|
PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"DTD/xhtml1-strict.dtd">
---------------------------------------------
RubyNetCDF is the Ruby interface of the NetCDF library. Ruby is a free
object-oriented scripting language and is available from
http://www.ruby-lang.org/. To handle numeric data, RubyNetCDF uses
NArray, which is the standard numeric multi-dimensional array class
for Ruby (http://www.ruby-lang.org/en/raa-list.rhtml?name=NArray).
An NArray object holds numeric data in a consecutive memory area
pointed by a C pointer. Thus, it is computationally efficient.
NArray is similar to NumPy for Python, but results of some benchmark
tests suggests that NArray is more efficient than NumPy.
RubyNetCDF consists of the four class in the following.
class NetCDF -- the file class
An NetCDF object represents a NetCDF file
class NetCDFDim -- the dimension class
Although in its C version a NetCDF dimension is represented by a
combination of a file ID and a dimension ID, it is represented by
only one NetCDFDim object in RubyNetCDF.
class NetCDFVar -- the variable class
Although in its C version a NetCDF variable is represented by a
combination of a file ID and a variable ID, it is represented by
only one NetCDFVar object in RubyNetCDF.
class NetCDFAtt -- the attribute class
Although in its C version a NetCDF attribute is represented by a
combination of file ID, variable ID, and its name, it is represented
by only one NetCDFAtt object in RubyNetCDF.
All the NetCDF variable types char, byte, short, int, float, and
double are supported in this Ruby interface. These types are called,
however, differently in it to adhere to the convention of Ruby, or,
more specifically, of NArray. These types are named to as "char",
"byte", "sint", "int", "sfloat", and "float", respectively. Therefore,
the vartype (=ntype) method of the NetCDFVar class returns one of these
strings. The def_var method of the NetCDF class also accepts one of
them to define a variable. It should be noted especially that "float"
in this library means the double in the NetCDF terminology. This is
due to the convention of Ruby -- the predefined Float class
corresponds to the double in C, not the float.
The "get" method of NetCDFVar class reads a variable in a NArray of
the same type as in the file, except for the "char" type which is read
into a "byte". This is because NArray does not have a "char" type.
However, it not is not supposed to be a problem, since a byte NArray
can easily handle string data.
Errors are basically handled by raising exceptions. However, light
errors in value-returning methods are handled by returning nil (e.g.,
if a non-existent attribute name is specified in attribute reading).
Those methods that returns nil on error are explicitly written as such
in the following.
Security handling is done just as in the pre-defined File class.
To use the RubyNetCDF library, load the library first by placing the
following line in your Ruby program to load the library:
require 'numru/netcdf'
Here, 'numru', which is taken from "Numerical Ruby", is the name of
the subdirectory in the user's load path where the RubyNetCDF library
is placed (by default). Then, it can be used as in the following:
file = NumRu::NetCDF.create('tmp.nc')
x = file.def_dim('x',10)
y = file.def_dim('y',10)
v = file.def_var('v','float',[x,y])
file.close
Here, NumRu is the module that has the library in it. The
reason why NetCDF library is wrapped in such a module is to avoid
conflicts in the name space. Without this kind of treatment,
problems happen if the user wants to use a library that happens to
have a class or module with the same name as even one of the classes
in this library.
If such a problem is not expected to happen, the prefix "NumRu::" can
be eliminated by "including" the NumRu module as in the following, so
that to place "NumRu::" is not needed anymore in the current scope:
include NumRu
file = NetCDF.create('tmp.nc')
...
For more examples, see demo and test programs included in the
distribution package.
---------------------------------------------
method_name(argument1, argument2, ...) -- arguments that can be omitted are expressed as Argument_name=Default_value
-
Explanation of its function
Arguments
- name of argument1 (its class or possible values): explanation
- name of argument2 (its class or possible values): explanation
- ...
Return value
- Explanation of the return value
Corresponding (dependent) function(s) in the C library of NetCDF
- Name(s) in NetCDF ver 3. The function equivalent to the current
method, if not in parenthesis. If no direct correspondence,
dependent functions are listed in parentheses.
---------------------------------------------
class NetCDF
Class Methods
Instance Methods
- close Closes the file.
