10.4.3. Browsing APQC outputs

Overview

It may be that you would like to open up one (or more) subject’s APQC-HTML files and jump to a certain block or image. What wizardry can accomplish this goal? That is the topic of this page.

IDs in the HTML page

Basically, most elements or objects in an HTML page can have an tag or “ID” attached to it. Using that ID within your browser’s address bar, you can jump to that location in the page.

For files on your computer, the address bar just contains the full path of that file, starting with file:// (in place of http:// or https://, which would be used for content on non-local servers). That instructs the browser to load that specific file (usually an HTML-formatted one, ending with the extension .html). For example, the following could be the address bar for the QC-directory’s index.html file for a subject:

file:///data/cereal_study/sub-001/sub-001.results/QC_sub-001/index.html

That will typically load a file in the browser at the top of the page.

You can then specify a place to jump within that page by including the element’s ID in the address bar, separated by a hash symbol # from the file name. For example:

file:///data/cereal_study/sub-001/sub-001.results/QC_sub-001/index.html#id_of_interest

Now the question becomes: how do you discover the key to this puzzle, the element’s ID?

Answer: mostly, in one of two ways:

  1. Each QC block’s label is also that block’s HTML page ID. So, you could use:

    file:///data/cereal_study/sub-001/sub-001.results/QC_sub-001/index.html#ve2a
    

    or

    file:///data/cereal_study/sub-001/sub-001.results/QC_sub-001/index.html#mot
    

    etc. to jump to the top of the given block.

  2. Check this list of QC block labels and images/text sections (which hopefully is kept uptodate...). The QC block IDs and individual image/text IDs are delimited slightly by column.

    Table of most QC-block and item IDs

    Block ID

    Item ID

    (item type) title string

    vorig

    Check: vols in orig space

    EPI

    (image) EPI in original space

    anat

    (image) Anatomical in original space

    ve2a

    (image) Check: vol alignment (EPI-anat)

    epi2anat

    (image) ulay: ..., olay: ...

    va2t

    Check: vol alignment (anat-template)

    anat2temp

    (image) ulay: ..., olay: ...

    vstat

    Check: statistics vols

    fstat

    (image) olay: ...

    mot

    Check: motion and outliers

    enormoutlr

    (image) Motion Euclidean norm (enorm) and outlier fraction

    VR6

    (image) 6 volume registration motion parameters

    regr

    Check: regressors (combined and individual)

    stims

    (image) regressors of interest ...

    df

    (text) Summary of degrees of freedom (DF) usage

    grayplot

    (image) Grayplot (‘-peelorder’) of residuals dset

    warns

    Check: all warnings from processing

    xmat

    (text) regression matrix correlation warnings

    press

    (text) pre-steady state warnings

    TENT

    (text) TENT warnings from timing_tool.py

    qsumm

    Check: summary quantities from @ss_review_basic

    ssrev

    (text) basic summary quantities from processing

Command line examples to view HTML output

Firstly, when modern afni_proc.py finishes, it will show users the @ss_review_basic output per usual, but then also (hopefully) note that the APQC HTML finished building successfully. It will directly prompt you with a command to open+view that hot-off-the-press file, namely with something like:

afni_open -b sub-001.results/QC_sub-001/index.html

Boom! The afni_open -b means that the following HTML file will be opened in your default browser.

You can also specify your own browser. For example, using firefox and utilizing either the absolute or relative path your QC-directory’s index.html file, respectively:

firefox /data/cereal_study/sub-001/sub-001.results/QC_sub-001/index.html

firefox sub-001.results/QC_sub-001/index.html

You can explicitly instruct firefox to open the file in either a new browser window or tab with the following (respectively):

firefox -new-window sub-001.results/QC_sub-001/index.html

firefox -new-tab sub-001.results/QC_sub-001/index.html

To jump to a certain location in the page, you can include that in the command line call directly, by including the desired element’s ID and the hash symbol:

firefox sub-001.results/QC_sub-001/index.html#vstat

firefox sub-001.results/QC_sub-001/index.html#enormoutlr

etc.

Any combination of the above (with/without ID hash, absolute/relative path, and -new-window/-new-tab/blank) can be used.

Using a script to load many APQC-HTML files

If you have several subjects that you have processed, you might want to view them all together quickly. In particular, you might want to:

  • open several subjects’ APQC HTML files;

  • have those files be in their own cluster of browser tabs so that you can use the Tab and ctrl+Tab keys too circulate through them forwards and backwards, respectively;

  • jump to a certain QC-block or QC-item for each HTML file, to be able to flip through the same image(s) easily.

This set of action items can be accomplished with the following scripts.

Warning

When opening new tabs, some browsers might exhibit a brief “flash”, because an empty tab gets created and then populated. Thus, using a script like the ones below to open many tabs in quick succession could lead to a flashing or strobe-like effect. Anyone who might be sensitive to such effects should take this into consideration.

