Doug Teixeira

When designing a website, I always try to delegate as much work to css as possible, as it's much more light weight than modifying styling using JavaScript.

When polishing up my [portfolio site]http://www.dougteixeira.com), I decided to switch from a rectangular avatar to a circular avatar (which I think looks more sleek).

My choices are: *crop the image file itself (using pixlr.com/editor), but that's way less versatile and I can't apply the crop effect to other images without editing the image file. * Use CSS to crop the image using the clip property.

Syntax

Example:

.example-element { clip: rect(); position: absolute; overflow: hidden; }

Styling Prerequisite:

In order for the clip property to be applied, your element must be styled with:

position: absolute or position: fixed and NOT with position: relative or position: static. This may be changed in the future, but as for now it remains one of the biggest downsides to the clip property.

Properties: Accepts 3 values:

auto: Default behavior is no clip effect. Same thing as not using clip at all.

inherit: Inherit the clip value from the parent container. a shape function. rect() is the only currently accepted function, but more may exist in the future.

Rect() Syntax:

Requires 4 length values, separated by commas starting from the top and continuing clockwise, similar to padding and margins.

Ex:

.example-element { clip: rect(top, right, bottom, left); }

top & bottom: define offset from top border.

I am currently freelancing as a web developer for hire.

I am a web developer with experience in Ruby on Rails, JavaScript, Backbone.js, jQuery, SQL, HTML, and CSS. I received a B.S. in Business Administration with specializations in marketing and a minor in psychology. I have taken coursework on database management, java programming, programming in visual c++, and visual basic.

I co-founded a health IT consulting company for the past 3 years and have taught myself web development with the help of appacademy.io.

My GitHub: http://www.github.com/teixeir3 http://www.DougTeixeira.com/ <-- currently under construction

Today we presented our final projects: part 2 from the past week! I started adding many more models today and am far from finished. But on Monday, we start the hiring bootcamp, where we'll get to hammer in our whiteboarding skills and fancy up my resume with all my updated tech skills. Woohoo!

I will continue working on my prototype, but will also focus on solidifying all the skills I've learned thus far with practice, practice, practice.

it was inevitable with everyone getting sick around me; I am sick and have felt pretty terrible since last night. But I trucked through the day despite my woes, albeit unproductively. I had planned on getting some much more done, oh well.

Today I got my diagnosis fetcher fully working. I search for a icd10 code in the database and then fetch it from the api if it's not found, THEN save a record to the current patient to add it to their list of diagnoses. Very cool.

Even cooler is I'm 90% of the way towards making the fancy selector search so patients can type in a substring query and the client makes an ajax call to the server for each character entered and appends it to the list of selection options. Just like google! Ahhh it's so awesome. I just have to add a handler for inputting the proper text into the text box once the item is selected from the list of choices. So close! I will finish it in the morning :)

I started the day tweaking some CSS, as it was fresh on my mind from studying over the weekend. I made everything look a bit nicer and added some accent colors. CSS isn't complicated, but it makes a lot more sense after putting the work into really learning it this weekend. I added a fancier nav bar with gradients and tried to make my font sizes and block spacing more consistent. It still needs a lot of work, but it's slowly getting better.

Then started the API search: diagnosis codes. I looked for a while to try to find an API that could offer me ICD10 diagnosis codes and descriptions, but couldn't find anything reliable or worthwhile, especially since they are still pretty new. I found a few that offered the old ICD9 codes, which would've worked, but I wanted to design for the future and since it'll be mandatory soon to switch to ICD10, the other APIs didn't cut it.

So I went to the source, CMS.gov. I figured I could parse the codes in XML and save them in my own database, but that turned out to be a little tricky. I actually figured out a way, using the ruby 'crack' gem to translate xml to a ruby hash, where I could then iterate through the parameters to pull out what I needed to save my objects. This was also a bit tricky, since the data contained many nested levels, but I figured that out as well. My idea came somewhat to a halt once I realized exactly how many codes there were (over 300,000) and my computer slowed down just by opening the XML file and I wasn't able to even input just the A codes for parsing without seemingly crashing the terminal. I then realized that my free heroku account would only allow me to save 10,000 DB rows and I decided this wasn't the way (for now). I'd like to eventually make my own API for serving up ICD9 and ICD10 codes, but that seemed a bit outside the scope of this project.

My app is finally coming together, on damp-reaches-3345.herokuapp.com (custom domain soon once domain.com finishes transferring my domain name to namecheap.com) I spent a decent amount of times this weekend refactoring code so that it's more efficient, shorter, and makes more sense. I finally implemented delayed-job, foreman, and scheduler and made a rake task / method that checks all my reminders and creates the entire app's alerts at once. This task is set to be run every 10minutes but I currently have the task turned off as to not waste my precious dyno hours. It works though!

I also beefed up my seed file to test the a much larger amount of information, so I implemented Faker so the info looks real.