- ndims Returns the number of dimensions in the file
- nvars Returns the number of variables in the file
- natts Returns the number of global attributes in the file
- unlimited Returns the unlimited dimension in the file
- path Returns the path of the file (contents of the filename specified when opened/created)
- redef Switches to the define mode. Does nothing if already in it.
- enddef Switches to the data mode. Does nothing if already in it.
- sync Synchronizes the disk copy of a netCDF dataset with in-memory buffer
- def_dim Define a dimension
- put_att Sets a global attribute
- def_var Defines a variable
- def_var_with_dim Same as def_var but defines dimensions first if needed
- var Opens an existing variable in the file
- dim Opens an existing dimension in the file
- att Opens an existing global attribute in the file
- fill= Sets a fill mode. (Default behavior of NetCDF is FILL.)
- each_dim Iterator regarding the dimensions in the file.
- each_var Iterator regarding the variables in the file.
- each_att Iterator regarding the global attributes of the file.
- dim_names Returns the names of all dimensions in the file
- var_names Returns the names of all variables in the file
- att_names Returns the names of all the global attributes of the file
class NetCDFDim
Class Methods
Instance Methods
- length Returns the length of the dimension
- length_ul0 Same as length but returns 0 for the unlimited dimension
- name= Rename the dimension
- name Returns the name of the dimension
- unlimited? Inquires whether the dimension is unlimited or not
class NetCDFVar
Class Methods
- NetCDFVar.new Combines NetCDF.open and NetCDF#Var to open a variable with one line (no need to use this).
Instance Methods
- dim Inquires the dim_num-th dimension of the variable (dim_num=0,1,2,..)
- dims Returns an array of all the dimensions of the variable
- shape_ul0 Returns the shape of the variable, but the length of the unlimited dimension is set to zero.
- shape_current Returns the current shape of the variable.
- url Returns a combination of filename (path) and variable name as path+'?var='+varname
- each_att Iterator regarding the global attributes of the variables.
- dim_names Returns the names of all the dimensions of the variable
- att_names Returns the names of all the attributes of the variable
- name Returns the name of the variable
- name= Rename the variable
- ndims Number of dimensions of the variable (which is rank of the variable).
- ntype Aliased to vartype
- vartype Data value type of the variable
- natts Returns the number of the attributes of the variable
- file Inquires the file that the variable is in
- att Returns the attribute specified by its name
- put_att Sets an attribute
- put Set the values of the variable
- scaled_put Same as put but interprets the attributes scale_factor and add_offset
- get Returns values of the variable
- scaled_get Same as get but interprets the attributes scale_factor and add_offset
class NetCDFAtt
Class Methods
Instance Methods
- name Returns the name of the attribute
- name= Rename the attribute
- copy Copies an attribute to a variable or a file. If file, becomes an global attribute
- delete Delete an attribute
- put Sets the value of the attribute
- get Returns the values of the attribute
- atttype Inquires the type of attribute values
---------------------------------------------
NetCDF.open(filename, mode="r", share=false)
-
Opens a file (class method). If mode="w" and non-existent, a new
file is created.
Arguments
- filename (String): file name (path)
- mode (String) : IO mode "r"(read only), "w"(writable)
- share (true or false) : Whether to use the "shared" mode or not
(set true if a file being written may be read from other
processes. See nc_open in Ch.5 of users' guide of the C version)
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
NetCDF.new
-
Aliased to NetCDF.open
NetCDF.create(filename, noclobber=false, share=false)
-
Creates a NetCDF file (class method)
Arguments
- filename (String) : file name (path)
- noclobber (true or false) : overwrite or not if the file exists
- share (true or false) : Whether to use the shared mode or not
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
NetCDF.create_tmp(tmpdir=ENV['TMPDIR']||ENV['TMP']||ENV['TEMP']||'.', share=false)
-
Creates a temporary NetCDF file (class method).
Its name is automatically generated, and it is deleted when closed.
Arguments
- tmpdir (String) : directory to place the temporary file.
By default, "." or a directory specified by an environmental
vairable (TMPDIR or TMP oe TEMP) is used. In a secure mode,
theses environmental variable is NOT used, and the default
value is '.'.
- share (true or false) : Whether to use the shared mode or not
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
close
-
Closes the file.
