Learning Swift Swiftly @learningswiftly-blog - Tumblr Blog

Posts

I'm done with the application! Finally. As told I removed both the forward/backwards button because they're useless in a clock, and you would get random tweets anyhow. I managed to fix the animation by replacing some values. The animation used the startpoint to calculate the fill, which caused it to actually look like it's being run backwards. I swapped this with the endpoint, and the animation ran smooth. I then had the challenge to fit the animation with the seconds that were left. This took me quite some work actually, since the animation is a seperate view in a seperate class. I eventually formatted the seconds I already had from the counter to a single digit number (e.g no more 01, 02 etc, but 1, 8, 12) and then I referenced to my class in my viewcontroller and updated the value of seconds with the fill of the animation. I'm kind of happy to have finished the app. I actually wanted to do more, but the more hours I put in the project the more I noticed my skill level wasn't as far as I'd like yet. I've put the bar too high. I'm now however able to start doing more of the courses that I stopped to do this app for, which I can't wait to start doing again. Total hours spend: +/- 85

It took me a hell of a long time to be able to create animation with Swift. It's far from perfect, but the base of the loader is there. I tried loads of different stuff, but eventually used the tool 'QuartzCode' to generate Swift code from my vectors. I proceeded by creating a custom class, linking it with my viewcontroller and adding a UIView inside my storyboard also linked to the class. This contained the layers with the circles and the animation. I then start the animation from my viewcontroller when the app opens. I'm not fully happy with how the loader looks, and I still need to connect it to the current seconds (which is harder than I expected). I'm somehow not able to mirror the whole animation, which is something I'd have to look into as well. I am however removing the icons which lets the user browse between older tweets. This is for two reasons: #1, it defeats the purpose of the clock. #2, the way I currently build the code every new request would be a random tweet. So if you would go back and forth, no tweet would be the same. To fix this, I'd have to build a caching system or use a backend. This is definitely too much work for me right now, and there's loads of stuff I could've done better in the whole process. But that's also a good way of learning things. I now know what do better next time.

Total hours spend: +/- 80

It's been quite a while since I was able to work on the project. I had to finish a lot of stuff for my internship, but thats done now luckily. This week I'll be focussing more on Swift. I've found a (complex) way of doing a custom loader bar like I designed. I'm currently reading the documentation for CAShapeLayer which supposedly is the best class to use for this purpose. I'm also trying out 'Paintcode' to make graphics which I can immediately turn into clean Swift code and then use in my app. This way I don't have to hardcode the circle myself. I'll update more later this week.

I'm really happy that I've actually made it this far already. I started on implementing the design in the actual app, and the current progress I've made is actually quite good. It's already really starting to look like the app I've designed. Currently, the next/prev buttons aren't doing anything at all yet. I have to copy my current NSURLSession and mod it so that it does a new one whenever the user presses there. I might also build a function that actually saves the tweets up until a certain amount, so there's no need to redo a GET when the user would've had the app open for a certain amount of time. It is however not really necessary I think since it's quite the hassle for such a small improvement.

The next step is doing the custom fonts and the circle loader. I don't think those are going to be a big issue, maybe the styling of the loader will be a bit hard but I'll see that when I get that far. I also have to figure out how to do padding inside the background of a label. As you can see most of the text almost hits the outsides of the label, which isn't really visually appealing. I found out today how to use the Twitter API a little bit. I didn't use it before since I've read that it got way more complex after they put in a certain AuthO system to prevent everyone being able to randomly pull JSON files. I've played around with their Fabric framework a little bit, and even though for this project it might've been perfect, I feel like the learning curve of implementing and using the framework is a little bit too much for me at this moment. I think by doing a custom URL GET, filtering the data from a string etc. i've chosen a good method to demonstrate my current skills. It's not perfect at all, but it does the job. As final comment on this post I would like to point out that I've also build in a sort of 'safety' feauture for whenever twitter would give me their 'error' page. This happened sometimes, maybe 1 of out 50 tweets, and then I would receive their page without any tweets and just a couple of HTML elements. This obviously has something to do with me not using the previously mentioned API and constantly firing get requests from a URL. Currently when this happens the user will be given an error message 'reload the app or wait a minute'. I might just build a button in that will be displayed which the user can tap to redo a GET.

The visual mockup I did for the application. I'm not entirely sure if this is actually do-able for me, but I'm going to try my best at least. What's new?