  1. Ex. 1.

    Jump to a particular QC block for all subjects.

     1
     2
     3
     4
     5
     6
     7
     8
     9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    #!/bin/tcsh
    
    # Construct a list of all subjects to view
    set all_files = `\ls sub*/*.results/QC_*/index.html`
    
    # Loop over all the subjects in the list
    foreach ii ( `seq 1 1 $#all_files` )
    
        set ff = ${all_files[$ii]}
        echo "++ Opening: $ff"
        sleep 0.1      # this helps *all* windows open properly
    
        # Open the first HTML a new window, the rest in a new tab
        if ( $ii == 1 ) then
            firefox -new-window ${ff}\#vstat
        else
            firefox -new-tab    ${ff}\#vstat
        endif
    
    end
    
  2. Ex. 2.

    Jump to a particular QC item for all subjects (identical to the previous example, just a different ID used after the #).

     1
     2
     3
     4
     5
     6
     7
     8
     9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    #!/bin/tcsh
    
    # Construct a list of all subjects to view
    set all_files = `\ls sub*/*.results/QC_*/index.html`
    
    # Loop over all the subjects in the list
    foreach ii ( `seq 1 1 $#all_files` )
    
        set ff = ${all_files[$ii]}
        echo "++ Opening: $ff"
        sleep 0.1      # this helps *all* windows open properly
    
        # Open the first HTML a new window, the rest in a new tab
        if ( $ii == 1 ) then
            firefox -new-window ${ff}\#enormoutlr
        else
            firefox -new-tab    ${ff}\#enormoutlr
        endif
    
    end
    

Closing browser tabs

Because information can be saved for each tab/index.html file (e.g., comments and ratings), there is a little bit of standard browser security in place so that changes might not get inadvertently lost. Sometimes, this might be helpful, while at other times, annoyanceful. We propose approaches to reduce the latter and increase convenience (hopefully).

Note

NB: the suggestions provided here refer to using a modern firefox browser (~63-5), where most of this has been tested. I don’t know if this behavior changes much over time or across browsers. Please be aware of your own browser’s behavior and/or quirks in order to prevent loss of work.

If you open+view a QC HTML page and then try to close it, you will likely get this message (along with a couple buttons):

This page is asking you to confirm that you want to leave - data you have entered may not be saved.

As noted above, this feature is to prevent loss of any entered-but-not-saved information in the QC buttons of the page (at present, even if there is no information entered– sorry!). If you have a lot of tabs open, this can be a bit annoying, moving a mouse back and forth.

However, in reality, you don’t have to move your mouse around to verify that you want to close the tab. That would be highly inefficient and unfortunate. Instead, you can proceed as follows:

  • If you are clicking the ‘x’ on the tab to close it, just click that again, and the tab should close.

  • If you are using a keyboard shortcut to close the tab (mine is ctrl+w), then just type that same shortcut again, and the tab should close.

Some comments on browsing

You, Dear Reader, might want to know, How many of these browser tabs could/should be opened at one time before causing computational woe?

Answer: The author of this webpage doesn’t know.

Said author has personally used this for 108 subjects in a group, with two sets of tabs open (i.e., 216 of the APQC HTML documents open), while also having a fair number of other firefox tabs open. This amount did not crash or really even slow down the desktop used for this task. However, an exact maximum is probably unascertainable (until that subtly unhappy point when aaallll memory is used up on a computer) and surely varies per machine. So, as ever, caveat browsor!

To select subsets of a group, one can use useful shell command line tricks. For example, let’s say that the subjects are named in the following style: sub-001, sub-002, sub-003, … sub-999. One could glob over a subset of subjects, say the first 99, by replacing line 4 in one of the examples above with:

set all_files = `ls sub-0*/*.results/QC_*/index.html`

Or, one could get subjects 120-149, inclusively, with:

set all_files = `ls sub-1[2-4]*/*.results/QC_*/index.html`

And there are surely other such “tricks”, too, but I don’t want to rob you of the pleasure of finding them.

Opening images in a new tab

Each individual image can be opened in a separate tab (for example, to zoom in a bit, potentially). Users can either:

  • right-click on an image and select “Open Link in New Tab”, or

  • middle-click on the image.

If you just left-click on the image, it would try to open in the current tab, thereby leaving that page.

Bonus fun fact: the file name of each image contains the HTML page’s ID for jumping to that location. So, if you ever forget that ID while offline or something, you can open the image in a new tab (as learnt above), and check out the file name at the end of the web address. Take the last part of the filename between the last underscore _ and extension .jpg, and there you have it!

For example, if the image’s address in the new tab is

file:///data/cereal_study/sub-001/sub-001.results/QC_sub-001/media/qc_05_mot_enormoutlr.jpg

we would see that the image’s file name is qc_05_mot_enormoutlr.jpg, and therefore its ID would be “enormoutlr”.