Undo Feature: Use ROI-Manager instead of Draw

[demisjohn]

Perhaps have Fiji place the text & lines etc. on an Overlay layer, instead of drawing directly to the image. Not sure if this would enable the user to effectively undo by deleting the overlaid layer.

Currently the lack of "undo" is mildly irritating, cauing me to Revert the image if I don't like the annotation.

[demisjohn]

Should see how the Scale Bar's [Overlay] checkbox achieves this, and copy it.

[demisjohn]

Some example code on how to place text on an overlay layer here:
https://imagej.nih.gov/ij/macros/python/Create_Overlay.py

[demisjohn]

I tried using Overlay manually in FIJI, as this is an option in the Text tool. Works exactly as expected - one can move the text around. The "Flatten" option or "save" to a JOEG etc. thus stamps the text onto the image permanently.

Should implement our "Draw Measurement" as an Overlay, would be very useful.

[demisjohn]

Manual method, from here: https://imagej.nih.gov/ij/docs/guide/146-11.html
(After Analyze > Set Scale...)

1. Open Analyze > Tools > ROI manager
   2a) Draw your ROI (eg. ellipse, text, line tools etc.)
   2b) In ROImgr, [add] & set properties - linewidth, color, name; ROImgr > More... > Labels > Font Size larger + Names as labels; then [Deselect]
   2c) Draw more ROIs, modify then [Deselect]
   3a) Image >Overlay > Overlay Options: Set Options (fontsize, color, Show Labels [√])
   3b) Image > Overlay > From ROI Manager - will overlay all lines and their names
2. Adjust ROI positions, by selecting correct ROI tool, and adjusting positions.
3. Image > overlay > Flatten will draw the overlay onto the image exactly as it'll appear in the export (applying text sizes to image dpi etc. - not same as vector versions)
4. Save As > JPEG... or any format that flattens the overlay into a raster version.
   (Optionally can Edit > Draw - although even without this, the Overlay appears to get drawn already when you Save As...)

[demisjohn]

For this plugin to not draw directly to the image, it'll have to instead add the ROI's (text annotations, lines) to the ROI manager, and pop open the ROI manager. (NOT overlay GUI - that doesn't allow you to move or adjust anything.)

At this point, the user can move ROIs (by choosing the correct ROI tool, then moving etc.).

But now to complete the process, we have to:

• Image >Overlay > Overlay Options
  -- using the settings file for font sizes etc.
• Image > Overlay > From ROI Manager
• Image > Overlay > Flatten

If we enable the "overlay" option, then we'll also need to add a new plugin called Draw Annotations which does these steps for the user (otherwise they may as well just learn all the details about FIJI, which is what we're trying to avoid).

---

Add option in settings.py file for Use Overlay or something similar, which enables this option.

[demisjohn]

On second thought - actually our scripts should finish by adding the ROIs to the Overlay, but not flattening or saving. Users can undo by using Image > overlay > Remove Overlay.

So if settings.py > Use_Overlay = True, then script's procedure would be:

• Acquire Line ROI from user's drawn line
• Generate Text ROI from above - fill color enabled

Then:

overlay.add(roi)  # for each roi generated
imageobj.setOverlay(overlay)
imageobj.show()

User can either Save, or "undo" via Image > Overlay > Remove Overlay.

The only drawback is that the text sizes shown of the ROI will not reflect the final 'drawn" text sizes, until the user does Image > Overlay > Flatten, which is then not undoable.

[demisjohn]

Maybe add a "show annotation preview" script, that pops open a new window showing the "drawn"/Flattened version of the image.