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
ndims
-
Returns the number of dimensions in the file
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
nvars
-
Returns the number of variables in the file
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
natts
-
Returns the number of global attributes in the file
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
unlimited
-
Returns the unlimited dimension in the file
Arguments
Return value
- a NetCDFDim if it exists; nil if not
Corresponding (dependent) function(s) in the C library of NetCDF
path
-
Returns the path of the file (contents of the filename specified when opened/created)
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
redef
-
Switches to the define mode. Does nothing if already in it.
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
enddef
-
Switches to the data mode. Does nothing if already in it.
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
sync
-
Synchronizes the disk copy of a netCDF dataset with in-memory buffer
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
def_dim(dimension_name, length)
-
Define a dimension
Arguments
- dimension_name (String) : Name of the dimension to be defined
- length (Integer) : length of the dimension. 0 for unlimited.
Return value
- defined dimension (NetCDFDim object)
Corresponding (dependent) function(s) in the C library of NetCDF
put_att(attribute_name, value, atttype=nil)
-
Sets a global attribute
Arguments
- attribute_name (String) : name of the global attribute
- value (Numeric, String, Array of Numeric, or NArray) : value of the attribute
- atttype (nil or String) : data type of the attribute value.
nil lets it automatically determined from the value.
"char" (or "string"), "byte", "sint", "int", "sfloat", or "float"
specifies the type explicitly (1,1,2,4,4,8 bytes, respectively)
Return value
- created attribute (NetCDFAtt object)
Corresponding (dependent) function(s) in the C library of NetCDF
def_var(name, vartype, dimensions)
-
Defines a variable
Arguments
- name (String) : Name of the variable to define
- vartype (String) : data type of the variable ("char", "byte", "sint",
"sint", "int", "sfloat", or "float")
- dimensions (Array) : Dimensions of the variable. An Array of
NetCDFDim, in the order from the fastest varying dimension to
the slowliest varying one; its length becomes the rank of the
variable.
Return value
- defined variable (NetCDFVar object)
Corresponding (dependent) function(s) in the C library of NetCDF
def_var_with_dim(name, vartype, shape_ul0, dimnames)
-
Same as def_var but defines dimensions first if needed.
Raise exception if it conflicts with the lengths of exisiting dimensions.
Arguments
- name (String) : Name of the variable to define
- vartype (String) : data type of the variable ("char", "byte", "sint",
"sint", "int", "sfloat", or "float")
- shape_ul0 (Array of Integer) : Shape of the variable, i.e.,
lengths of dimensions. The unlimited dimension is specified by zero.
The length of shape_ul0 determines the rank of the variable.
- dimnames (Array of String) : Names of the dimensions. Its length
(=>rank) must be equal to that of shape_ul0
Return value
- defined variable (NetCDFVar object)
Corresponding (dependent) function(s) in the C library of NetCDF
var(var_name)
-
Opens an existing variable in the file
Arguments
- var_name (String) : Name of the variable to open
Return value
- a NetCDFVar object if the variable exists; nil if not
Corresponding (dependent) function(s) in the C library of NetCDF
dim(dimension_name)
-
Opens an existing dimension in the file
Arguments
- dimension_name (String) : Name of the dimension to open
Return value
- a NetCDFDim object if the dimension exists; nil if not
Corresponding (dependent) function(s) in the C library of NetCDF
att(attribute_name)
-
Opens an existing global attribute in the file
Arguments
- attribute_name (String) : Name of the global attribute to open
Return value
- a NetCDFAtt object if the attribute exists; nil if not
Corresponding (dependent) function(s) in the C library of NetCDF
- (nc_inq_attid used for inquiry)
fill=(filemode)
-
Sets a fill mode. (Default behavior of NetCDF is FILL.)
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
each_dim{ ... }
-
Iterator regarding the dimensions in the file.
Ex.: {|i| print i.name,"\n"} prints names of all dimensions
Arguments
- { ... } : Block for the iterator. A "do end" block is the alternative.
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- (dependent on nc_inq_ndims)
each_var{ ... }
-
Iterator regarding the variables in the file.
Ex.: {|i| print i.name,"\n"} prints names of all variables
Arguments
- { ... } : Block for the iterator. A "do end" block is the alternative.
