4:37 p.m.
I was trying to estimate a state space model, using MLE and I noticed that even though
these can be saved to the session some models are lost when reopening the session. In the
example below I'm trying to save the MLE estimation to the session with the name
"model1", which is successful but then it's lost at reopening.
Any thoughts on why this happens and on an alternative to save the session?
<hansl>
clear
nulldata 200
set seed 101010
setobs 1 1 --special-time-series
/* set the true variance parameters */
true_s1 = 0.5
true_s2 = 0.25
/* and simulate some data */
v = normal() * sqrt(true_s1)
w = normal() * sqrt(true_s2)
mu = 2 + cum(v)
y = mu + w
/* starting values for variance estimates */
s1 = 1
s2 = 1
/* state-space model set-up */
bundle kb = ksetup(y, 1, 1, s1)
kb.obsvar = s2
kb.diffuse = 1
/* ML estimation of variances */
model1 <- mle ll = ERR ? NA : kb.llt
ERR = kfilter(&kb)
params kb.statevar kb.obsvar
end mle
/* compute the smoothed state */
ksmooth(&kb)
series muhat = kb.state
</hansl>
6:35 a.m.
Am 29.09.2020 um 18:37 schrieb Filipe Costa:
I was trying to estimate a state space model, using MLE and I
noticed
that even though these can be saved to the session *some models are lost
when reopening the session*. In the example below I'm trying to save the
MLE estimation to the session with the name "model1", which is
successful but then it's lost at reopening.
Any thoughts on why this happens and on an alternative to save the session?
<hansl>
...
(First, very useful that you provided the example code!)
Not sure what you mean. In the session view I can re-open the "model1"
by double-clicking alright. Either you must be talking about something
else, or it's not working in your setup. Platform, version numbers?
thanks
sven
9:16 a.m.
Hi again,
The model appears in the session icons. Now save the session and close Gretl. Then reopen
the session and you'll see the following message:
1 session object(s) could not be rebuilt.
And then the model isn't there on the session icons. My gretl version is 2020d running
on Arch Linux.
> On 30/09/2020 07:35 Sven Schreiber <svetosch(a)gmx.net> wrote:
>
>
> Am 29.09.2020 um 18:37 schrieb Filipe Costa:
> > I was trying to estimate a state space model, using MLE and I noticed
> > that even though these can be saved to the session *some models are lost
> > when reopening the session*. In the example below I'm trying to save the
> > MLE estimation to the session with the name "model1", which is
> > successful but then it's lost at reopening.
> >
> > Any thoughts on why this happens and on an alternative to save the session?
> >
> > <hansl>
> ...
> (First, very useful that you provided the example code!)
>
> Not sure what you mean. In the session view I can re-open the "model1"
> by double-clicking alright. Either you must be talking about something
> else, or it's not working in your setup. Platform, version numbers?
>
> thanks
> sven
> _______________________________________________
> Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
> To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
> Website: https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
10:07 a.m.
Am 30.09.20 um 11:16 schrieb Filipe Costa:
The model appears in the session icons. Now save the session and
close Gretl. Then reopen the session and you'll see the following message:
1 session object(s) could not be rebuilt.
And then the model isn't there on the session icons. My gretl version is 2020d
running on Arch Linux.
Right, I see what you mean. As further output on stderr I get:
parser error : EntityRef: expecting ';'
ERR = kfilter(&kb)
^
(where the caret ^ is under the closing parenthesis)
From the outside it looks like a relatively easy-to-fix bug...
In the meantime you could perhaps save the $model bundle to a file with
the bwrite() function and later read it back in with bread(). Don't know
if that serves your ultimate purpose.
thanks for the report
sven
12:31 p.m.
On Wed, 30 Sep 2020, Sven Schreiber wrote:
Am 30.09.20 um 11:16 schrieb Filipe Costa:
> The model appears in the session icons. Now save the session and close
> Gretl. Then reopen the session and you'll see the following message:
>
> 1 session object(s) could not be rebuilt.
>
> And then the model isn't there on the session icons. My gretl version is
> 2020d running on Arch Linux.
Right, I see what you mean. As further output on stderr I get:
parser error : EntityRef: expecting ';'
ERR = kfilter(&kb)
(where the caret ^ is under the closing parenthesis)
From the outside it looks like a relatively easy-to-fix bug...
Yes, it's now fixed in git and snapshots. Thanks for the report,
Filipe.
(The fix means that newly-created session files containing this sort
of model will work OK. Unfortunately it does not mean that
previously created session models containing unescaped "&" can be
read successfully, though they could be hand-edited to make them
readable.)
Allin
8:40 a.m.
Glad to help uncover any bug Allin!
I think I'm not able to install the update from git on Arch (eventually I don't
have the skill). Any workaround on the code that can make it work if the session is saved
and reopened later? You wrote about the unescaped "&" as being the problem.
What can I do there exactly?
