Class table3d (o2scl)¶
-
class table3d¶
A data structure containing one or more slices of two-dimensional data points defined on a grid.
Todo
In class table3d:
Future: Improve interpolation and derivative caching, possibly through non-const versions of the interpolation functions.
Future: Should there be a clear_grid() function separate from clear_data() and clear()?
Future: Allow the user to more clearly probe ‘size_set’ vs. ‘xy_set’? (AWS 07/18: This is apparently resolved.)
Iterator types
-
typedef std::map<std::string, size_t, std::greater<std::string>>::iterator map_iter¶
-
typedef std::map<std::string, size_t, std::greater<std::string>>::const_iterator map_const_iter¶
Data storage
-
std::map<std::string, double> constants¶
The list of constants.
-
size_t numx¶
The size of the x grid.
-
size_t numy¶
The size of the y grid.
-
std::map<std::string, size_t, std::greater<std::string>> tree¶
A tree connecting column names to list indexes.
-
std::string xname¶
The name for the x grid.
-
std::string yname¶
The name for the y grid.
-
bool xy_set¶
True if the grid has been set.
-
bool size_set¶
True if the size of the grid has been set.
-
bool has_slice¶
True if the table has at least one slice.
-
size_t itype¶
The interpolation type.
Initialization
-
template<class vec_t, class vec2_t>
inline void set_xy(std::string x_name, size_t nx, const vec_t &x, std::string y_name, size_t ny, const vec2_t &y)¶ Initialize the x-y grid.
This function will not allow you to redefine the grid when there is data in the if a grid of a different size was already set from a previous call to either set_xy() or set_size(). However, you may freely redefine the grid after a call to clear_data() or clear_table(). You may change individual grid points at any time with set_grid_x() and set_grid_y().
-
void set_xy(std::string x_name, uniform_grid<double> gx, std::string y_name, uniform_grid<double> gy)¶
Initialize the x-y grid with uniform_grid objects.
This function will not allow you to redefine the grid when there is data in the if a grid of a different size was already set from a previous call to either set_xy() or set_size(). However, you may freely redefine the grid after a call to clear_data() or clear_table(). You may change individual grid points at any time with set_grid_x() and set_grid_y().
-
void set_size(size_t nx, size_t ny)¶
Initialize size.
This function will not allow you to resize the if it already has data or if the size has already been set with the set_xy() function, unless you clear the data with clear_data() or the with clear_table() first.
On-grid get and set methods
-
void set(size_t ix, size_t iy, std::string name, double val)¶
Set element in slice
name
at locationix,iy
to valueval
.
-
void set(size_t ix, size_t iy, size_t z, double val)¶
Set element in slice of index
z
at locationix,iy
to valueval
.
-
double &get(size_t ix, size_t iy, std::string name)¶
Get element in slice
name
at locationix,iy
-
const double &get(size_t ix, size_t iy, std::string name) const¶
Get element in slice
name
at locationix,iy
(const version)
-
double &get(size_t ix, size_t iy, size_t z)¶
Get element in slice of index
z
at locationix,iy
-
const double &get(size_t ix, size_t iy, size_t z) const¶
Get element in slice of index
z
at locationix,iy
(const version)
Off-grid get and set methods
These methods return the value of a slice on the grid point nearest to a user-specified location. For interpolation into a point off the grid, use table3d::interp().
-
void set_val(double x, double y, std::string name, double val)¶
Set element in slice
name
at the nearest location tox,y
to valueval
.
-
void set_val(double x, double y, size_t z, double val)¶
Set element in slice of index
z
at the nearest location tox,y
to valueval
.
-
double &get_val(double x, double y, std::string name)¶
Get element in slice
name
at location closest tox,y
-
const double &get_val(double x, double y, std::string name) const¶
Get element in slice
name
at location closest tox,y
-
double &get_val(double x, double y, size_t z)¶
Get element in slice of index
z
at location closest tox,y
-
const double &get_val(double x, double y, size_t z) const¶
Get element in slice of index
z
at location closest tox,y
Off-grid get and set methods returning nearest point
-
void set_val_ret(double &x, double &y, std::string name, double val)¶
Set element in slice
name
at the nearest location tox,y
to valueval
.
