1-Wire Weather Station

This is a real-time weather station I built using an AAG electronica instrument and an EEE 701 for the server.

I coded the server in Java using Apache’s Mina library for non-blocking sockets. The client is FLEX, and uses Socket (rather than HTTP). The webcam service consists essentially of a Java class that reads the video4linux data as raw RGB and saves it to a JPEG which is then streamed over the socket along with all the other data (I built the server to be pluggable, so each of these little add-ons is a plugin that shares the stream with the other plugins). The archival data lives in MySQL.

I’ve more or less lost interest in this project so it’s not pretty, but on windy days, it is fairly enjoyable to watch! The server itself can be used for just about anything, from a weather station, to a GPS satellite tracker, to the user map that adorns the top-left corner of this site.

Data updates every 30 seconds and comes from here: http://earthquake.usgs.gov/earthquakes/catalogs/

AppendToGatewayUrl modifies the gateway of a NetConnection object with the session id (in this case “jsessionid”) of the current session. If you reconnect with this value appended to the gateway URL, your session can be identified, even on clients that do not use cookies.

There are two ways to handle this:

• set NetConnection.client to an object that implements the AppendToGatewayUrl function, or;
• subclass NetConnection to implement it (since NetConnection.client points to the NetConnection instance by default, you can not reassign the client field if you want this method to work).

There’s one caveat to both approaches, though. It has been my experience that if you call NetConnection.connect while you are receiving a response to bundled remote methods, the data may be lost or a “Client.Data.UnderFlow” error thrown. This occurs because AppendToGatewayUrl gets called after the first result, and the immediate call to NetConnection.connect aborts the rest of the transfer.

Here’s a sample implementation of option A:

To get around the problem with bundled calls, you could save the session id in a field and reconnect before the next call.

Usually what happens is, you’ve tried to cast the reference to the loaded swf to its document type, like so:

If doSomeMethod, for example, plays an animation on the timeline that you have governed with “stop” commands, or that calls methods in the document class, you’ll find that it doesn’t work.

Why? Because you’ve embedded the class definition in the parent movie, and it doesn’t know about the timeline code. Think about it: adding timeline code to a movie effectively makes that movie a subclass of its document class. When you embed a reference to the document class in the parent, the version associated with the child movie (with the timeline code) isn’t used.

An associated problem is that you’ve defeated some of the key reasons for modularizing your sites – to make updates easier, and load code only when it’s needed, not before. Embedding the document classes of your loaded movies defeats the former – because you have to recompile the parent as well as the child when you update the child, and the latter – because you’re loading all the child classes in the parent.

How to solve the problem, then? use interfaces. Make all of your document classes implement a single interface and make absolutely sure that no references to the document classes exist anywhere in the parent movie.

For example, here’s a child movie class:

As you can see, MyChild implements IChild, an interface:

Now, in your parent movie, you’ll use something like this:

Notice that you’re calling a method on the interface, not the document class. In fact, the parent movie should be completely agnostic where the doc classes are concerned. The interface is enough.

Of course, you’re not allowed to output 2- or 1-element pixels, so you have to use image3 for input. What to do? It’s simple really: Say you have a Vector of numbers representing pairs (coordinates, in my case). The first pixel with contain x,y,x and the second y,x,y and so on. All you have to do is check the current pixel’s x-coordinate (assuming the height of your PixelBender input is set to 1) and do the calculations in one or the other order.

Note that in PixelBender, the x,y of the pixel (as reported by outCoord()) is the midpoint of the pixel, so x=0.5,1.5,2.5 and so on. To determine which order to take the calculations in, you just need to know whether x is even or odd.

Kernel Code:

There are a few tricks to using it, though:

Pixel Bender can only accept floats. If you want to operate on doubles, you might be able to work out something with a ByteArray (which is also permitted as Shader input) and Pixel Bender code, but that’s beyond the scope of this article, if it’s even possible.

Though you can feed a Vector of Numbers into your kernel, they are treated internally as pixels. A pixel is an array of floats with one to four elements, but the output pixel cannot have less than three elements. What this means is that you’ll be feeding three or four Numbers into the kernel at a time, executing a computation on each, then returning the results as a 3- or 4-element pixel. In the case of the map projections, I’m feeding in four numbers at a time representing two lat/lon sets.

Because you’ll be operating on chunks of three or four Numbers, the length of your vector must be a multiple of the size of each pixel. For example, if your vector is ten elements long, you must either stop after processing three pixels, or pad out the vector to a length of 12.

Because Pixel Bender considers the data you pass to be a 2-dimensional grid, you have to compute a width and height value for your data. You can use 1 for the height, but only if your data set is small. The kernel will throw an error, saying that the input is too large, otherwise. The best way to fit it all in is to break your data up into a square.

So let’s see an example. Say you want to compute the square root of a list of Numbers. You create a Vector. to store them. Assuming nVec is your vector:

The shader kernel can be loaded with URLLoader, or embedded and passed to the shader as a class instance. To embed the pbj file, you’ll do something like this:

Here’s the Pixel Bender kernel:

So, does using Pixel Bender for math really help? Yes! In my simple tests, using Pixel Bender for collections of Numbers larger than a few hundred is at least 3x as fast and much better still for larger sets and more complex calculations. There is some overhead associated with setting up the shader objects, but depending on what you’re doing, the savings are definitely worth it.