> On 30/09/2020 13:31 Allin Cottrell <cottrell(a)wfu.edu> wrote:
>
>
> On Wed, 30 Sep 2020, Sven Schreiber wrote:
>
> > Am 30.09.20 um 11:16 schrieb Filipe Costa:
> >> The model appears in the session icons. Now save the session and close
> >> Gretl. Then reopen the session and you'll see the following message:
> >>
> >> 1 session object(s) could not be rebuilt.
> >>
> >> And then the model isn't there on the session icons. My gretl version is
> >> 2020d running on Arch Linux.
> >
> > Right, I see what you mean. As further output on stderr I get:
> >
> > parser error : EntityRef: expecting ';'
> > ERR = kfilter(&kb)
> >
> > (where the caret ^ is under the closing parenthesis)
> >
> > From the outside it looks like a relatively easy-to-fix bug...
>
> Yes, it's now fixed in git and snapshots. Thanks for the report,
> Filipe.
>
> (The fix means that newly-created session files containing this sort
> of model will work OK. Unfortunately it does not mean that
> previously created session models containing unescaped "&" can be
> read successfully, though they could be hand-edited to make them
> readable.)
>
> Allin
> _______________________________________________
> Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
> To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
> Website: https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
9:01 a.m.
Am 01.10.20 um 10:40 schrieb Filipe Costa:
Glad to help uncover any bug Allin!
I think I'm not able to install the update from git on Arch (eventually I don't
have the skill). Any workaround on the code that can make it work if the session is saved
and reopened later? You wrote about the unescaped "&" as being the problem.
What can I do there exactly?
The .gretl session files are "just" a gzipped collection of XML files
AFAIK. So unzip the myfile.gretl archive file to get a folder; inside
that folder there should be a "model.1" file, and inside that file you
have the problematic line with the unescape ampersand:
ERR = kfilter(&kb)
I _think_ it should be enough the replace the "&" with the corresponding
xml code: "&" (without the quotes). Then save the file and gzip the
whole previously extracted folder again into a .gretl archive file; try
to reload with gretl.
hth,
sven
11:55 a.m.
On Thu, 1 Oct 2020, Sven Schreiber wrote:
Am 01.10.20 um 10:40 schrieb Filipe Costa:
> Glad to help uncover any bug Allin!
>
> I think I'm not able to install the update from git on Arch (eventually I
> don't have the skill). Any workaround on the code that can make it work if
> the session is saved and reopened later? You wrote about the unescaped
"&"
> as being the problem. What can I do there exactly?
The .gretl session files are "just" a gzipped collection of XML files
AFAIK. So unzip the myfile.gretl archive file to get a folder; inside
that folder there should be a "model.1" file, and inside that file you
have the problematic line with the unescape ampersand:
ERR = kfilter(&kb)
I _think_ it should be enough the replace the "&" with the corresponding
xml code: "&" (without the quotes). Then save the file and gzip the
whole previously extracted folder again into a .gretl archive file; try
to reload with gretl.
Since you're on Linux, you may find this little bash script helpful:
<shell-script>
#!/bin/sh
if [[ -z $1 ]]
then
echo "Usage; fix-sessionfile <session file name>"
exit 1
fi
FN=$(basename $1 .gretl)
DIR=.$FN
rm -rf $DIR
unzip $1
LIST=$(grep -l \& $DIR/*)
for f in $LIST
do
sed $f -i -e 's/&/&/g'
done
zip tmp.zip -r $DIR
mv tmp.zip "$FN-fixed.gretl"
</shell-script>
-------------------------------------------------------
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
-------------------------------------------------------
12:53 p.m.
Am 01.10.2020 um 13:55 schrieb Riccardo (Jack) Lucchetti:
On Thu, 1 Oct 2020, Sven Schreiber wrote:
>
> I _think_ it should be enough the replace the "&" with the
corresponding
> xml code: "&" (without the quotes). Then save the file and gzip
the
> whole previously extracted folder again into a .gretl archive file; try
> to reload with gretl.
Since you're on Linux, you may find this little bash script helpful:
...
sed $f -i -e 's/&/&/g'
I'm barely literate in bash or sed, I hope this doesn't replace the &
character in an already-escaped & sequence!
thanks
sven
12:58 p.m.
On Thu, 1 Oct 2020, Sven Schreiber wrote:
Am 01.10.2020 um 13:55 schrieb Riccardo (Jack) Lucchetti:
> On Thu, 1 Oct 2020, Sven Schreiber wrote:
>
>>
>> I _think_ it should be enough the replace the "&" with the
corresponding
>> xml code: "&" (without the quotes). Then save the file and gzip
the
>> whole previously extracted folder again into a .gretl archive file; try
>> to reload with gretl.
>
> Since you're on Linux, you may find this little bash script helpful:
> ...
> sed $f -i -e 's/&/&/g'
I'm barely literate in bash or sed, I hope this doesn't replace the &
character in an already-escaped & sequence!
