[GRASS-dev] storing full script as metadata

Hello,

Just wanted to keep exact track of the processing leading to a new layer.

1 - would it be feasible to store the whole processing line with
parameters (i.e. r.something -a input=file_in output-file_out etc...)
as metadata in the output file of a module ?

2 - would it be interesting to make a scripts/ place inside a mapset
where scripts would get automatically saved with something like:
YYYYMMMDDHHHHH_MODULE_NAME.sh or the like?

just an idea.
Yann

--
Yann Chemin
International Rice Research Institute
Office: http://www.irri.org/gis
Perso: http://www.freewebs.com/ychemin

Yann Chemin wrote:

Just wanted to keep exact track of the processing leading to a new layer.

1 - would it be feasible to store the whole processing line with
parameters (i.e. r.something -a input=file_in output-file_out etc...)
as metadata in the output file of a module ?

G_command_history() already does this, i.e. it appends the result of
G_recreate_command() to the history.

2 - would it be interesting to make a scripts/ place inside a mapset
where scripts would get automatically saved with something like:
YYYYMMMDDHHHHH_MODULE_NAME.sh or the like?

I'm not sure what you're suggesting here.

GRASS (the libraries and modules) only knows which commands are run;
it doesn't know *how* they are run, e.g. if they are part a script, if
they were typed into the shell, executed by the GUI, etc.

--
Glynn Clements <glynn@gclements.plus.com>

2008/6/9 Glynn Clements <glynn@gclements.plus.com>:

Yann Chemin wrote:

Just wanted to keep exact track of the processing leading to a new layer.

1 - would it be feasible to store the whole processing line with
parameters (i.e. r.something -a input=file_in output-file_out etc...)
as metadata in the output file of a module ?

G_command_history() already does this, i.e. it appends the result of
G_recreate_command() to the history.

G_recreate_command()
OK I'll look into it.

2 - would it be interesting to make a scripts/ place inside a mapset
where scripts would get automatically saved with something like:
YYYYMMMDDHHHHH_MODULE_NAME.sh or the like?

I'm not sure what you're suggesting here.

GRASS (the libraries and modules) only knows which commands are run;
it doesn't know *how* they are run, e.g. if they are part a script, if
they were typed into the shell, executed by the GUI, etc.

well, since G_recreate_command passed it into the metadata,
I guess already that is good.
I was thinking more of a kind of "command log", but with time-based
separation of commands used in txt files.
Whether they are stored inside a mapset would just be convenient.

thanks
Yann

--
Glynn Clements <glynn@gclements.plus.com>

--
Yann Chemin
International Rice Research Institute
Office: http://www.irri.org/gis
Perso: http://www.freewebs.com/ychemin

Yann Chemin wrote:

well, since G_recreate_command passed it into the metadata,
I guess already that is good.
I was thinking more of a kind of "command log",
but with time-based
separation of commands used in txt files.
Whether they are stored inside a mapset would just be
convenient.

see also 'v.info -h <mapname>' for some developed <mapname> which is the result of a number of vector module processing steps.

and the raster-metadata rewrite task suggestion box on the grass wiki grass7 ideas page.

Hamish