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- (dependent on nc_inq_nvars)
each_att{ ... }
-
Iterator regarding the global attributes of the file.
Ex.: {|i| print i.name,"\n"} prints names of all of them.
Arguments
- { ... } : Block for the iterator. A "do end" block is the alternative.
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- (dependent on nc_inq_natts, nc_inq_attname)
dim_names
-
Returns the names of all dimensions in the file
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- (nc_inq_ndims, nc_inq_dimname)
var_names
-
Returns the names of all variables in the file
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- (dependent on nc_inq_nvars, nc_inq_varname)
att_names
-
Returns the names of all the global attributes of the file
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- (dependent on nc_inq_natts, nc_inq_attname)
---------------------------------------------
length
-
Returns the length of the dimension
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
length_ul0
-
Same as length but returns 0 for the unlimited dimension
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
name=(dimension_newname)
-
Rename the dimension
Arguments
- dimension_newname (String) : new name
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
name
-
Returns the name of the dimension
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
unlimited?
-
Inquires whether the dimension is unlimited or not
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- (dependent on nc_inq_unlimdim)
---------------------------------------------
NetCDFVar.new(file,varname,mode="r",share=false)
-
open a NetCDF variable. This can also be done with NetCDF#var
(instance method of NetCDF class), which is recommended over
this method.
Arguments
- file (NetCDF or String) : a NetCDF file object (NetCDF)
or the path of a NetCDF file (String).
- varname (String) : name of the variable in the file
- mode (String) : IO mode "r"(read only), "w"(writable)
- share (true or false) : Whether to use the "shared" mode or not
(set true if a file being written may be read from other
processes. See nc_open in Ch.5 of users' guide of the C version)
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- (dependent on nc_open, nc_create, nc_inq_varid etc.)
dim(dim_num)
-
Inquires the dim_num-th dimension of the variable (dim_num=0,1,2,..)
Arguments
- dim_num (Fixnum) : 0,1,... 0 is the fastest varying dimension.
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- (dependent on nc_inq_vardimid)
dims
-
Returns an array of all the dimensions of the variable
Arguments
Return value
- Array of NetCDFDim objects.
Corresponding (dependent) function(s) in the C library of NetCDF
shape_ul0
-
Returns the shape of the variable, but the length of the unlimited dimension is set to zero.
Good to define another variable.
Arguments
Return value
- Array. [length of 0th dim, length of 1st dim,.. ]
Corresponding (dependent) function(s) in the C library of NetCDF
- (dependent on nc_inq_vardimid, nc_inq_unlimdim etc)
shape_current
-
Returns the current shape of the variable.
Arguments
Return value
- Array. [length of 0th dim, length of 1st dim,.. ]
Corresponding (dependent) function(s) in the C library of NetCDF
- (dependent on nc_inq_vardimid etc)
each_att{ ... }
-
Iterator regarding the global attributes of the variables.
Ex.: {|i| print i.name,"\n"} prints names of all of them.
Arguments
- { ... } : Block for the iterator. A "do end" block is the alternative.
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- (dependent on nc_inq_natts, nc_inq_attname)
dim_names
-
Returns the names of all the dimensions of the variable
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- (dependent on nc_inq_varndims, nc_inq_vardimid, nc_inq_dimname)
att_names
-
Returns the names of all the attributes of the variable
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- (dependent on nc_inq_natts, nc_inq_attname)
name
-
Returns the name of the variable
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
name=(variable_newname)
-
Rename the variable
Arguments
- variable_newname (String) : new name
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
ndims
-
Number of dimensions of the variable (which is rank of the variable).
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
ntype
-
Aliased to vartype
vartype
-
Data value type of the variable
Arguments
Return value
- String ("char","byte","sint","int","sfloat", or "float")
Corresponding (dependent) function(s) in the C library of NetCDF
natts
-
Returns the number of the attributes of the variable
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
file
-
Inquires the file that the variable is in
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
att(attribute_name)
-
Returns the attribute specified by its name
Arguments
- attribute_name (String) : Name of the attribute
Return value
- a NetCDFAtt object if the attribute exists; nil if not
Corresponding (dependent) function(s) in the C library of NetCDF
- (nc_inq_attid is used for inquiry)
put_att(attribute_name, value, atttype=nil)
-
Sets an attribute
Arguments
- attribute_name (String) : name of the attribute
- value (Numeric, String, Array of Numeric, or NArray) : value of the attribute
- atttype (nil or String) : data type of the attribute value.
nil lets it automatically determined from the value.
