Goodness. During my G'MIC hibernation (Four years? Really?), I had forgotten about the experimental enthusiasms of the GimpChat crowd. Thank you all for all your examples. It is very useful to me to see how the script performs under the hands of others. Addressing some of your particular remarks:
Lyle: Not showing up iny G'MIC even after refresh. Use 2.9.7.My bad. I forget that when we script/plug-in authors commit a new script or plug-in to the
gmic-community code repository, it is initially only available to those who have migrated to the bleeding-edge pre-release version, currently 2.98-prerelease. For those using 2.9.7, such as yourself, the plug-in will not be a part of the update file — even when you deliberately press the circular arrow button at the bottom of the gmic-qt dialog box. For those using 2.9.7 and earlier, Testing -> Gmic Tutorials -> Rectangular Tiling won't appear. This is good policy, on the whole, because even when plug-in authors have the best intentions (I have only the best intentions!
) mistakes can happen on the early releases (me?!? mistakes?? NEVER!!!!
(then, why am I writing this post?
)) and production users of 2.9.7 should not have to suffer these.
I don't believe I can post the relevant script files here without triggering a GimpChat Security Alert. For the adventurous among you, — there are a few — these can be obtained from the release announcement at
Post 51: G’MIC Tutorial Fragments. There is a "Hide Detail" drop-down at the bottom of the post. Expand it and copy the revealed text into your
user.gmic file. For Windows folk, that
usually resolves to something like
C:\Users\<yourwindowsaccountname>\AppData\Roaming\user.gmic, where you replace
<yourwindowsaccountname> with your Windows user account name (without angle brackets, no quotes). Note: "usually," because your specific Windows installation may vary from the norm. On Windows 10, you may paste the Windows symbolic name
%AppData% in the search bar and File Explorer will open the relevant folder for you. See
Where to Find the AppData Folder in Windows 10 by Vijit Ail. Linux and MacOS users would append the revealed text to their
$HOME/.gmic file.
Make a backup of your gmic user file first, just in case mistakes happen. If you don't have a
user.gmic .gmic, then create one from scratch to contain the revealed text. You now have your very first personal G'MIC command. When 2.98 goes into production, remove this revealed text; the 2.9.8 production version will likely have an updated tiling script.
dinasset: Thank you for the python/Gimp plug-in wrapper (DIEGO_MOSAIC_GARRY_A.7z). Alas! I don't think it will help Lyle or anyone else with 2.9.7 or earlier - the code calls
pdb.plug_in_gmic_qt(GarryA_Image, GarryA_BG, 1,2 , "-v - -_rec_tileit 0.2975,10,0.25,1,32,1,0,0 "), the last parameter being a G'MIC command line that references
-_rec_tileit — and
that is part of the
gtutor_tileit package that is not available in updates for 2.9.7 and earlier. A noble effort. Thank you, sir, but I fear that the adventurous among you will still need to trek over to the watering hole I mentioned above. Sorry for the inconvenience.
level_0: I get a different result.????Probably. The
Disrupt Orientation widget references a random number generator and there is no telling at what random number that generator starts. It
can be asked to start at a specific point in the generator cycle, thereby giving rise to reproducible results. On the current release, I'm not asking, being predisposed to variety in life. I could put in a "seed" field in the UI, though I try to keep UI's as free of gadgets as possible — I get quirky if there are more than seven or eight things to adjust on a plug-in, and with this I'm at that limit already. Do you
really want one? Dinasset will fine me fifty euros if the UI gets too cluttered (
). Or I could internally fix the seed and
everyone has reproducible results; perhaps no one would notice — or care much.
I am charmed by the Mother-and-Child rendition, by the way. Thank you.
dinasset: "often the holes are filled with too dark colour values with respect to the area colour."There are color shifts when a photographic source also has an alpha channel. In any case, I agree that holes/hole-filling could use some work, but I am not sure when. I don't have full grasp of the nature of that color shift yet. And - as you know - there is precious little scripting documentation. That's my row I've got to hoe.