curious bug
by Riccardo (Jack) Lucchetti
Hi all,
being the king of the typo has its advantages: for example, you find weird
bugs:
<hansl>
nulldata 60
setobs 4 1:1
y = normal()
ols y 0 y(0 -to -1) # note the typo
ols y 0 y(1)
</hansl>
The two models are identical.
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------
2 years, 10 months
quiet option issues with 'append' and 'store'
by Sven Schreiber
Hi,
I'm noticing the following. First:
<hansl>
open australia
append denmark --quiet # still get message!
</hansl>
... despite the doc saying that the quiet option should lead to no printout.
Secondly, it would also be useful if 'store' had such an option in the
first place. The use case is the same as for append, namely that in some
scripts/packages the printout shouldn't be cluttered with more or less
meaningless path echo information.
This is with a snapshot from two weeks ago.
thanks
sven
2 years, 11 months
$sysinfo.os vs. $windows (and Mac?)
by Sven Schreiber
Hi all,
the doc says that $sysinfo.os gives "linux", "osx", "windows" or other
as a piece of introspection. An older accessor is $windows, which
returns 0 or 1.
So the $windows accessor appears redundant and doesn't cover the Mac -
should it perhaps be discouraged for new hansl scripts? (I guess
$sysinfo.os has been around long enough such that there are no real
compatibility issues.) At least I guess the doc for $windows should also
point to $sysinfo.os.
As a further minor point, AFAIK MacOS is not named "OS X" anymore. But
obviously, changing the return string to "macos" or something like that
would raise backwards compatibility problems. Again, maybe the easiest
thing is simply to add a note to the doc?
thanks
sven
2 years, 11 months
unusable combined plot of VAR residuals n=6
by Sven Schreiber
Hi, I was somewhat surprised that something quite basic didn't work out
well: Estimating a 6-dim VAR with roughly T=700 I'm selecting (in the
VAR model/system window) the combined residual plot, and I'm getting a
relatively small window --compared to the screen size-- where the
heights of each of the six plots are very small, and almost nothing can
be seen, really.
Another partly related small remark: In the VAR window, the plot menu
has entries for the IRF labeled with the endogenous variables, which is
good, but the residuals sub-menu just talks of "Equation 1" through
"Equation 6". It would be good to also have the variable names inserted
automatically. (Oh, and it doesn't seem to be marked for translation.)
thanks
sven
2 years, 11 months
weird function arg bug
by Sven Schreiber
Hi,
more or less minimal example:
<hansl>
function list heylist (const series x)
print "aha"
list out = x # L
return out
end function
open denmark
heylist(diff(LRM)) # works!
list oi = heylist(diff(LRM)) # fails
</hansl>
thanks
sven
2 years, 11 months
Current state with distance()
by Artur T.
Hi,
I am currently testing the new distance() function under development.
Thank you again for this new feature.
I am focusing on the pairwise case where both matrices X and Y are
given. Here is a toy example:
<hansl>
clear
set verbose off
matrix X = {1, 2; 3, 4}
matrix Y = {1, 2; 0, 3}
expected_distances = {0, 2; 4, 4}
matrix actual = distance(X, "manhattan", Y)
print actual
print expected_distances
actual (4 x 1)
0
2
4
4
expected_distances (2 x 2)
0 2
4 4
</hansl>
The matrix expected_distances is replicating sklearn's behavior. As you
can see, sklearn returns a matrix instead of a column vector. I am not
sure what the standard is and/ or other software packages return. I just
want to mention this. Being in line with some widely-applied "standard"
is useful, I would say. I also think that the matrix version is easier
to read - but that might be a matter of taste ;-)
Also, the help text does not say anything on the return dimension in
case both matrices X and Y are given.
Artur
2 years, 11 months
Edge cases for "end" operator
by Artur T.
Hi all,
I wrote some unit-tests for the recent "end" operator. See here:
https://sourceforge.net/p/gretl/workspace/ci/09881e0435c18730b1a8b54210dc...
I found some edge cases for which the current behavior may be discussed:
<hansl>
clear
set verbose off
WHICH = 1
if WHICH == 1
bundles B = array(0)
eval B[end]
elif WHICH == 2
matrix m = {}
eval m[end]
elif WHICH == 3
strings Sempty = array()
eval Sempty[end]
elif WHICH == 4
open denmark.gdt --quiet --preserve
list L0 = null
eval L0[end]
elif WHICH == 5
open denmark.gdt --quiet --preserve
eval LRM[end] # Index value 2147483647 is out of bounds
elif WHICH == 6
open denmark.gdt --quiet --preserve
eval LRM[(end-1)]
/*
DUM_END: not interpreted for upsym SERIES
> LRM[(end-
The symbol 'end' is undefined
Data error
*/
endif
</hansl>
For the first four cases, I get the message (by the way, no error or
warning is flagged): "Index value 0 is out of bounds". Is this expected,
or should it be handled differently?
For case 5, I would have expected the last entry (conditional on the
currently active sample) of series LRM.
Hope that's useful.
Best,
Artur
2 years, 11 months
persistent.tex ?
by Sven Schreiber
Hi,
it took me a while to find the tex source for the Named lists and
strings chapter of the guide -- why on earth is the file called
persistent.tex ?? I propose to rename it unless I'm missing something.
cheers
sven
2 years, 11 months
adding obs not noticed by model forecast GUI
by Sven Schreiber
Hi,
the following happened with 2021d: I open denmark.gdt and estimate a toy
model in the GUI. Then I go to Analysis/Forecasts in the model window.
Gretl complains about unavailable out-of-sample obs. I dismiss this
message window and I switch to the main window and add 5 obs. Switching
back to the still open model window I attempt forecasting again. -> But
I get the same complaint, which now is no longer true.
thanks
sven
2 years, 12 months
Ubuntu LTS
by Sven Schreiber
Hi,
after the Ubuntu question on the users list it occurred to me that a new
long-term support version of Ubuntu is coming up in April 2022.
The feature freeze seems to be in February. Not sure if there is another
relevant deadline for packages like gretl before that, but I think we
should keep that in mind, for example in terms of potential new hansl
language features that might be needed for function packages later.
cheers
sven
3 years