-
void set_val_ret(double &x, double &y, size_t z, double val)¶
Set element in slice of index
z
at the nearest location tox,y
to valueval
.
-
double &get_val_ret(double &x, double &y, std::string name)¶
Get element in slice
name
at location closest tox,y
, and also return the corresponding values ofx
andy
.
-
const double &get_val_ret(double &x, double &y, std::string name) const¶
Get element in slice
name
at location closest tox,y
, and also return the corresponding values ofx
andy
.
-
double &get_val_ret(double &x, double &y, size_t z)¶
Get element in slice of index
z
at location closest tox,y
, and also return the corresponding values ofx
andy
.
-
const double &get_val_ret(double &x, double &y, size_t z) const¶
Get element in slice of index
z
at location closest tox,y
, and also return the corresponding values ofx
andy
.
-
void add_slice_from_table(table3d &source, std::string slice, std::string dest_slice = "", int verbose = 0)¶
This function adds a slice from a different table3d object, interpolating the results into the current table3d object.
Grid information get and set methods
-
void set_grid_x(size_t ix, double val)¶
Set x grid point at index
ix
.
-
void set_grid_y(size_t iy, double val)¶
Set y grid point at index
iy
.
-
double get_grid_x(size_t ix) const¶
Get x grid point at index
ix
.
-
double get_grid_y(size_t iy) const¶
Get y grid point at index
iy
.
-
std::string get_x_name() const¶
Get the name of the x grid variable.
-
std::string get_y_name() const¶
Get the name of the y grid variable.
-
void set_x_name(std::string name)¶
Set the name of the x grid variable.
-
void set_y_name(std::string name)¶
Set the name of the y grid variable.
Size get methods
-
void get_size(size_t &nx, size_t &ny) const¶
Get the size of the slices.
-
size_t get_nx() const¶
Get the x size.
-
size_t get_ny() const¶
Get the y size.
-
size_t get_nslices() const¶
Get the number of slices.
-
bool is_size_set() const¶
True if the size of the table has been set.
-
bool is_xy_set() const¶
True if the grid has been set.
Slice manipulation
-
void line_of_names(std::string names)¶
Create a set of new slices specified in the string
names
.
-
std::string get_slice_name(size_t z) const¶
Returns the name of slice with index
z
.
-
void new_slice(std::string name)¶
Add a new slice.
-
void set_slice_all(std::string name, double val)¶
Set all of the values in slice
name
toval
.
-
size_t lookup_slice(std::string name) const¶
Find the index for slice named
name
.
-
bool is_slice(std::string name, size_t &ix) const¶
Return true if slice is already present.
-
void rename_slice(std::string olds, std::string news)¶
Rename slice named
olds
tonews
.This is slow since we have to delete the column and re-insert it. This process in turn mangles all of the iterators in the list.
-
void delete_slice(std::string sl)¶
-
void copy_slice(std::string src, std::string dest)¶
Make a new slice named
dest
which is a copy of the slice with name given insrc
.
-
void init_slice(std::string slice, double val)¶
Initialize all values of slice named
scol
toval
.Note
This will call the error handler if the value
val
is not finite (i.e. eitherInf
orNaN
).
Lookup and search methods
-
void lookup_x(double val, size_t &ix) const¶
Look for a value in the x grid.
-
void lookup_y(double val, size_t &iy) const¶
Look for a value in the y grid.
-
void lookup(double val, std::string slice, size_t &ix, size_t &iy) const¶
Look for a value in a specified slice.
Interpolation, differentiation, and integration
-
void set_interp_type(size_t interp_type)¶
Specify the interpolation type.
-
size_t get_interp_type() const¶
Get the interpolation type.
-
double interp(double x, double y, std::string name) const¶
Interpolate
x
andy
in slice namedname
.
-
double deriv_x(double x, double y, std::string name) const¶
Interpolate the derivative of the data with respect to the x grid at point
x
andy
in slice namedname
.
-
double deriv_y(double x, double y, std::string name) const¶
Interpolate the derivative of the data with respect to the y grid at point
x
andy
in slice namedname
.
-
double deriv_xy(double x, double y, std::string name) const¶
Interpolate the mixed second derivative of the data at point
x
andy
in slice namedname
.
-
double integ_x(double x1, double x2, double y, std::string name) const¶
Interpolate the integral of the data respect to the x grid.
-
double integ_y(double x, double y1, double y2, std::string name) const¶
Interpolate the integral of the data respect to the y grid.
-
template<class vec_t>
inline void interp_slices(double x, double y, size_t nv, vec_t &v)¶ Fill a vector of interpolated values from each slice at the point
x,y
-
void deriv_x(std::string fname, std::string fpname)¶
Create a new slice, named
fpname
, containing the derivative offname
with respect to the x coordinate.
-
void deriv_y(std::string fname, std::string fpname)¶
Create a new slice, named
fpname
, containing the derivative offname
with respect to the y coordinate.
Extract 2-dimensional tables
Clear methods
-
void zero_table()¶
Zero the data entries but keep the slice names and grid.
-
void clear()¶
Clear everything.
-
void clear_data()¶
Remove all of the data by setting the number of lines to zero.
This leaves the column names intact and does not remove the constants.
Summary method
-
void summary(std::ostream *out = 0, int ncol = 79) const¶
Output a summary of the information stored.
Outputs the number of constants, the grid information, and a list of the slice names
Contour lines method
-
template<class vec_t>
inline void slice_contours(std::string name, size_t nlev, vec_t &levs, std::vector<contour_line> &clines)¶ Create contour lines from the slice named
name
.This uses contour to compute contour lines (stored in
clines
) from slicename
givennlev
contour levels inlevs
.
Manipulating constants
-
virtual void add_constant(std::string name, double val)¶
Add a constant, or if the constant already exists, change its value.
-
virtual void remove_constant(std::string name)¶
Remove a constant.
-
virtual int set_constant(std::string name, double val, bool err_on_notfound = true)¶
Set a constant equal to a value, but don’t add it if not already present.
If
err_on_notfound
istrue
(the default), then this function throws an exception if a constant with namename
is not found. Iferr_on_notfound
isfalse
, then if a constant with namename
is not found this function just silently returns o2scl::exc_enotfound.
-
virtual bool is_constant(std::string name) const¶
Test if
name
is a constant.
-
virtual double get_constant(std::string name)¶
Get a constant.
-
virtual void get_constant(size_t ix, std::string &name, double &val) const¶
Get a constant by index.
-
inline virtual size_t get_nconsts() const¶
Get the number of constants.
Miscellaneous methods
-
int read_gen3_list(std::istream &fin, int verbose = 0, double eps = 1.0e-12)¶
Read a generic table3d object specified as a text file.
This function reads a set of columns of numerical values, presuming that the first column is the x-grid value, the second column is the y-grid value, and the remaining columns are slices to be added. If the first row appears to be strings rather than numerical values, then the first row is used for the x name, y name, and slice names. Values in the first two columns which differ by less than
eps
are assumed to refer to the same grid point. If not all combinations of x and y are found, then those entries are left unchanged in all slices.\verbatim embed:rst .. todo:: In class table3d::read_gen3_list(): Future: It would be great to add a function which generates a text file in this format as well. Future: This function is very similar to from_table() below, so it might be possible just to avoid code duplication between these two functions. \endverbatim
-
template<class vec_t>
inline int read_table(const o2scl::table<vec_t> &tab, std::string xname2 = "", std::string yname2 = "", double empty_value = 0.0, int verbose = 0, bool err_on_fail = true, double eps = 1.0e-12)¶ Set the current table3d object by reading a o2scl::table.
The function reads the table \c tab and attempts to convert it to a table3d object by using column \c xname2 and \c yname2 as the values for the x- and y-grids. This function is particularly useful for a table which has the following structure: \verbatim x y z 1 1 4.5 1 2 2.0 1 3 1.6 2 1 1.5 2 3 4.7 3 1 3.9 3 2 4.5 3 3 4.8 \endverbatim If \c xname2 or \c yname2 are empty strings, then the first or second column is presumed to hold the value for the x- or y-grid, respectively. The grids in the table3d object are set to include all possible values in the associated column, treating any values with a relative deviation of \c eps as identical. In the example above, using "x" and "y" as the columns for the table3d grid, the grids would both be [1,2,3]. New slices are created in the table3d for each column in the table object which is not part of the grid. Any current data stored in the table3d object is destroyed. All slices are filled with \c empty_value before being assigned values from \c tab. For example, in the table above, the slice named "z" in the new table3d object would have a final entry of 0.0 for (x,y) at (2,2) because there is no entry in the original table for that location.