Next came the big styling review. I originally rushed through it too much so I went back to organize the different sections, factor out some partials of repeated templates, and just make it look better and more functional overall. It's coming together, but still has a long way to go. Studying the CSS over and over again this weekend did help with my styling technique overall as well.

I also refactored my routes so they're not needlessly nested and verbose, while changing my before_filters to enforce the correct authentication and privileges.

I spent the majority of the day on 3 things: beautifying my app with CSS, refactoring, and alerts!

Now that my reminders work well and self generate themselves (besides a fun cookie overflow error when I generated too many at once ) I started focusing on creating alerts and some methods that I can run scheduled tasks on to create alerts based on the state of the population's reminders. It was actually quite enjoyable, but I didn't quite get to finish testing it all.

I separated out the users controller so that patients and doctors have their own routes and actions, which really simplified things and allowed me to refactor.

Things are starting to come together, but I still feel like I could be working faster. I spent most of the day figuring out general issues with my models and how set reminders / alerts, which are the core functionality of the application. The cool part was setting reminders to auto generate 1 year in advance and to create a new one in a week advance of the last when it is due so that they completely generate automatically. I still have some bugs to work out, but I am progressing. I can't help but feel like I'm behind most other people who are cloning their apps, but I suppose that's the nature of designing something new. I've had to make decisions at each step along the way.

Final project season is here! I've spent all weekend drawing out the wire frame for my prototype health application. It was intense. Like camping ( ^.^ ) I definitely have enough for a minimum viable product, but the crux of the prototype is in the reports / alerts functionality, which I have found a few gems for, but have yet to try. We shall see. So far it seems like my project is way more expansive than the regular final projects, but that's because I'm creating it from scratch instead of cloning something. This will definitely be a test of my skills learned thus far.

Tom and I worked together today cloning gist, which was pretty cool for the most part, but we ran into some trouble with jbuilder. Jbuilder itself isn't too difficult to understand, but I'd highly recommend using the jbuilder templates instead of writing to_builder methods in each of your models. We had some silly bugs overwriting the as_json method instead of the to_json method. There's no real need to overwrite these methods anyway since it blurs the lines between the traditional MVC pattern anyway and the jbuilder templates are easy enough.

Today was an exercise in getting used to a backbone.js heavy app that helped us practice more of the typical syntax, but also demonstrated how we can translate our typical rails designs into backbone. For example, it was the firs time we tied associations between our backbone models. NewsApp is born!

We used simple-rss to fetch rss feeds. We started with the normal db tables, models, associations and were given some basic methods with our feeb and entry models. A feed has many entries and an entry belongs to a feed. The associations are pretty standard, but we had never modeled them in backbone before. The only real trick to incorporating the association to the backbone models / collections was including them on our initial fetch from the database. When we grabbed all the feeds as json, we included all of their associated entries:

{ render :json => Feed.all.as_json(include: :entries) }

The next hurdle was overwriting our Backbone Feed model's parse method, which is called whenever fetching data from the database. We changed our rails controller to send over entry json objects so we had to use our custom parse method to grab those out of the response and instantiate a new collection with them before returning the response. Super cool :)

The next trick was refreshing the entries in the database to include new entries that are updated to each RSS feed, so we added a refresh button to the feed show page which displayed an index of that feed's associated entries. We didn't want to however fetch new entries every time any user presses that button; that would overload our server with more than a handful of users refreshing their feeds. To delegate the task, we made sure to only update the database itself every two minutes by checking that the updated_at column in the database was more than two minutes old before updating our feed in the database, and returned the only data if it was within two minutes. It was a pretty cool app all together, but we could definitely put a lot more work into cleaning it up and adding more features plus extra styling.

I am slightly giddy because today went pretty well! Alex G and I made our first backbone.js app today, which was very cool. Journal App is born!

https://github.com/teixeir3/Journal_App.git

This was our first time exploring backbone models views and controllers, writing the majority of our app in javascript and leaving only the setup, the initial request, and database storage to rails. Requests only hit the server to fetch our initial collection of posts when you visited the root page and to actually create, destroy, and update data in the database itself.

The rest of the work was done by the client itself in the javascript that we sent it on the initial request.

We used the listenTo() method in our View constructors to wait for appropriate events on the collection and to render an updated version of the page simply by resetting or appending to the DOM (not really all the simple but it looks that way in action).

We also did some CSS styling to turn our index page into a sidebar that scrolls with the page. When we had the sidebar functionality worked out, clicking a post link from the index sidebar displayed the show page for that specific post.

In preparation for starting the backbone.js curriculum, we wrote our own lite version of a MVC pattern in Javascript. We created Models, Views (and templates) and methods that essentially performed the same function as routers.

We started with a basic rails app, user authentication, db tables, models etc with all the usual associations and validations. For the rest of the project we focused more on javascript. We tried to keep most of our javascript logic in the client side, to minimize calls to the actual server (mostly just save [create / update]). We made our save, create, and update methods on the javascript models themselves, which performed AJAX calls to the server to perform the necessary requests and then synced up the client side data on a successful db insertion / update.