Hm, I think you're right, I hadn't thought of that.
OTOH, what cases do we possibly have of escaped sequences in the xml files
contained in a session file?
-------------------------------------------------------
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
-------------------------------------------------------
1 p.m.
Am 01.10.2020 um 14:58 schrieb Riccardo (Jack) Lucchetti:
On Thu, 1 Oct 2020, Sven Schreiber wrote:
> Am 01.10.2020 um 13:55 schrieb Riccardo (Jack) Lucchetti:
>>
>> Since you're on Linux, you may find this little bash script helpful:
>> ...
>> sed $f -i -e 's/&/&/g'
>
> I'm barely literate in bash or sed, I hope this doesn't replace the &
> character in an already-escaped & sequence!
Hm, I think you're right, I hadn't thought of that.
OTOH, what cases do we possibly have of escaped sequences in the xml
files contained in a session file?
" ?
-s
1:13 p.m.
On Thu, 1 Oct 2020, Sven Schreiber wrote:
Am 01.10.2020 um 14:58 schrieb Riccardo (Jack) Lucchetti:
> On Thu, 1 Oct 2020, Sven Schreiber wrote:
>
>> Am 01.10.2020 um 13:55 schrieb Riccardo (Jack) Lucchetti:
>>>
>>> Since you're on Linux, you may find this little bash script helpful:
>>> ...
>>> sed $f -i -e 's/&/&/g'
>>
>> I'm barely literate in bash or sed, I hope this doesn't replace the
&
>> character in an already-escaped & sequence!
>
> Hm, I think you're right, I hadn't thought of that.
>
> OTOH, what cases do we possibly have of escaped sequences in the xml
> files contained in a session file?
" ?
You're right. However, if we had a list of those, it'd be possible to
instruct sed to leave them alone (provided there's not too many).
-------------------------------------------------------
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
-------------------------------------------------------
4:41 p.m.
On Thu, 1 Oct 2020, Riccardo (Jack) Lucchetti wrote:
On Thu, 1 Oct 2020, Sven Schreiber wrote:
> Am 01.10.2020 um 14:58 schrieb Riccardo (Jack) Lucchetti:
>> On Thu, 1 Oct 2020, Sven Schreiber wrote:
>>
>>> Am 01.10.2020 um 13:55 schrieb Riccardo (Jack) Lucchetti:
>>>>
>>>> Since you're on Linux, you may find this little bash script helpful:
>>>> ...
>>>> sed $f -i -e 's/&/&/g'
>>>
>>> I'm barely literate in bash or sed, I hope this doesn't replace the
&
>>> character in an already-escaped & sequence!
>>
>> Hm, I think you're right, I hadn't thought of that.
>>
>> OTOH, what cases do we possibly have of escaped sequences in the xml
>> files contained in a session file?
>
> " ?
You're right. However, if we had a list of those, it'd be possible to
instruct sed to leave them alone (provided there's not too many).
See gretl_xml_encode() in lib/src/strutils.c
Allin
1:27 p.m.
I tried to replace "&" by "&" in the xml .gretl files and
it works well. But I tried to go one step behind and model this way in the original file:
ERR = kfilter(&kb)
But unfortunately it doesn't work!!!! Is that expected for you?
> On 01/10/2020 10:01 Sven Schreiber <svetosch(a)gmx.net> wrote:
>
>
> Am 01.10.20 um 10:40 schrieb Filipe Costa:
> > Glad to help uncover any bug Allin!
> >
> > I think I'm not able to install the update from git on Arch (eventually I
don't have the skill). Any workaround on the code that can make it work if the session
is saved and reopened later? You wrote about the unescaped "&" as being the
problem. What can I do there exactly?
>
> The .gretl session files are "just" a gzipped collection of XML files
> AFAIK. So unzip the myfile.gretl archive file to get a folder; inside
> that folder there should be a "model.1" file, and inside that file you
> have the problematic line with the unescape ampersand:
>
> ERR = kfilter(&kb)
>
> I _think_ it should be enough the replace the "&" with the
corresponding
> xml code: "&" (without the quotes). Then save the file and gzip
the
> whole previously extracted folder again into a .gretl archive file; try
> to reload with gretl.
>
> hth,
>
> sven
> _______________________________________________
> Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
> To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
> Website: https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
1:29 p.m.
On Thu, 1 Oct 2020, Filipe Costa wrote:
I tried to replace "&" by "&" in the
xml .gretl files and it works well. But I tried to go one step behind and model this way
in the original file:
ERR = kfilter(&kb)
But unfortunately it doesn't work!!!! Is that expected for you?
Yes, that cannot work.
-------------------------------------------------------
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
-------------------------------------------------------
1547
days inactive
1549
days old
14 comments
4 participants
participants (4)
-
Allin Cottrell -
Filipe Costa -
Riccardo (Jack) Lucchetti -
Sven Schreiber