[kst] [Bug 372265] New: "Change Data File" does not update file name references in text box

[Ben Lewis]

https://bugs.kde.org/show_bug.cgi?id=372265

Bug ID: 372265
Summary: "Change Data File" does not update file name
references in text box
Product: kst
Version: 2.0.8
Platform: MS Windows
OS: MS Windows
Status: UNCONFIRMED
Severity: normal
Priority: NOR
Component: general
Assignee: ***@kde.org
Reporter: ***@gmail.com
Target Milestone: ---

I would like to display some header information on a plot and have that
information update as the underlying data file changes. Currently this does not
work for me.

Using a text box I can display information (string data) contained in the
header of my data file. The link to this header information requires the file
name of the data file. When using the "Change Data File" function the file
name reference in the text box is not updated to the new data file name and the
link is broken.

[b***@kde.org]

https://bugs.kde.org/show_bug.cgi?id=372265

***@astro.utoronto.ca changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |***@astro.utoronto.
| |ca

--- Comment #1 from ***@astro.utoronto.ca ---
These automatic strings automatically created by the data source are the
problem. This is not (under the hood) a DataScalar, so it never shows up in
the list of things to change when you change datasources.

A work around for now is to create->string and then create a data string from
the data source from field FILE. This properly updates when you "change
datasource".
However, you have to hit reload to actually redraw it. I was working on fixing
that on the way into work this morning.

What to do with the automatic scalars and automatic vectors is more
problematic.
-How valuable are these automatic scalars/vectors. Would it be better to just
create them with create->scalar and create->string so they are consistent with
everything else? There should be create scalar and create string shortcut
icons in the label dialogs in any case.

[Ben Lewis]

https://bugs.kde.org/show_bug.cgi?id=372265

--- Comment #2 from Ben Lewis <***@gmail.com> ---
Hi Barth,

WRT the automatic scalars and vectors, I have not found them to be very useful
so, if it makes the code easier, then I would be okay with creating them
manually.

I have tried your work around and it kind of works but not quite. If I change
data files manually the "data strings" update as expected but if I run Kst from
the command line using the following script the "data strings" are not updated.

@echo off
for %%f in (*.csv) do (
echo "%%f"
"path\to\kst\kst2.exe" -F "%%f" --png "%%~nf.png" --pngWidth 3840 --pngHeight
2160 "kstfilename.kst"
)

Is there a way to force Kst to refresh the "data strings" prior to exporting an
image file?

Regards, Ben

[Nicolas Brisset]

https://bugs.kde.org/show_bug.cgi?id=372265

Nicolas Brisset <***@airbus.com> changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |***@airbus.com

--- Comment #3 from Nicolas Brisset <***@airbus.com> ---
Hi all,

I believe automatic scalars are useful and I would not like having to create
them manually. For instance, they can be used to display some stats in text
form along with a graphical plot. For me they should be preserved, and of
course recomputed and labels using them (if any) refreshed automatically when
the data is reloaded from the same file - or from another file.

The same should essentially happen with strings from datasources in my opinion.

Where it gets more tricky is when you have more than one datasource. In that
case we have to keep track of which datasource provides which object to be able
to update the right one(s) in case one of the datasources change.

[C Barth Netterfield]

OK.

as of the next commit, I (will) have fixed data strings and data scalars so
that they are saved properly, and so that -F and the change data source both
work for them as expected.

Now we have to decide what to do with the datasource automatic primitives:

I think there are three options:
1) they are permanently associated with the datasource. eg:
[tmp.dat (DS1):FRAMES (X1)]
is the number of frames for the file tmp.dat - now and forever. -F doesn't
change this. Change datasource doesn't change this. Saving and reloading
doesn't change this. The primitives can not be edited in any way.
With the exception of some minor bugs, this is the logic of how they have been
implemented.

