merge in Jack's new fitter class into spinmob. From "02 Easy Gui Generator v2014-09-07.zip"

ability to set a settings kwarg to not display the spinmob welcome upon import. Not sure if this can happen before spinmob starts easily.

Perhaps we should have a "silent" option on loading files to suppress warnings like:

Could not fine a line of pure data! Perhaps check the delimiter?
The default delimiter is whitespace. For csv files, set delimiter=','

This is helpful when hitting some silly problem, but it would be nice not to have it when loading files that just container header information, for example.

It would also be nice to suppress the
"loaded full_file_path"
text output.

Clicking it seems to only toggle, not bring up the directory like it should.

It would be nice to include an example where a couple of data sets are simultaneously fit to a set of free parameters.

Can't do this yet.

Look at changing "studentized residuals" to (y_i - f)/sigma.

The old spinmob (hosted on google code) is the current master. It does not include the class based fitter object. All new code is in the bleeding branch. Feel free to merge this with master, or from master we could create a new "old spinmob" branch. I think I will do that.

If there is no figure plotted, coarsen_all_traces probably shouldn't open a new figure.

Also, the window has "block_events" while objects have "block_signals", which seems inconsistent.

but other functions are working.

Hitting an error when loading a file that has a single element for each column of data. I'm printing what looks like relevant information in this branch: https://github.com/Spinmob/spinmob/tree/load_single_line_of_data

In _functions.py elimante wx dependence, put in pyqt instead.

For some reason subsequent f.set_data() calls results in a memory leak? Unconfirmed.

See line 144.

From the comment it looks like self.xdata should be hidden.

See https://github.com/Spinmob/spinmob/blob/master/_data.py#L820.

Same for ydata, and eydata (I think).

self._xdata, self._ydata, self._eydata.

When plotting it would be great if when making a plot the new plot is always placed ontop of any other open windows as a default.

As far as I can tell _functions.py is unused. I'm fine with deleting this, please go ahead.

It would be nice to have an option to suppress the warning displayed from the fitter when performing a fit with the default value eydata=None.

There is a LOT of > 10 year-old code in this file. We should comb over it while writing tests and eliminate silly / useless functions.

Where should integrate_shown_data be implemented/re-implemented?

Current functionality:
sm.plot.xy.files(xscript='c(0)', yscript='c(1)/41.')

Test this with new databox class.
sm.plot.xy.files(xscript='c[0]', yscript='c[1]/41.')

something about _egg

Try pushing auto-save

Ability to do

sm.settings['delimiter'] = ','

so that the default limiter is CSV, for example.

See line 1064 nm.

Make sure that the path is right justified when plotted.

In the "plotting" page of the wiki there are no actually plots (yet).

There is a LOT of > 10 year-old code in this file. We should comb over it while writing tests and eliminate silly/slow functions that don't fully exploit numpy.

import spinmob as sm; sm.plot.magphase.files(0,'d[1]+1j*d[2]')

Works for 1 file but not for 2 or more.

It might be nice to have a function

sm.data.fitter.set_parameters()

where the user can set the parameters for a fitter instance.

If the cfg file has a value outside the bounds of a float, the control will show the bounded value but the actual get_value() will be the one from the file!!

Everything that is seen in a plot should be printed.