Data Unbound

Here's some code I figured out for doing so that I believe should work:

var tabBrowser = getTabBrowser(chromeWindow);
var tabs = tabBrowser.mTabBox._tabs.childNodes;

for (var i=0; i <tabs.length; i++) {
  var browser = tabBrowser.getBrowserForTab(tabs[i]);
  output(browser.contentWindow.location);
}

It took me a while to gather the pieces, and I'll have to explain how I did it — but for now, let me just note the code here.

Yesterday, I thought that my installations of WordPress did not allow the use of the class attribute in my span HTML elements. To fix the problem, I followed the trail of references leading from » WordPress Strips Classnames, And How To Fix It MaisonBisson.com. I learned that WordPress uses the kses PHP library to filter HTML of possibly invalid and non-secure constructs. It turned out that the kses configuration in WordPress 2.6.3 wasn't responsible for filtering out the class attributes after all — it was the HTML editor in visual mode.

Yesterday, I came across two behaviors of WordPress (v 2.6.3), that I wanted to change:

• the HTML editor sanitizes code in a post if you use the visual editing mode.
• WordPress changes regular single and double quotation marks into "smart quotes", which causes major problems for snippets of code

One way to change this behavior is to edit some of the WordPress source code (the first issues could be resolved by edits described in WordPress Editor Fix to Stop Stripping or Changing Code and the second, by changes shown in phpied.com » Blog Archive » Smart quotes in WP).

However, I was hoping that someone had come up with an approach that would not involve changing WordPress code directly, such as writing up WordPress plug-in to make the change. In the second case (of disabling smart quotes), I did find and install the Unfancy Quote Plugin For WordPress, which "lets you override WordPress' quote fancification". In the first case, I was not able to find such an appropriate plug-in and so would have to write one myself. (WordPress › TinyMCE Advanced « WordPress Plugins might be close to what I want.)

Another possible approach is to mess with my-hacks.php — although that's clearly a legacy approach. A final approach would be to stop using the visual editing mode altogether — that's the workaround that I'll use right now.

(I definitely would like to learn how to write a WordPress plug-in (Writing a Plugin « WordPress Codex) — though not today. There's a good chance I will want to render my mashup book in the form of a WordPress site — and I figure I'll have to write a plug-in or two to make that happen.)

I'm currently learning how to program Zotero, specifically how to integrate Zotero with other applications. I document my learning experience to make it easier for others to learn what I've learned. Note that I'm still learning (I'm far from an expert) — so I think my advice will improve over time. But it decided not to wait until I am an expert before I share my experiences.

You can write Zotero plugins (a Firefox extension that accesses the Zotero extension) as a way of extending Zotero — see plugins [Zotero Documentation].

In a series of posts, I would like to explore how to use Chickenfoot as a scripting framework for Zotero and as a way to explore a working Zotero instance. Why Chickenfoot? Chickenfoot is appealing because it's the closest thing we have to a Web automation framework running within a web browser — and hence something that can take advantage of all the scriptability and context of the Web browser. (Besides, much of the work we do with Chickenfoot will be transferable to writing a Zotero plug-in.)

A goal I've set for myself to focus my learning — and to provide a narrative for the series: create a Chickenfoot script to grab all the references I added on a given date and format those items to be sent out as HTML, wiki formatting, and to be uplaoded to some social bookmarking systems, including delicious and Connotea.

Note: I'm using v 1.0.7 of Zotero, running on Firefox 2.0.0.17 on Windows XP. I'm also using Chickenfoot v 1.0.4. I assume that you will have some basic knowledge of JavaScript and Firefox extensions. (I plan to provide more background later.) The overall approach we take here is a combination of experimenting with bits and pieces of source code, combined with reading the original Zotero source code.

In this first post, we warm up by studying how to access with Chickenfoot Zotero and ZoteroPane, two important Zotero objects, and perform some basic tasks with them.

First, the Zotero object is arguably the main JavaScript object, one that gives you access to the underlying functionality of Zotero. You can access it in this way (see include.js):

var Zotero = Components.classes["@zotero.org/Zotero;1"].
getService(Components.interfaces.nsISupports).wrappedJSObject;

or, in the context of Chickenfoot script:

var Zotero = chromeWindow.Zotero;

(See Rewrite the Web with Chickenfoot [JavaScript & Ajax Tutorials] for a description of the chromeWindow, which is the top-level object of the Firefox browser.)

From a thread on the Zotero forums, I learned an incantation (courtesy of Dan Stillman) to access the ZoteroPane:

var wm = Components.classes["@mozilla.org/appshell/window-mediator;1"]
.getService(Components.interfaces.nsIWindowMediator);
var browserWindow = wm.getMostRecentWindow("navigator:browser");
var ZoteroPane = browserWindow.ZoteroPane;


It turns out but you can also use chromeWindow to get at ZoteroPane:

var ZoteroPane = chromeWindow.ZoteroPane;

What can we do with Zotero and ZoteroPane?