I’m working on a parser in AS3 that will load a GPX file and expose all the information inside as ActionScript objects. See a test using Google Maps below.

Click “Load” to load the file from my server (this is a file containing a short bike ride around Victoria, BC), or click “Browse…” to load your own GPX file.

I haven’t tested this very extensively, so if your file has features that mine doesn’t, it may fail. If it does, please leave a comment! I’ll be uploading this to Google Code as soon as I do some more tests.

Note: you’ll need Flash Player 10 to use the browse button.

Here’s how it works…

Somewhere in your code, you must set up your languages:

import com.robskelly.text.LocalTextField;

LocalTextField.addLanguage('en_us','lang/en_us.xml');
LocalTextField.addLanguage('es','lang/es.xml');

Now you can set a language:

LocalTextField.setLanguage('en_us');

LocalTextField won’t load the language file until you call setLanguage.

LocalTextField itself is a component that wraps a flash.text.TextField instance. You can drag an instance of LocalTextField onto the stage, set its ID parameter, and off you go!

The language files are XML and look like this:

<strings>
	<string id="1">This is a string.</string>
	<string id="2">This is also a string.</string>
	<string id="3" html="true">This is a string with <b>HTML<b /></string>
</strings>

Download a sample here.

This is a pipe-delimited string of integers, so for one character set, say, “Basic Latin,” enter:

For more than one character set, say, “Basic Latin,” plus “Armenian,” enter:

This script uses the IDs from the UnicodeTable.xml file in C:\Documents and Settings\[username]\Local Settings\Application Data\Adobe\Flash CS3\en\Configuration\FontEmbedding. You can create your own character sets by copying or adding glyphRange elements.

For example, I wanted to extend the Basic Latin character set with some European characters. I simply copied the Basic Latin glyphRange element and gave it a new name and ID. Then I opened the charmap utility in Windows (C:\WINDOWS\system32\charmap.exe) and found the Unicode codes for the characters I wanted. The codes in charmap are in the format “0+0000.” Just change the “+” to an “x.” Since the characters in charmap are listed in order of their Unicode codes, you can use the first and last IDs of a range of characters and all those in between will be included. Here’s my “Basic Latin plus European” glyphRange:

<glyphRange name="Basic Latin Plus European" id="30" >
<!-- Ranges copied from "Basic Latin" -->
<range min="0x0020" max ="0x002F" />
<range min="0x0030" max ="0x0039" />
<range min="0x003A" max ="0x0040" />
<range min="0x0041" max ="0x005A" />
<range min="0x005B" max ="0x0060" />
<range min="0x0061" max ="0x007A" />
<range min="0x007B" max ="0x007E" />
<!-- New ranges -->
<range min="0x00C0" max="0x0259" />
<range min="0x1E80" max="0x1EF9" />
</glyphRange>

Download the script here.
Place the file in C:\Documents and Settings\[username]\Local Settings\Application Data\Adobe\Flash CS3\en\Configuration\Commands and run it from the “Commands” menu in Flash.

Update: If you get an error like this:

At line 1 of file "/Users/dan/Downloads/ReplaceTextFormat.jsfl":
TypeError: fl.findObjectInDocByType is not a function

there’s a quick fix. Find the file ObjectFindAndSelect.jsfl in en/Configuration/Javascript and edit the function FlashUtils_findObjectInTimeline, adding this in the first line of the body (around line 29):

if(!objTimeline)return;

This method performs a bunch of operations on that object but fails to check if it’s null first. According to a user in the comments on the livedoc entry for fl.findObjectInDocByType(), pngs in the fla timeline will cause objTimeline to be null. I tested it with a png and, by golly, it’s true. Thanks Dan!

There are several problems with creating loadable font modules for Flash:

• If you embed fonts in a TextField, there’s no way to access the Font instance or class from the swf.
• If you create a font library symbol, there’s no way to alter the character set, other than changing your machine’s locale.
• If you set the same font in a library symbol and a TextField, the charset in the symbol will actually override that in the TextField, and an incomplete charset will be baked into the swf (whaa?)

So, how do you create an accessible font in an swf? FLEX!

Flex allows you to specify specific charsets for embedded fonts, and requires that you define a class for the embedded font. The font can easily be retrieved from the swf after loading. Here’s a usage example.

The font is now registered in the current swf and ready for use.

Now you can do this:

There’s only one catch here. You have to be careful what font name you use. It must be the font’s internal name. Fortunately, there’s an easy way to retrieve it:

You can use fontName in the TextFormat object.

Now, some code:

Create an interface that your target movie will use to retrieve fonts.

Now make an abstract class for the font swfs. This will be a base class for the MXML application. You can use an interface in the application’s default declaration, but if you inherit from Sprite, you can drop some baggage.

Now, create an mxml file that implements everything. The application inherits from FontSource.

The new swf contains the font class (with glyphs) and a mechanism to retrieve it.

In order to publish the font swf, you have to edit your [flex sdk]/frameworks/flex-config.xml file so that it contains the appropriate character ranges. Some charsets are already provided in the [flex sdk]/frameworks/flash-unicode-table.xml file. You simply copy and paste elements from this file (or create your own) and paste them into flex-config.xml.

In flash-unicode-table.xml, you’ll copy the language-range element (in this case, all uppercase latin characters):

Paste the copied element into flex-config.xml: