Loading a .model file in dinver

[alekos]

Dear Marc,
On dinver, after you load the ground
profiles from the inversion, you can not load the a ".model" file!!
although this option exists on the menu.

Best Wishes
Alekos

[admin]

This option is still to be implemented... sorry for that.

Meanwhile, you can create the .layer containing your theoretical model for comparison with the inverted ones:

Code: Select all

$ cat my_theoretical.model | gpprofile -vs | figue -c 

Save the layer and add it inside dinver. If you want another color or change the weight of the line, click on properties for the graphcontent and move to tab "Lines". There you can adjust the legend. You can also save the legend and use it directly in the command line (if you must do this job hundreds of time):

Code: Select all

$ cat my_theoretical.model | gpprofile -vs | figue -c -l mylegend.legend

You can do the same with Vp:

Code: Select all

$ cat my_theoretical.model | gpprofile -vp | figue -c 

And save it in another .layer file.

Currently, you can skip the graphical user interface of figue when you want to produce jpeg or .page file directly (options -e or -s). I should add an option to save the layer directly (avoiding right clicks and so on).

[alekos]

Hi Marc,
the same can be done for the dispersion curve. In a similar simple way. However, what one can do with the Autocorrelation theoretical curves from a .model file? You can have them in figue but if one needs to save all these layers (one for each ring) and add them in the Autocorrelation results of dinver (or in a page file with the autocorrelation results) it will take him a long time. Imagine that someone might have a lot of dinver files for the same site and wants to add the .model theoretical autocorrelation curves.

[admin]

Hi Alekos,

For saving layers automatically without entering the gui, you can use option -layer in figue (see -help for details). Then you will have a collection of .layer files.

It would be handy to be able to add layers to plots in a straight way: select n plots, select n .layer files, but how to attribute each .layer to a particular plot? Which order? Do you have any suggestion? A first implementation can follow plot order and alphabetical order for .layer file names.

"Add model" menu option should be added for Profiles, DC, SPAC and Ell. For the last three the forward computation is automatically started for the given model.

Ok I can handle these requests for future releases.

[alekos]

Hi Marc,
I have a suggestion. If you sort the rings in the ring file you use when calculate the Autocorrelation for the reference model on the same way that are stored in the target file you use for dinver then things might be easy. If you put the autocorrelation lines of the reference model on figue than you will see that the name of each object is 'object_n' (where the number of the ring in the ring file). However if you load the page file with the dinver results (or open the autocorrelation results in dinver) in the properties of each figure (Tab Object) the name is 'spac_n'.. So what we are talking about can be easily done if one merges 'spac_n' with 'object_n'.

I hope I have helped.

[admin]

I can work on page files having the same structure and merge their plots but I think it is not explicit enough. Keeping your idea that rings can be organized by the user, I would prefer to export all .layer files and provide a quick import of several .layer in figue (or in any sheet). This is relatively fast to implement. It will be available in next release, today? Hum... not sure but it would be nice.

[alekos]

Hi Marc,
I would appreciate if you add this feature in the release you are preparing now. It would be nice if we could merge the 10ths of different dinver models we have as page files with the reference model Autocorrelation Curves which we have as page files as well.

Thank you

[admin]

This is implemented now... How does it work?

I add two items in menu "Format": add/save layers

They both interact with selected graphs only. .layer files are saved in the order of plots. "Add layers" sort the file names you select before adding.

You can view the graph order by clicking on menu "Edit/Order". Blue bullets should appear. On texts, the bullets may be not clearly visible. In this mode, clicking on one graph change the order. You can easily reorganize the order by clicking on graphs one after the other. "Send to back" and "Bring to front" from object contextual menus act also on the same order of plots. To avoid uncontrolled alterations of this order, click again on menu item "Edit/Order" once you've checked that the order is ok. The blue bullets should disappear.

[sbonnefo]

Hi Marc,

I have the same problem as Alekos, I can not load a .model file.

I try your suggestion and save a layer file. However, I can not visualize the Vs model in figue using the in line command:
$ cat my_theoretical.model | gpprofile -vs | figue -c

I've got this error message: "Missing value for Vs at layer 0"
Maybe there is something wrong with my model file? (I attach it below)
I looked at the gpprofile help to write my model file.

Sylvette

[admin]

You .model file looks ok. I tested here with success under Linux (I did the same command line).

It shouldn't be better with an upgrade. I did not change any of these lines for loading model files.
You get the same message for all softwares: gpdc, gpell,... do you?

I suspect the stupid Window termination (\r\n or Carriage Return/Line Feed) of lines to be responsible, but I can't figure out why. Would it be possible to remove the \r (ASCII char 0D in hex or 13 in dec)?

If you don't know how to do it, here is a small awk script:

Code: Select all

cat yourfile | awk '{
  gsub("\r","",$0)
  print $0
  }' > yourfilewithoutCR

Is it better? If so, the next time I run under Windows, I'll fix it.

[sbonnefo]

Please forget my model file ... I simply forgot to indicate the model file name ... OUPS!
I apologize for this very stupid error ...

Just to know, have you implemented the 'add model' option in Dinver in the latest release?

[admin]

Not yet implemented.