The first thing to do is to examine the two objects. You can examine them the output panel by toggling the output variable to see the children of the output object. For example, if you run the following code, you can get a list of all the children of Zotero and ZoteroPane:


function list_props(obj) {
var name, props;
var props;
props = "";
for (name in obj) {
props = props + " " + name;
}
return props;
}
var Zotero = chromeWindow.Zotero;
var ZoteroPane = chromeWindow.ZoteroPane;
list_props(Zotero);


to generate the following list:


init stateCheck getProfileDirectory getZoteroDirectory getStorageDirectory getZoteroDatabase chooseZoteroDirectory debug log getErrors getSystemInfo varDump safeDebug getString localeJoin getLocaleCollation setFontSize flattenArguments getAncestorByTagName join inArray arraySearch arrayToHash hasValues randomString getRandomID moveToUnique initialized skipLoading startupError startupErrorHandler Prefs Keys Hash Text Date Browser UnresponsiveScriptIndicator WebProgressFinishListener JSON DBConnection DB Schema Item Items Notes Collection Collections Creators Tags CachedTypes CreatorTypes ItemTypes FileTypes CharacterSets ItemFields getCollections Attachments Notifier History Search Searches SearchConditions Ingester OpenURL Translate Cite CSL QuickCopy Report Timeline Utilities Integration File Fulltext MIME ItemTreeView ItemTreeCommandController CollectionTreeView CollectionTreeCommandController ItemGroup ProgressWindowSet ProgressWindow Annotate Annotations Annotation Highlight version isFx2 isFx3 platform isMac isWin isLinux locale dir Maps IconFactory


Calculating the number of top level items

The following simple Chickenfoot script returns the number of top-level items in my Zotero collection:


var Zotero = chromeWindow.Zotero;
var items = Zotero.Items.getAll(true);
items.length + 1;


Toggle the Zotero Panel

The following two-liner shows how to turn the Zotero display on and off (equivalent to hitting Ctrl-Alt-Z)

var ZoteroPane = chromeWindow.ZoteroPane;
ZoteroPane.toggleDisplay();

printing the title of the first selected item

Finally, a three-liner to print out the title of the first selected item:


var ZoteroPane = chromeWindow.ZoteroPane;
var selectedItems = ZoteroPane.getSelectedItems();
var title = selectedItems[0].getField("title");
title;


Conclusion

This post is meant only to lay the foundation for the ultimate goal which is to integrate Zotero and other systems using Chickenfoot. I hope it also whet your appetite for more -- and if you're a Zotero user, that you would go right away to install Chickenfoot to explore Zotero in a deeper way.

This afternoon, I undertook some fall cleanup of my computer's hard drive, a task but became urgent as I start running out of space. Last night, I ran WinDirStat – Windows Directory Statistics to figure out what was hogging up the disk space. I was not surprised to learn my photos, videos, and podcasts ( which I was downloading from iTunes, sometimes indiscriminately) were taking up gigabytes. Every time I buy a new digital camera, I end up taking high-resolution photos, which just exacerbates the problem.

As I was cleaning up my hard drive and backing up photos to an external drive, I found a lot of old photos that I had yet to upload to Flickr. I work hard to properly timestamp my photos so that I can make use of Flickr's ability to let me browse my photos by not only when I upload them but also when I took them. The key is to have the right timestamps in the EXIF headers of my photos. Flickr would then read and use those headers to figure out when the pictures were taken.

The problem is that some of my photos, specifically the ones that came from an old webcam, did not have a timestamp in the EXIF headers. In that case, I could still upload the photos into Flickr and then manually adjust each of the photo's timestamps using Flickr. To avoid using such a tedious and error-prone process, I had written a script in Python that would write the EXIF timestamp from each photo, using the modification date of the photo as an approximation for when the photo was taken.

The last time I looked, I could not find a Python library for doing extensive EXIF manipulation. I ended up trying to different commandline utilities: jhead and exiv2. Today, I modified my script to use exiv2. Let me record a few one-liners that I found useful for the task:

exiv2 filename

reads off the EXIF headers for the file

exiv2 -pt filename

writes out in fuller detail the EXIV headers

exiv2 -da filename

deletes the EXIF headers

exiv2 -k -M"set Exif.Image.DateTime Ascii 2002:01:02 22:15:26" Pict0001.JPG

sets the timestamp for Pict0001.jpg while preserving the file timestamp.

What I'd like to find out now is how to write a timestamp into a video file? My understanding is that EXIF doesn't apply to video files — so what is EXIF for video?