Progress 'bar' indicating how long it takes for the next tweet to appear. I've found some online (obj-c, which I'd have to convert) tuts on how to create a circular progress bar based on a timer. I've already got the timer so this should be do able. Previous/Next Tweet buttons

I think it'll definitely make the app more usable if you can 'scroll' through previous tweets. However, this means I'd have to pull tweets again. I've already found a method on how to set back my NSDate object 1 minute at a time, so creating the request should be pretty do able. I will not let the user be able to scroll further than the current time, since that would kind of ruin the application. I might put a limit on the 'previous' tweets as well, just to keep the purpose of the app there. Info button

I just want to create a second view containing some info about this project. It's nice to practice, doesn't really add loads of functionality to the app but it's fun. Custom fonts

This shouldn't be too hard. There's quite some information to find online on how to do this.

Substring coloring

This should be the hardest thing to do I think. I want to make the time in the current tweet be a different color. I'll have to play around a lot with substring options to be able to get this done. I have no clue on how to do this yet, but I do think it makes the readability of the time heaps better.

Watchkit Extension

I'd still have to visualize it, but I think it shouldn't really be that hard to build a Watchkit extension as soon as I've finished this app. It really is just recreating the current style in there and displaying the main tweet.

More progress. I’ve decided what I want to put in the MVP of the app and what not:

Previous tweet displaying a tweet one minute earlier

Current tweet being ‘typed’

Timer counting down until the next tweet

I fixed the current tweet quite a lot today. I was noticing that I was getting quite some tweets that weren’t really any good at all, lots of companies and spammers tweeted with the time in their tweets. So I added the words ‘its’ & ‘and’ to the searchquery. This means I’ll get tweets containing the time, ‘its’ & ‘and’ which make it more likely to be a person tweeting instead of some spamaccount. The previous tweet was just a extra that was really easy to implement. Before doing a new search I set a label to the current tweet. That’s all there is to it. The next step is fixing a countdown timer, doing some visual mockups for the app and trying to recreate those visuals. That might be harder than I expected, since I’d like to use custom fonts and pictures. Which are subjects I don’t have a lot of experience with yet.

Hours spend (on this project as of yet): 20

Total hours spend: 63.6

So I’ve picked the possibly worst tweet ever to demo my next progress, but who cares. I’m quitting the twitter ID for now since I don’t think it’s a necessary step to take now, it takes me too much time to fix and I just want to have an MVP out. I’m also now aware of apple’s guidelines for apps, and I’m actually quite sure that it won’t make it past their appstore inspection, but I’m still going to try to finish this project. I’ve build a typewriter effect to display the tweet, and it works pretty good actually. What it does is creating a array out of the string, adding a letter to a string every x milliseconds (again a random number to generate the typing effect), and then display the string after every new character. When a new tweet is pulled from the internet, I start the timer which does all of this, it ends whenever the last character from the string is reached.

Little Update

I'm having quite the issues with figuring out how to display the users twitter ID. I supposed it was as easy as a tweet, but I'm somehow having quite the struggle with grabbing it. Currently I am getting it as console output, but it somehow always contains a few ['s before and lots of stuff after. Maybe I'll redo that part completely. When it's done I want to really start doing the UI since the based of the app will be finished.

Managed to put in 2 hours more work today. I did 2 major things that will help the result of the visible tweet be more of use. I've managed to randomize the tweet thats being taken from the twitter search, this because theres quite some companies tweeting with the time in their tweets so I'd regulary get multiple tweets from the same company in a row. I used arc4random to generate a random number from 1-9, so it will pick a random entry out of the array containing the tweets. I've also been able to write a small loop that checks the current time, if a certain string occurs (e.g 12 up until 23) it will put PM behind the string thats being used for the source, else it puts AM behind there. I might try to grab the twitter users name as well, this seems like a nice way of enriching the app. It won't be included in the MVP I think, but it'll be a nice exercise afterwards. I've also quickly linked a Label to my code so I can see the tweets in real time being updated in stead of just looking at my console output the whole time.

Got it!

I was finally able to solve the annoying error I received all the time. It had nothing to do with the function that was performed, that one worked flawless. The error occurred because the string that was used with the function was incorrect. Pulling information from twitter gets you loads of HTML elements, including lots of double quotes and backslashes. Swift however wasn't able to make a correct string when there's about 20 backslashes and double quotes in the string, so I was trying to parse a invalid string the whole time. I didn't notice the String was invalid since it somehow is valid when it's in an array (which I constantly logged to the console), but when I tried to do other stuff with the string besides parsing I noticed all the errors popping up. I just wrote 2 lines of Swift code that erase all the double quotes and backslashes in my String, and I was then able to parse the string so all the HTML would disappear as well.

Current console output. Only the lines beneath the twitter URL are really relevant, the rest is just some console logging for myself. It's the full tweet containing hashtags and the time! The thing I noticed was that before I put 'am' after the tweet, it wasn't able to pull correct tweets with time in it. So i'll have to write a small piece of code that puts am or pm (whichever is relevant to that time) behind the search string. I'll then write a piece of code that highlights the time in the string so users will be able to see the time before they have to scan the whole tweet. After that, it's mostly interface building and the Minimale Viable Product of the app is done! I'm planning extra stuff obviously, like a watchkit extension and a possibility to save favorite tweets. But that stuff isn't relevant yet. I want to finish an MVP first. The total amount of hours I've put in this piece of code are about 12 already, but it's a really good feeling seeing your app fall together. Even though it's this damn small.

More progress

It's a really cool feeling seeing your app getting together piece by piece, without using a tutorial. I'm however still experiencing the issue with the HTML parsing like I've noted in the past post, but I decided to keep on working on the rest to not get demotivated. The HTML parsing issue will be fixed later, or I will try to create some workaround for it. Currently my app checks every second if the seconds are '00', if so it will initiate a twitter search based on the current hours and minutes. This will eventually return the first tweet thats visible on that twitter search.

This is the console output I'm currently getting. The first thing are the seconds, I print them to the console whenever they are '00' just as a personal check. The second part is the search term I add to the URL, twitter '%3A' is a encoding for ':' which then gets put behind a var containing a string with the twitter URL. These two combined are a new string thats being used by the NSURLSession to pull the information from twitter. I'm aware that everything possibly can be done in a way more efficient way, but I'm already really stoked on the progress I'm making and my own capability of writing this piece of code. The next step for me is building a small part of the interface so I can see labels being updated instead of getting only console output. Then I'll have to face the HTML Parsing eventually, but Stackoverflow gives me some hope that it is actually do able.

Progress

So I started working on the app, and it's been harder to do than I've expected it to be. I've been able to pull a twitter search in using NSURLSession. I put all the data I get in a variable, and using 'componentsSeparatedByString' I'm able to filter out the whole page except for the first tweet that I need. The biggest problem I'm currently facing is that I have a tweet as a String, but this string is filled with HTML elements I don't need. Using the same method as before i'd be able to filter all these elements out, but every tweet differs so it would only work for a single tweet. So I'm trying to write a function that filters out all the HTML elements from the string, which turns out to be a whole lot harder than I've expected. I keep getting kernal errors even though I'm not sure whats wrong with the code. The only good news is whenever I get these fixed, the hardest part of the app is done.

Too Long

It's been too long since I've posted here and also since I've done anything with Swift. At the moment I'm way too busy with my internship and work to be doing anything with Swift, but since the holidays are coming up I'm really hoping to pick it up again. I've got this cool idea I want to turn into an app. I'm not sure where, but I saw art website that displayed a clock based on tweets. Every minute another tweet containing the timestamp of that relevant time was displayed. Even though an app will do literally the same and has no extra value whatsoever (besides being opened up quicker on your phone), it seems like a cool project for me to recreate as an app. It shouldn't be too hard to make either. I'm guessing I will do a NSURLSession containing a twitter search once every minute, filter the content that gets pulled in and display it on the screen. I do however plan to make it a visually good looking app, and maybe even release it as my first app if I'm happy enough with it. This should be fun.

So the Udemy course covered some stuff about Segues, it actually was pretty late in the course (later than I expected it to be). The content of the course wasn't really clear to me either, so I looked somewhere else to find more and easier formatted information. Lynda had a new Swift video course available, and had a nice small chapter about segues and transferring data between them. I decided to follow it, and actually learned way more than I expected. First of all, this teacher uses custom classes for everything he does. This applies to the MCV model way more as well, since you don't code everything in your viewcontroller but actually do stuff outside of your view. Segues are not that hard I found out, you create storyboards and press the element you want to be able to make the seque, ctrl drag it to the new storyboard and voila. You do have to do some simple coding to make the transitions look good, but nothing too hard. Inside a 'prepareForSegue' function you can enter stuff that will be loaded upon using the segue. In this example, I used a array named currentphoto in the tableview. I created a var named secondScene which contains the information about the segue.

var secondScene = segue.destinationViewController as DisplayViewController

// Pass the selected object to the new view controller.

if let indexPath = self.tableView.indexPathForSelectedRow() {

let selectedPhoto = photos[indexPath.row]

secondScene.currentPhoto = selectedPhoto

Hours Spend: 2.5 Total Hours Spend: 43.5

I had a little hour of fun with the new WatchKit. It's not that difficult as I expected it to be. Building layouts for the watch is extremely easy since you don't have to mess with Auto Layout and stuff, theres just basic elements that expand to full width/height or you can give them a prefixed width/height. Then you're able to center them vertically/horizontally and thats really al there is for now. Its easy to build, but it's certainly harder to design for since there's not that much space to use. I haven't really doven into the developing part for the watch, I just did a basic 'how many fingers' app but replaced the fingers with watches. I did however notice that some of the stuff is a little bit different. For example, if I want to change the text on a label on a iPhone app i use varname.text = "string". Here I had to use varname.setText = "String". Not sure why. You have to implement watchkit as a target in your current iPhone app, this creates two new folders. In one of those folders you have a swift file for the view and a swift file for the notifications. In the other folder you have a sort of storyboard as you're used to with iPhone apps. Except that it's way smaller of course. I'm continuing my regular Swift course before I start playing with WatchKit more, but it's definitely fun and easier than I expected it to be. Time spend: 1 hour Total Time Spend: 41 hours

Where Am I

Finally had some time to do some cool stuff again. Finished the whole geolocation part in the udemy course, and build a small app with the learned skills. I found out that it's a real hassle to use the simulator when you're doing stuff with GPS. This is because the simulator itself obviously has no GPS location, the phone isn't anywhere and the simulator is sandboxed so it cant access (and probably will never, for obvious reasons) the GPS module in my macbook. So I had to simulate my GPS using the debug mode on the simulator, which works pretty shit. But I got the job done nonetheless. Excuse the ugly layout, i really couldn't stand the basic labels in xCode anymore so I just HAD to do something. This isn't much better either.

So the app is really basic, it just takes lots of input it gets from the GPS sensor and translates these to strings that fill labels and thus inform the user of some of it's GPS information. I'll explain the code briefly:

manager = CLLocationManager()

manager.delegate = self

manager.desiredAccuracy = kCLLocationAccuracyBest

manager.requestWhenInUseAuthorization()

manager.startUpdatingLocation()

After declaring a varibale with the CLLocationManager type (in this case it's called manager) you give it some properties so it starts initiating. Because the delegate is the viewcontroller, you simply set the delegate to self (referring to the VC). The desiredaccuracy is set to best. The request is basically the app asking the user for permission to use GPS. The startUpdatingLocation part initiates every second (I believe thats a standard perimeter) and updates the location. You then use a function called locationManager to start putting the data in a variable. I created a variable, gave it the type of CLLocation, and that's almost enough. I printed out this variable and it gave me loads of options, for example I could use println('var.coordinates.latitude') and it gave me the user's lat. For the city I had to reverse the geolocation, which can be done with the reverseGeocodeLocation method.

CLGeocoder().reverseGeocodeLocation(userLocation, completionHandler:{(placemarks, error) in

if (error) != nil {

println(error)

} else {

let p = CLPlacemark(placemark: placemarks?[0] as CLPlacemark)

println(p)

self.address.text = "\(p.country)"

}

})

Hours Spend: 2.5 Total Hours Spend: 40

Small Update

I didn't have lots of time to do lots of Swift the past few days unfortunately. My internship currently requires me to learn Axure and Sketch 3, which also take up most of my free time when I'm at home. Although I wasn't able to do Swift, I think both Axure and (mostly) Sketch 3 will come in handy together with my iOS development skills. Sketch 3 allows me to design screens really easily and also export them in the right formats so I can import them in xCode, theres lots of possibilities to design icons, bootscreens etc. so I will definitely use that a lot. I'm also really happy that Apple released WatchKit. Developing for the Apple Watch was one of the main reasons why I wanted to learn Swift. I already took a look at some of the options in the new xCode beta, and they look really promising. I also learned that it's not possible to build stand-alone Apple Watch apps untill late 2015, therefor the first year all the apps I'm able to make will be extensions of an iPhone app, This makes me extra happy that I've started to learn developing for the iPhone at first. I've also pre-ordered the Apple Watch course by the same guy that made the Swift course that I'm currently following. Can't wait for it to release in January. On a final note, I've been doing some small swift-ing as well obviously. I've managed to be able to let a user set a point on a map by pressing the location for 2 seconds. I'm currently busy with using the CoreLocation framework to detect the users current location. The next step is building a small app based on all the Mapkit/CoreLocation functions i've learned about. I'm not too stoked on this subject since I don't really like doing stuff with maps that much, but it's surprisingly more easy than I've expected. And being able to grab the user's location is always a good skill to have.

Hours Spend: 1.5

Total Hours Spend: 37.5

Trending Blogs

Recently Viewed Blogs

Learning Swift Swiftly