Showing posts with label web development. Show all posts
Showing posts with label web development. Show all posts

Monday, February 22, 2010

CakePHP & User Uploads

The first time i tackled this problem i did the obvious thing, and placed the uploaded files in the webroot folder inside of the app folder. However the key problem with this is that you lose the ability to quickly move the app folder around without fear. This is important for me, as sometimes i'm working with bargin basement hosting, so "cap depoly" isn't always an option. Leaving the quickest way to deploy to simply replace the entire app folder via FTP. (We all have to start somewhere!)

Thankfully cakes media views allow you to specify a folder anywhere on the webserver!

What i decided to do was to make a simple controller called media and route all of my USG images through it.



class MediaController extends AppController {
/**
 *
 * @var string
 * @access public
 */
var $name = 'Media';

var $uses = array();

/**
 * Index action.
 *
 * @access public
 */
function index( $file = null, $size = 's'  ) {

$this->view = 'Media';

$components = split('\.',$file);

$params = array(
'id'=> $components[0],
'name'=> $components[0],
'extension'=> $components[1],
'path' => ROOT . DS. 'media' . DS .'filter'. DS . $size .DS . 'transfer' . DS . 'gen' . DS
);


$this->set($params);

}

}


So the controller looks a little something like this in it's raw form. I'm using the Media plugin, hence the addition folder paths. The key thing to note is the 'Path' key/val pair in the $params array. Notice you can pass an absolute file path! so in theory you could even mount another drive, and serve your media from there, pretty awesome.

This way, i can use a url like this

example.com/media/filename.jpg/l


To get a large image, and

example.com/media/filename.jpg/s


And this to get a small image.

Quite handy!

Wednesday, July 23, 2008

OSS: Python webpage link checker

Another piece of python to go. Nothing special, but something fairly useful that i've thrown together. 
The following code samples are released under the terms of the GPL 2.0

It's URL link checker, it's not fully recursive (working on it, they don't teach recursive algorithms on my course). But checks all the anchor tags on the given page, returning warnings and the like depending on what it finds.

Usage: To use this just run it like so "python [scriptname]"
It will then ask for the url of the page you wish to check, and then it will print a report, like so.

Python

#Link Checker
#Brazen attempt to write one in 15 minutes
# Jonathan Dalrymple
# July 15,2008
# Start 11:52
# End 12:35

import httplib
import sgmllib
import re

#HTML Parser
class LinkChecker( sgmllib.SGMLParser ):
def __init__(self, verbose = 0):
sgmllib.SGMLParser.__init__( self, verbose )

self.linkList = []
self.inLink = False
self.lastHref = None
self.hostName = None

print 'Parsing file...'

def parse( self, fileStr ):
self.feed( fileStr )
self.close()

def start_a( self, attr ):

#Show user that the parser is working
#print '*'

self.inLink = True

for name, val in attr:
if name == 'href':
self.lastHref = val

def end_a( self):
self.inLink = False

def handle_data(self, str):

if self.inLink:

tmp = self.__parseUrl( self.lastHref )
self.linkList.append( (str,tmp['host'],tmp['path']) )

def __parseUrl( self, str ):
ret = {}

#slice of preceeding http
#if str[0:7] == 'http://':
# str = str[7:len(str)]

# Extract path regex "\w+\.(\w*)\.(\w{2}\.)?(\w{2,3})"
m = re.compile("\w+\.(\w*)\.(\w{2}\.)?(\w{2,3})").match( str )

if m == None:
#print 'Error in handle url, regex failed'

ret['host'] = None
ret['path'] = str
else:
ret['host'] = m.string[0:m.end()]

if m.end() <>
ret['path'] = str[ m.end(): len(str) ]
else:
ret['path'] = None

return ret
#Check link
def __checkLink( self, displayName, host = None, path = '/' ):
if host == None:
host = self.hostName

#slice path and determine if it is a full url
reqObj = httplib.HTTPConnection( host, 80 )

#print '---Requesting %s' % url+path
reqObj.request('get', path )

response = reqObj.getresponse()

if response.status == 200:
retVal = "SUCCESS |%s returned %d" % (displayName, response.status )
elif response.status == 404:
retVal = "FAILURE |%s returned %d, (%s)" % (displayName, response.status, path )
else:
retVal = "WARNING |%s returned %d, (%s)" % (displayName, response.status, path )

return retVal
def testUrlParser( self, list ):
for u in list:
print '--' + str(self.__parseUrl( u ))
def runReport( self, urlStr ):
urlDict = self.__parseUrl( urlStr )

if urlDict != None:
self.hostName = urlDict['host']
req = httplib.HTTPConnection(urlDict['host'], 80)

req.request('get',urlDict['path'])
response = req.getresponse()

if response.status == 200:

htmlStr = response.read()

self.parse( htmlStr )

print "%d links found" % len( self.linkList )

for v in self.linkList:
#print v
print self.__checkLink( v[0], v[1], v[2] )
else:
print "Download Request for %s failed: %d" % ( response.reason, response.status)


def main():

#urlStr = "http://www.google.co.uk/search?hl=en&q=bar&btnG=Google+Search&meta="
urlStr = raw_input('Url you wish to check:')
if urlStr != None:
bar = LinkChecker()
#foo = ('www.google.com','http://www.google.com','http://docs.python.org/lib/lib.html','docs.python.org/test')
#bar.testUrlParser( foo )
bar.runReport( urlStr )

if __name__ == '__main__':
main()


Quote of the Day
I now finally understand why those guys buy £100,000  cars, only to sit in traffic. It's because they remember the days of running for the train, only to find that it's standing room only and their new best friend is a significantly taller gentleman's (or lady!!) armpit and or sweaty back.

