Text in infobox:
The plotting implemtation in ANAPHE is only rudimentary, so I never really
tried that. But in the code I see that the text is just appended to the legend
items. It seems that there only seems to be one 'Infobox' which can contain
text (user-defined or statistics/fitting params (i.e. annotations) of a histogram)
or 'legends' (i.e. text with some graphics thing). This is also the impression
I get from the AIDA definitions (which, I agree, are not very clear..)
I looked a little bit around to find ways to suppress this problem: it seems
that in the parser class (http://www.python.org/doc/current/lib/xmlparser-objects.html)
there is a function StartDoctypeDeclHandler which may need to be overwritten
in order to stop it from looking up the dtd file on the web?
I don't know but maybe this hint can help you..
I think it would useful to have the possibility to do simple error calculation
only (if it speeds things up). Especially when you are in an interactive
session you don't want to wait more than a few seconds for a (simple) fit
result; at the same time you don't need the most sophisticated errors.
'New' plotting bug:
Plotting (1D) histograms with zero entries results in some math exception
because it tries to have a Y axis between 0 and 0. I don't know what the
optimal behaviour should be, but checking upperY and lowerY and setting them
to 1 and 0, respectively can be a solution (other plotting tools write a
big 'no entries' in the middle of the plot which I personally don't like
I didn't check the behaviour of 2D and 3D.
From: Koji KISHIMOTO <korry@us...> - 2004-09-26 08:50:13
I'm sorry for my late reply. I've been absent for a week.
(i) Text in infobox
I reached a determination. Now I think it will be better for PAIDA to
be able to display three boxes: statistics, legends and texts. The
texts box will display user-defined texts like fitted parameter values.
The Python documentation you suggested seems to be very helpful to this
issue. I'll try to overwrite the handler.
(iii) Fitting speed
OK, in the next release, PAIDA will accept a option argument to limit
the asymmetric error calculation. The fitting time will be reduced by
about 50% at least.
(iv) Plots with zero entries
That's right. I had also found this bug and modified local copies to,
as you mentioned, check the lower and upper values.