(Modified 2009 Oct 31)

Recent modifications:

• 2009-10-31 -- can read pre-existing selection doc using CTRL-r
• 2009-05-11 -- looks for filenumber in directory if not in filename
• 2009-04-23 -- linked CTRL-v to invert (CTRL-i still works)
• 2008-09-29 -- sorts keys in output doc file
• 2008-09-29 -- image size recalculated for each new page
• 2008-09-29 -- file-number from command-line doc-file input carried to other files
• 2007-12-20 -- attached GNU GPL
• 2007-12-17 -- added 1/4th size display option
• 2007-12-14 -- restored brightness slider
• 2007-11-27 -- shows pop-up window if doc file not found
• 2007-11-27 -- shortcuts CTRL-N and CTRL-P increment and decrement filenumber in Open Selections
• 2007-11-10 -- can use arbitrary doc-file column for particle-number
• 2007-10-11 -- can label with arbitrary column
• 2007-10-05 -- can use stacks
• 2007-06-29 -- added "smooth" button

A more complete list of changes is shown in the source code.

Rationale:

Often, I would like to display more images than fits on a screen. JWEB is inconvenient, for several reasons. montage.py recently added the ability to open subsets of images, but it isn't straightforward to manually page through screenfuls of images. The ability to page through screenfuls was the mean feature I wanted to implement.

Getting started:

Documentation for the accompaying batch files is here.

The tarballs containing SPIDER batch files and scripts are here.

The source code is here.

You'll need SPIRE, the SPIDER Reconstruction Engine, installed in order to run this program. If it isn't already installed, I recommend the full installation.

If the network is slow, and if you're running my particle-verification procedure, copying the directory Reconstruction/select/ locally should suffice. If you need to copy the resulting output doc files, I'd recommend using tar, which will preserve the directories, e.g.:

tar -cvf new.tar select/prj*/not*.dat

Procedure:

Assuming the Python script is in your $PATH, run it by typing:

montagefrom.py &

If you're running it for the first time, a pop-up window will appearwith some pre-set defaults. When you change or update the filenames, they get saved. In the following example, I've picked up from a previous session, and the programs recalls those filenames:

If your data extension is not .dat, be sure to change it here.

In the example above, I had been using unstacked images. montagefromdoc.py also works with stacked images:

It may be preferable to enter the input file templates on the command line. (Navigating through the file-browser button can be cumbersome.) The first optional cammand-line argument is the doc file:

montagefromdoc.py select/prj001/goodsel.dat &

The second optional cammand-line argument is the image stack/template:

montagefromdoc.py select/prj001/goodsel.dat select/prj001/stkfilt.dat &

When you click on Continue, a montage of images will pop up:

With the menus overlain:

The selection part should be intuitive. Sometimes, clicking on the bad particles and inverting the selection is easier. Shift-clicking will select the contiguous stretch of particles from the last one you selected.

There are some keyboard shortcuts here (also available in Help-> Keyboard shortcuts:

• Page Down == next screenful of images
• Page Up == previous screenful of images
• CTRL-a == Select all
• CTRL-v == Invert selection (or CTRL-i)
• DEL == Clear all
• CTRL-f == toggle labels
• CTRL-s == Save selections
• CTRL-r == Read selections
• CTRL-w == Close window
• CTRP-o == Open file

From the Open Selections (CTRL-o) menu:

• CTRL-n == increase filenumber
• CTRL-p == decrease filenumber

To do:

I have a queue of TO DOs (some listed in the source code), but if there are any other suggestions, let me know.