Realizations, Andorra here i come

Thursday, February 21, 2008

Introducing JGrid 0.01: Simple inline Table Editing

Introducing Jgrid.

Basically i got fed up of the sheer amount of configuring that was required by most inline data grid editing systems. And because programmers are lazy, and big headed, i wrote my own.

Currently it's very rough around the edges, and i would imagine it's got a few bugs. Not to mention all the console/debuging code still in place. But i am releasing more for feedback purposes.

I'm always looking to improve my development practices, so please tell me what you think.

Jgrid was built on top of jquery, and is a jquery plugin. I've built it using jquery 1.2.3, which is included in the release package below. A Documentation document (PDF/ODF) is also included as well as a basic html page, which draws data from a database, [HINT: it will break when you run it on your machine!], i plan to update in version 0.02 when i have more than a single callback.

Enjoy.



Download

Sunday, January 27, 2008

The gift that just keeps on giving

After my 14 hour stint yesterday, i'm back on the trail.

Today i've cooked up a Javascript shopping cart. In true fashion it's all OO, and shiny. However very unlike me, it's JS library agnostic! Yes, thats right, i wrote by very own 'getElementById' staements!. I've only provided methods for adding, removing, and getting the price of all the objects. There is also a method for updating a ordered or unordered list, with the items stored in the object.

Javascript (No Library required!)

//Basket Object
//Library agnostic
var BasketObj = function(){
this.items = new Array();
this.bindToElementId = null

//Add an object to the basket
this.addItem = function( pObj ){
retVal = false;
if( pObj.name.length > 1 ){
this.items.push( pObj );
console.log('Object added to basket');
retVal = true;
}
else{
console.log('Not a valid object');
}

return retVal;
}
//Remove the first matching item from the basket and return it
this.removeItem = function( pItemName ){
var retVal = null;

for(i=0; i < this.items.length; i++){

if( this.items[i].name == pItemName ){

console.log('Object removed from basket')
// reference the obj
retVal = this.items[i];
//dereference it
this.items[i] = null;
//exit the loop
break;
}

}

return retVal;
}

//Calculate the total basket value
this.total = function(){
var retVal = 0.00;

for(i=0; i < this.items.length; i++){
retVal += this.items[i].price;
}

return retVal;
}

//Updates the gui display of the object
this.updateDisplay = function(){
if( this.bindToElementId != null ){
//Empty the element first
document.getElementById( this.bindToElementId ).innerHTML = '';

//Sort the array before we print
//this.items.sort();

//Now proceed to fill it
for(i=0; i < this.items.length; i++){
document.getElementById( this.bindToElementId ).innerHTML += '<li>' + this.items[i].name + '</li>';
}
}
}
}


To use it, you must first create the object,


x = new BasketObj();


Once thats done you can set it up, the only thing that you really need to set, is the id name of the object you wish to have the results shown in. This is store in a variable called bindToElementId.

x.bindToElementId = 'basketList';


Now you can add your objects to it. Your objects can be as complex as you want, however they must have the following properties,
name: This is the display name for the object, and the name that will be used for internal lookups
price: This is the value that will be added to create your total,

You use the addItem() to add items to the object, and the removeItem( nameOfObject ) to remove items.

To update your list with the current objects you call updateDisplay. And to get the total value of all the objects you can call total().

One other thing to mention, is that i use console.log statements, so be aware of those, and also i check for the name attribute before adding objects to the internal array, so don't try and add them, and prototype the method to your object later on.

Anyways, this object features in my DFTI coursework, which i've been streaming though over this past weekend, i've now implemented the Drag and drop shopping and it looks a little something like this...



I still have to do the admin page, and stock control logic, however imagine this could be done in a day or so.

Saturday, January 26, 2008

Json Encoding

For a enterprising young developer like myself, i don't take advantage of json enough. So since i'm in the mix at the moment, with my DFTI course work, i thought now would be a good time.

Unfortunately the json_encode() function that i had seen in the PHP documentation, is only in the CVS distributions at the moment (v 5.1.2), so i came up with the below.

