value of two indicates that two levels, the native plus the next coarsest are available, and so on.

Parameters

[in] varname Data or coordinate variable name.

Return values

num	If `varname` is unknown zero is returned. if `varname` is not compressed (has no multi-resolution representation) one is returned. Otherwise the total number of levels in the multi-resolution hierarchy are returned.

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual int VAPoR::DC::GetNumTimeSteps ( string varname ) const

virtual

Return the time dimension length for a variable

Returns the number of time steps (length of the time dimension) for which a variable is defined. If varname does not have a time coordinate 1 is returned. If varname is not defined as a variable a negative int is returned.

Parameters

[in] varname A string specifying the name of the variable.

Return values

count The length of the time dimension, or a negative int if varname is undefined.

See also: IsTimeVarying()

virtual int VAPoR::DC::GetVar	(	string	varname,
		int	level,
		int	lod,
		float *	data
	)

pure virtual

Read an entire variable in one call

This method reads and entire variable (all time steps, all grid points) from a DC. This is the simplest interface for reading data from a DC. If the variable is split across multiple files GetVar() ensures that the data are correctly gathered and assembled into memory Any variables currently opened with OpenVariableRead() are first closed. Thus variables need not be opened with OpenVariableRead() prior to calling GetVar();

It is an error to call this method in define mode

Parameters

[in]	varname	Name of the variable to write
[in]	level	Refinement level of the variable. Ignored if the variable is not compressed.
[in]	lod	Approximation level of the variable. A value of -1 indicates the maximum approximation level defined for the DC. Ignored if the variable is not compressed.
[out]	data	Pointer to where data will be copied. It is the caller's responsbility to ensure `data` points to sufficient memory.

Return values

status A negative int is returned on failure

Implemented in VAPoR::VDC, VAPoR::VDCNetCDF, and VAPoR::DCWRF.

virtual int VAPoR::DC::GetVar	(	size_t	ts,
		string	varname,
		int	level,
		int	lod,
		float *	data
	)

pure virtual

Read an entire variable at a given time step in one call

This method reads and entire variable (all grid points) at time step ts from a DC. This is the simplest interface for reading data from a DC. Any variables currently opened with OpenVariableRead() are first closed. Thus variables need not be opened with OpenVariableRead() prior to calling GetVar();

It is an error to call this method in define mode

Parameters

[in]	ts	Time step of the variable to write. This is the integer offset into the variable's temporal dimension. If the variable does not have a temporal dimension `ts` is ignored.
[in]	varname	Name of the variable to write
[in]	level	Refinement level of the variable. Ignored if the variable is not compressed.
[in]	lod	Approximation level of the variable. A value of -1 indicates the maximum approximation level defined for the DC. Ignored if the variable is not compressed.
[out]	data	Pointer to where data will be copied. It is the caller's responsbility to ensure `data` points to sufficient memory.

Return values

status A negative int is returned on failure

Implemented in VAPoR::VDC, VAPoR::VDCNetCDF, and VAPoR::DCWRF.

virtual int VAPoR::DC::Initialize ( const vector< string > & paths )

pure virtual

Initialize the DC class

Prepare a DC for reading. This method prepares the master DC file indicated by path for reading. The method should be called immediately after the constructor, before any other class methods. This method exists only because C++ constructors can not return error codes.

Parameters

[in] path Path name of file that contains, or will contain, the DC master file for this data collection

Return values

status A negative int is returned on failure

Implemented in VAPoR::DCWRF.

virtual bool VAPoR::DC::IsCompressed ( string varname ) const

virtual

Return a boolean indicating whether a variable is compressed

This method returns true if the variable named by varname is defined and it has a compressed representation. If either of these conditions is not true the method returns false.

Parameters

[in] varname A string specifying the name of the variable.

Return values

bool	Returns true if variable `varname` exists and is compressed

virtual bool VAPoR::DC::IsCoordVar ( string varname ) const

inlinevirtual

Return a boolean indicating whether a variable is a coordinate variable

This method returns true if a coordinate variable is defined with the name varname. Otherwise the method returns false.

Return values

bool	Returns true if `varname` names a defined coordinate variable

Definition at line 1347 of file DC.h.

virtual bool VAPoR::DC::IsDataVar ( string varname ) const

inlinevirtual

Return a boolean indicating whether a variable is a data variable

This method returns true if a data variable is defined with the name varname. Otherwise the method returns false.

Return values

bool	Returns true if `varname` names a defined data variable

Definition at line 1334 of file DC.h.

virtual bool VAPoR::DC::IsTimeVarying ( string varname ) const

virtual

Return a boolean indicating whether a variable is time varying

This method returns true if the variable named by varname is defined and it has a time axis dimension. If either of these conditions is not true the method returns false.

Parameters

[in] varname A string specifying the name of the variable.

Return values

bool	Returns true if variable `varname` exists and is time varying.

virtual int VAPoR::DC::OpenVariableRead	(	size_t	ts,
		string	varname,
		int	level = `0`,
		int	lod = `0`
	)

pure virtual

Open the named variable for reading

This method prepares a data or coordinate variable, indicated by a variable name and time step pair, for subsequent read operations by methods of this class. The value of the refinement levels parameter, level, indicates the resolution of the volume in the multiresolution hierarchy as described by GetDimLensAtLevel().

The level-of-detail parameter, lod, selects the approximation level. Valid values for lod are integers in the range 0..n-1, where n is returned by DC::BaseVar::GetCRatios().size(), or the value -1 may be used to select the best approximation available.

