cma.logger

module documentation

logger class mainly to be used with CMAEvolutionStrategy

Class	`CMADataLogger`	data logger for class `CMAEvolutionStrategy`.
Class	`Logger`	log an arbitrary number of data (a data row) per "timestep".
Class	`LoggerDummy`	use to fake a `Logger` in non-verbose setting
Function	`custom_default_matplotlib`	reduce wasted margin area from 19.0% to 4.0% and make both grids default and
Function	`disp`	displays selected data from (files written by) the class `CMADataLogger`.
Function	`plot`	plot data from files written by a `CMADataLogger`, the call `cma.plot(name, argsdict)` is a shortcut for `cma.CMADataLogger(name).plot(argsdict)`
Function	`plot_zip`	create tar file `filename [+ ...] + '.tar.gz'` of folder `name`.
Function	`smartlogygrid`	turn on grid and also minor grid depending on y-limits
Variable	`__author__`	Undocumented
Variable	`last_figure_number`	Undocumented
Function	`_fix_lower_xlim_and_clipping`	minimize space wasted below x=0
Function	`_monotone_abscissa`	make x monotone if `not iabscissa`.

def custom_default_matplotlib(): ¶

reduce wasted margin area from 19.0% to 4.0% and make both grids default and

show legend when labels are available.

def disp(name=None, idx=None): ¶

displays selected data from (files written by) the class CMADataLogger.

The call cma.disp(name, idx) is a shortcut for cma.CMADataLogger(name).disp(idx).

Arguments

name: name of the logger, filename prefix, None evaluates to the default 'outcma/cma'
idx: indices corresponding to rows in the data file; by default the first five, then every 100-th, and the last 10 rows. Too large index values are removed.

The best ever observed iteration is also printed by default.

Examples

import cma
from numpy import r_
# assume some data are available from previous runs
cma.disp(None, r_[0, -1])  # first and last
cma.disp(None, r_[0:int(1e9):100, -1]) # every 100-th and last
cma.disp(idx=r_[0, -10:0]) # first and ten last
cma.disp(idx=r_[0:int(1e9):1000, -10:0])

See Also
`CMADataLogger.disp`

def plot(name=None, fig=None, abscissa=0, iteridx=None, plot_mean=False, foffset=1e-19, x_opt=None, fontsize=7, downsample_to=3000.0, xsemilog=None, xnormalize=None, fshift=0, addcols=None, **kwargs): ¶

plot data from files written by a CMADataLogger, the call cma.plot(name, **argsdict) is a shortcut for cma.CMADataLogger(name).plot(**argsdict)

Arguments

name: name of the logger, filename prefix, None evaluates to the default 'outcmaes/'
fig: filename or figure number, or both as a tuple (any order)
abscissa: 0==plot versus iteration count, 1==plot versus function evaluation number
iteridx: iteration indices to plot
x_opt: If isscalar(x_opt) it is interpreted as iteration number and the difference of x to the respective iteration is plotted. If it is a negative scalar the respective index rather than the iteration is used. Namely in particular, x_opt=0 subtracts the initial solution X0 and x_opt=-1 subtracts the final solution of the data. If len(x_opt) == dimension, the difference to x_opt is plotted, otherwise the first row of x_opt are the indices of the variables to be plotted and the second row, if present, is used to take the difference.
xsemilog: customized semilog plot for x-values
fshift: is added to plotted and shown f-values

Return None

Examples

cma.plot()  # the optimization might be still
            # running in a different shell
cma.s.figsave('fig325.png')
cma.s.figclose()

cdl = cma.CMADataLogger().downsampling().plot()
# in case the file sizes are large

Details

Data from codes in other languages (C, Java, Matlab, Scilab) have the same format and can be plotted just the same.

See Also
`CMADataLogger`, `CMADataLogger.plot`

def plot_zip(name=None, filename=None, unique=True): ¶

create tar file filename [+ ...] + '.tar.gz' of folder name.

The resulting tar file serves to recreate the current cma.plot elsewhere.

name defaults to the folder plotted with cma.plot by default. filename defaults to name while adding a unique time stamp if unique is true.

Return the path of the created file (may be absolute or relative).

Details

This is a convenience replacement for executing tar -czf filename.tar.gz name in a system shell where name defaults to the default plot data folder and filename is created to be unique.

This function calls CMADataLogger.zip to do the real work.

def smartlogygrid(**kwargs): ¶

turn on grid and also minor grid depending on y-limits

__author__: str = ¶

Undocumented

last_figure_number: int = ¶

Undocumented

def _fix_lower_xlim_and_clipping(): ¶

minimize space wasted below x=0

def _monotone_abscissa(x, iabscissa=0): ¶

make x monotone if not iabscissa.

Useful to plot restarted runs where the iteration is reset after each restart.

TODO: handle same iteration number better, but how? Currently we assume it was written "on purpose" twice.