PHP

/*
This function converts an array to a Json string.
*/
function jsonEncode( $pArr ){
$arr = Array('{');

foreach( $pArr as $val ){

if( is_numeric( $val) ){
array_push($arr, $val );
}else{
array_push($arr, '"');
array_push($arr, $val);
array_push($arr, '"');
}
array_push($arr,',');
}
//Remove trailing comma
array_pop($arr);

return implode('',$arr);
}

/*
This function converts an associative array to a Json string.
*/
function jsonEncodeAssoc( $pArr ){
$arr = Array('{');

foreach( $pArr as $key => $val){
array_push($arr, '"');
array_push($arr, $key);
array_push($arr, '"');
array_push($arr, ':');

if( is_numeric( $val ) ){
array_push($arr, $val);
}
else{
array_push($arr, '"');
array_push($arr, $val);
array_push($arr, '"');
}

array_push($arr, ',');
}

//Remove trailing comma
array_pop($arr);

array_push($arr,'}');

return implode('',$arr);
}


Both functions take either a regular array, or a Associative and converts it to a json string, which looks like ...


{"id":1,"name":"Ham And Pineapple","description":"<p>Fusce magna sem, gravida in, feugiat ac, molestie eget, wisi.</p>","categoryId":1,"hasNuts":0,"vegetarians":0,"imagePath":"ham-and-pineapple.gif","price":4.95}



The above is just some data from my database. It's a pizza ordering system. I plan to go all ajaxan with a Drag and drop shopping cart, as well as some other shiny stuff.

I haven't a clue what i'm going to do concerning javascript, i'm looking at mootools, mochikit, and of course jquery and maybe even yahoo ui. Might be a last minute thing.

Wednesday, October 31, 2007

Mobile Web development

In recent months I've seriously considered getting into mobile application development. My platform of choice has been symbian thus far. However yesterday something dawned on me. Most people don't understand that modern phones are miniature computers hiding in pink cases. As such the idea of additional applications is moot. However what everyone is now seeing is that the internet on mobile devices is possible. In this crazy online world of ours that means one thing web 2.0!

So from today, i'm gonna put some fuel into creating mobile web 2.0 applications. I have a vague idea of where i'm going, but only time will tell.

In other news, this is my 250th post, meh.

I'm slowly becoming myself again, best reflected by the fact that i have time to write this monologue!

Plus i got the 24" in the end, Benq fp241. ....

Well until another time




Quote of the Day

Caption

Tuesday, August 07, 2007

Database Table Size Calculations

As a professional Developer, every once in a while you'll need to create a Database, for one reason or another. It's always nice to apply structured thought to your design process and work out your proposed database size before who release it upon the unsuspecting database server.

It's kinda obvious to do, but just in case heres how i do it.

First create your data dictionary. In my case it was for a questionnaire management system, (QMS, might release it as OSS when i'm done).

i wanted to dynamically generate a new table for each questionnaire, so i need to work out the system limits before hand.

first we would have our "user id's"

id - UInt
hostName - 128 byte (unicode 64 character string)
ip addr = 15 bytes (it's ip v6)
assocQuestionnaire = 4 bytes aka 32bit unsigned integer

then depending on the questionnaire you could have upto 32 questions/ comment, (32 columns). These could then store upto 1024 characters so thats basically double in UTF-8, so 2048 bytes, bringing in each row at almost 65Kb, now the company i work for has 40 000 members of staff, so lets assume they all anwser the same questionnaire (Great advertising for me), the database would swell to 2.5 Gb!

Well i think thats sight over kill, so by tweaking the the system to only implement the larger fields when requested by a user, and assuming that they only want one comment field per questionnaire, we get a much more wholesome figure of 100 Mb.

So to summerize the formula is

sum( column types ) * num of rows.

Remember this is only an estimate, and doesn't account for the master tables or relationships or any of that other RDMS goodness. So it's a guide not, E=MC2

Jonathan

Thursday, July 19, 2007

CSS research

I'm completely knackered, but research is my life. One of my many clients wanted a semi transparent div. Now i say div, but they just want it to look shiny.

Basically then want an element with appox 20% opacity. Easy right, use a PNG. But that cancels out all ie versions apart from 7. How about a gif instead, you dont get alpha transparency! Well genius how about css, great, there are ways to do it on webkit, gecko, and ie, however they apply the same amount of transparency to their children.

the trick here is to think outside the box, literary...



Thanks to relative positioning, i made a solid div below my transparent div, and moved it over the transparent div, creating a the above effect. The cool things is that by modifing the z-index you can create a ghosting effect!

Best of all, there are no images used to create the above, (well apart form the obvious background pattern on the body), more testing is needed, but i think i might be on to something here.

The CSS
*{
text-align:center;
margin:0;
padding:0;
}
body{
background: url("pattern.gif");
}
div#wrapper{
margin:0 auto;
width:70%;
text-align:center;


padding:1em;
}
div#panel{
opacity: .50;
height:200px;
background: #ddd;
width:100%;
margin:0 auto;
}
div#wrapper p{
padding:1em;
background: #000;
color:#fff;
width:90%;
position:relative;
margin:0 auto;
top:-180px;
z-index:1;
}

The html


Transparent div

content goes here






Now if you excuse me, i'm off to bed