An error occurs, indicated by a negative return value, if the volume identified by the {varname, timestep, level, lod} tupple is not available. Note the availability of a volume can be tested with the VariableExists() method.

Parameters

[in]	ts	Time step of the variable to read. This is the integer offset into the variable's temporal dimension. If the variable does not have a temporal dimension `ts` is ignored.
[in]	varname	Name of the variable to read
[in]	level	Refinement level of the variable. Ignored if the variable is not compressed.
[in]	lod	Approximation level of the variable. A value of -1 indicates the maximum approximation level defined for the DC. Ignored if the variable is not compressed.

Return values

status Returns a non-negative value on success

See also: GetNumRefLevels(), DC::BaseVar::GetCRatios(), OpenVariableRead()

Implemented in VAPoR::VDC, VAPoR::VDCNetCDF, and VAPoR::DCWRF.

static bool VAPoR::DC::ParseDimensions	(	const vector< DC::Dimension > &	dimensions,
		vector< size_t > &	sdims,
		size_t &	numts
	)

static

Parse a vector of DC::Dimensions into space and time dimensions

This is a convenience utility that parses an ordered vector of Dimensions into a vector of spatial lengths, and the number of time steps (if time varying). The dimension vector axis must be ordered: X, Y, Z, T

Parameters

[in]	dimensions	An ordered vector of dimensions
[out]	sdims	Ordered vector of dimension lengths extracted from `dimensions`
[out]	numts	The number of time steps if `dimensions` contains a time dimension, otherwise `numts` will be one

Return values

status A value of true is returned if the dimensions contains a correctly sized and ordered vector of dimensions

virtual int VAPoR::DC::Read ( float * data )

pure virtual

Read all spatial values of the currently opened variable

This method reads, and decompresses as necessary, the contents of the currently opened variable into the array data. The number of values read into data is given by the product of the spatial dimensions of the open variable at the refinement level specified.

It is the caller's responsibility to ensure data points to adequate space.

Parameters

[out] data An array of data to be written

Return values

status Returns a non-negative value on success

See also: OpenVariableRead()

Implemented in VAPoR::VDC, VAPoR::VDCNetCDF, and VAPoR::DCWRF.

virtual int VAPoR::DC::ReadRegion	(	const vector< size_t > &	min,
		const vector< size_t > &	max,
		float *	region
	)

pure virtual

Read in and return a subregion from the currently opened variable

This method reads and returns a subset of variable data. The min and max vectors, whose dimensions must match the spatial rank of the currently opened variable, identify the minimum and maximum extents, in grid coordinates, of the subregion of interest. The minimum and maximum valid values of an element of min or max are 0 and n-1, respectively, where n is the length of the associated dimension at the opened refinement level.

The region returned is stored in the memory region pointed to by region. It is the caller's responsbility to ensure adequate space is available.

Parameters

[in]	min	Minimum region extents in grid coordinates
[in]	max	Maximum region extents in grid coordinates
[out]	region	The requested volume subregion

Return values

status Returns a non-negative value on success

See also: OpenVariableRead(), GetDimension(), GetDimensionNames()

Implemented in VAPoR::VDC, and VAPoR::DCWRF.

virtual int VAPoR::DC::ReadRegionBlock	(	const vector< size_t > &	min,
		const vector< size_t > &	max,
		float *	region
	)

pure virtual

Read in and return a blocked subregion from the currently opened variable.

This method is identical to ReadRegion() with the exceptions that for compressed variables:

The vectors start and count must be aligned with the underlying storage block of the variable. See DC::SetCompressionBlock()

The hyperslab copied to region will preserve its underlying storage blocking (the data will not be contiguous)

Implemented in VAPoR::VDC, VAPoR::VDCNetCDF, and VAPoR::DCWRF.

virtual int VAPoR::DC::ReadSlice ( float * slice )

pure virtual

Read a single slice of data from the currently opened variable

Decompress, as necessary, and read a single slice (2D array) of data from the variable indicated by the most recent call to OpenVariableRead(). The dimensions of a slices are NX by NY, where NX is the dimension of the array along the fastest varying spatial dimension, specified in grid points, and NY is the length of the second fastest varying dimension at the currently opened grid refinement level. See OpenVariableRead().

This method should be called exactly NZ times for each opened variable, where NZ is the dimension of third, and slowest varying dimension. In the case of a 2D variable, NZ is 1.

It is the caller's responsibility to ensure slice points to adequate space.

Parameters

[out] slice A 2D slice of data

Return values

status Returns a non-negative value on success

See also: OpenVariableRead()

Implemented in VAPoR::VDC, VAPoR::VDCNetCDF, and VAPoR::DCWRF.

virtual bool VAPoR::DC::VariableExists	(	size_t	ts,
		string	varname,
		int	reflevel = `0`,
		int	lod = `0`
	)		const

pure virtual

Returns true if indicated data volume is available

Returns true if the variable identified by the timestep, variable name, refinement level, and level-of-detail is present in the data set. Returns false if the variable is not available.

Parameters

[in]	ts	A valid time step between 0 and GetNumTimesteps()-1
[in]	varname	A valid variable name
[in]	reflevel	Refinement level requested.
[in]	lod	Compression level of detail requested. refinement level contained in the DC.

Implemented in VAPoR::VDC, VAPoR::VDCNetCDF, and VAPoR::DCWRF.

The documentation for this class was generated from the following file:

/Users/clyne/src/vapor-3.0/include/vapor/DC.h

Classes

Public Types

Public Member Functions

Static Public Member Functions

Additional Inherited Members

Detailed Description

Member Enumeration Documentation

Constructor & Destructor Documentation

Member Function Documentation