"char" (="string"), "byte", "sint", "int", "sfloat", or "float"
specifies the type explicitly (1,1,2,4,4,8 bytes, respectively)
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
put(value, option=nil)
-
Set the values of the variable
Arguments
- value : value to set (Numeric, Array of Numeric (1D only), or
NArray (possibly multi-D))
option (Hash) : Optional argument to limit the portion of the
variable to output values. If omitted, the whole variable is
subject to the output. This argument accepts a Hash whose keys
contain either "index" or a combination of "start","end", and
"stride". The value of "index" points the index of a scalar
portion of the variable. The other case is used to designate a
regularly ordered subset, where "start" and "end" specifies
bounds in each dimension and "stride" specifies intervals in
it. As in Array "start", "end", and "index" can take negative
values to specify index backward from the end. However,
"stride" has to be positive, so reversing the array must be
done afterwards if you like.
Example: If the variable is 2D:
{"start"=>[2,5],"end"=>[6,-1],"stride"=>[2,4]} -- Specifies a
subset made as follows: the 1st dimension from the element 2
to the element 6 (note that the count starts with 0, so that
the element 2 is the 3rd one) with an interval of 2;
the 2nd dimension from the element 6 to the last element
(designated by -1) with an interval of 5.
{"index"=>[0,0]}: Scalar of the fist element
{"index"=>[0,-2]}: Scalar from the 1st element of with
respect to the 1st dimension and the 2nd element from the last
with respect to the 2nd dimension
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- nc_put_var_<type>, nc_put_vars_<type>, nc_put_var1_<type>
scaled_put(value, option=nil)
-
Same as put but interprets the attributes scale_factor and add_offset
If scale_factor and/or add_offest is defined, does
(value-add_offset)/scale_factor before writing. (If not, exactly
the same as put.)
If one of the attributes is not defined, assumes 1 for the
factor and 0 for the offset.
See the document for put for arguments etc.
get(option=nil)
-
Returns values of the variable
Arguments
- option (Hash) : Optional argument to limit the portion of the
variable to get values. Its usage is the same as in the method
put.
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
- nc_get_var_<type>, nc_get_vars_<type>, nc_get_var1_<type>
scaled_get(option=nil)
-
Same as get but interprets the attributes scale_factor and add_offset
If scale_factor and/or add_offest is defined, returns
value*scale_factor+add_offset. (If not, exactly the same as get.)
If one of the attributes is not defined, assumes 1 for the
factor and 0 for the offset. The variable type of the return
value (NArray) is float (which is double precision) if the NetCDF
variable is int or float; else, it is sfloat (single presision).
See the document for get for arguments etc.
---------------------------------------------
name
-
Returns the name of the attribute
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
name=(attribute_newname)
-
Rename the attribute
Arguments
- attribute_newname (String) : New name
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
copy(var_or_file)
-
Copies an attribute to a variable or a file. If file, becomes an global attribute
Arguments
- var_or_file (NetCDFVar or NetCDF)
Return value
- Resultant new attribute (NetCDFAtt)
Corresponding (dependent) function(s) in the C library of NetCDF
delete
-
Delete an attribute
Arguments
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
put(value, atttype=nil)
-
Sets the value of the attribute
Arguments
- value (Numeric, String, Array of Numeric, or NArray) : value of the attribute
- atttype (nil or String) : data type of the attribute value.
nil lets it automatically determined from the value.
"char" (="string"), "byte", "sint", "int", "sfloat", or "float"
specifies the type explicitly (1,1,2,4,4,8 bytes, respectively)
Return value
Corresponding (dependent) function(s) in the C library of NetCDF
get
-
Returns the values of the attribute
Arguments
Return value
- String or an NArray object (NOTE: even a scalar is returned as
an NArray of length 1)
Corresponding (dependent) function(s) in the C library of NetCDF
atttype
-
Inquires the type of attribute values
Arguments
Return value
- "char","byte","sint","int","sfloat","float"
Corresponding (dependent) function(s) in the C library of NetCDF
|