2) When a datasource is first used, all of the primitives associated with the
datasource are automatically created as normal datascalars or datavectors.
They now show up as normal primitives in all of the dialogs, and can be
optionally edited and changed just like any other dataprimitive, including in
the change datasource dialog, or -F. However, since they are just
automatically created normal data primitives, "purge" would delete them
(unless we add some sort of flag to them).

We would have to decide which strings and scalars from a datasource should be
created automatically, and which ones should be left to manual creation.
This requires some (tractable) new code.

3) automatic strings and scalars are removed alltogether. If you want them,
you create them, just like any other datascalar or datastring.

(In any case we should add create string/scalar icons to text entry dialogs.)

thoughts?

cbn

Post by Ben Lewis
https://bugs.kde.org/show_bug.cgi?id=372265
What |Removed |Added
----------------------------------------------------------------------------
Hi all,
I believe automatic scalars are useful and I would not like having to create
them manually. For instance, they can be used to display some stats in text
form along with a graphical plot. For me they should be preserved, and of
course recomputed and labels using them (if any) refreshed automatically
when the data is reloaded from the same file - or from another file.
The same should essentially happen with strings from datasources in my opinion.
Where it gets more tricky is when you have more than one datasource. In that
case we have to keep track of which datasource provides which object to be
able to update the right one(s) in case one of the datasources change.

[b***@kde.org]

https://bugs.kde.org/show_bug.cgi?id=372265

--- Comment #4 from ***@astro.utoronto.ca ---
Git commit 1a1005423d9eb7f9f20bb2edc7f98ffc000c1552 by Barth Netterfield.
Committed on 30/11/2016 at 23:56.
Pushed by netterfield into branch 'master'.

-F also changes data strings, data scalars, & data matrixes.

M +5 -0 src/libkst/matrixfactory.cpp
M +4 -0 src/libkst/scalarfactory.cpp
M +4 -0 src/libkst/stringfactory.cpp

https://commits.kde.org/kst-plot/1a1005423d9eb7f9f20bb2edc7f98ffc000c1552

[b***@kde.org]

https://bugs.kde.org/show_bug.cgi?id=372265

--- Comment #5 from ***@astro.utoronto.ca ---
OK.

as of the next commit, I (will) have fixed data strings and data scalars so
that they are saved properly, and so that -F and the change data source both
work for them as expected.

Now we have to decide what to do with the datasource automatic primitives:

I think there are three options:
1) they are permanently associated with the datasource. eg:
[tmp.dat (DS1):FRAMES (X1)]
is the number of frames for the file tmp.dat - now and forever. -F doesn't
change this. Change datasource doesn't change this. Saving and reloading
doesn't change this. The primitives can not be edited in any way.
With the exception of some minor bugs, this is the logic of how they have been
implemented.

2) When a datasource is first used, all of the primitives associated with the
datasource are automatically created as normal datascalars or datavectors.
They now show up as normal primitives in all of the dialogs, and can be
optionally edited and changed just like any other dataprimitive, including in
the change datasource dialog, or -F. However, since they are just
automatically created normal data primitives, "purge" would delete them
(unless we add some sort of flag to them).

We would have to decide which strings and scalars from a datasource should be
created automatically, and which ones should be left to manual creation.
This requires some (tractable) new code.

3) automatic strings and scalars are removed alltogether. If you want them,
you create them, just like any other datascalar or datastring.

(In any case we should add create string/scalar icons to text entry dialogs.)

thoughts?

cbn

Post by Ben Lewis
https://bugs.kde.org/show_bug.cgi?id=372265
What |Removed |Added
----------------------------------------------------------------------------
Hi all,
I believe automatic scalars are useful and I would not like having to create
them manually. For instance, they can be used to display some stats in text
form along with a graphical plot. For me they should be preserved, and of
course recomputed and labels using them (if any) refreshed automatically
when the data is reloaded from the same file - or from another file.
The same should essentially happen with strings from datasources in my
opinion.
Where it gets more tricky is when you have more than one datasource. In that
case we have to keep track of which datasource provides which object to be
able to update the right one(s) in case one of the datasources change.