-
template<class vec_t, class vec2_t, class vec3_t>
inline void create_table_hist_set(vec_t &grid, std::string direction, std::string name, vec2_t &bin_edges, vec3_t &bin_grid, std::string bin_name, o2scl::table_units<> &t, std::string pattern, std::string slice, bool use_regex = false, int verbose = 0)¶ Create a table3d object by histogramming a series of columns from a o2scl::table_units object.
Create a new table3d object from a histogram a series of columns from a table_units object. If
direction
is “x”, then arrange these histograms “vertically”, so that the x-coordinate (namedname
) of the ith column is taken from the ith entry ofgrid
. Ifdirection
is “y”, then arrange these histograms “horizontally”, so that the y-coordinate (namedname
) of the ith column is taken from the ith entry ofgrid
. The histograms (in either case) are all created using the bin edges frombin_edges
. Whendirection
is “x” (“y”),bin_grid
is used for the y-coordinate (x-coordinate) of the new table3d object. This coordinate is namedbin_name
. The columns are taken from all those columns int
which match the regular expression inpattern
. All of the new histogram data is imported into a slice namedslide
in a new table3d object. oneThe vector
bin_grid
must have a size which is exactly 1 smaller than the size of the vectorbin_edges
. The number of columns matched from tablet
by the pattern specified inpattern
must be exactly equal to the size of the vectorgrid
.Any data in the current table3d object is destroyed.
-
template<class vec_t>
inline void create_table_hist_set_minmax(vec_t &grid, std::string direction, std::string name, size_t n_bins, std::string bin_name, o2scl::table_units<> &t, std::string pattern, std::string slice, double factor = 1.0e-10, bool use_regex = false, int verbose = 0)¶ Create a table3d object by histogramming a series of columns from a o2scl::table_units object.
This function works very similarly to the more detailed function with the same name, but it uses the minimum and maximum values of the table columns in order to automatically create the histogram bin edges from a set of
n_bins
bins. It uses these bin edges to create thebin_grid
andbin_edges
objects.
-
template<class vec_t, class vec2_t>
inline void create_table_hist_set_edgeonly(vec_t &grid, std::string direction, std::string name, vec2_t &bin_edges, std::string bin_name, o2scl::table_units<> &t, std::string pattern, std::string slice, bool use_regex = false, int verbose = 0)¶ Create a table3d object by histogramming a series of columns from a o2scl::table_units object.
This function works very similarly to the more detailed function with the same name, but it uses
bin_edges
to automatically compute thebin_grid
argument. Ifbin_edges
appears logarithmic, it uses the geometric mean of adjacent edges, otherwise it uses the arithmetic mean.
-
inline virtual const char *type()¶
Return the type,
"table3d"
.
Parsing mathematical functions specified as strings
-
template<class resize_mat_t>
inline int function_matrix(std::string function, resize_mat_t &mat, bool throw_on_err = true)¶ Fill a matrix from the function specified in
function
.
-
void function_slice(std::string function, std::string col)¶
Make a column from
function
and add it to the table.If the column already exists, the data already present is overwritten with the result.
-
table3d slice_to_uniform_grid(std::string slice, size_t xpts, bool log_x, size_t ypts, bool log_y)¶
Copy slice named
slice
to a new o2scl::table3d object with a uniform grid using the current interpolation type.
-
table3d table_to_uniform_grid(size_t xpts, bool log_x, size_t ypts, bool log_y)¶
Copy entire table to a new o2scl::table3d object with a uniform grid using the current interpolation type.
Tree iterator boundaries
-
inline map_const_iter const_begin() const¶
Return the beginning of the slice tree.
-
inline map_const_iter const_end() const¶
Return the end of the slice tree.
Public Types
-
typedef boost::numeric::ublas::vector<double> ubvector¶
-
typedef boost::numeric::ublas::matrix<double> ubmatrix¶
Public Functions
-
table3d()¶
Create a new 3D .
-
virtual ~table3d()¶
-
table3d(o2scl::table_units<> &t, std::string colx, std::string coly)¶
Create a table3d object from a table, assuming
scolx
andscoly
store the x- and y-grid data, respectively.