Planet Banshee: from the great minds of our community

February 23, 2017

Talk Python Training: Consuming HTTP Services in Python Review

Summary / tl;dr: Consuming HTTP Services in Python is a great addition to the training courses from Talk Python and Michael Kennedy. You’ll come away with a thorough knowledge of the best way to get data from the internet using the requests module; you’ll use real world examples and APIs from Basecamp, Github and a custom API Michael built just from the course; Michael will explain and show the concepts in an easy to learn manner with a little humor and recap each concept to make sure you understand.

In addition to being host of the well known Talk Python podcast, Michael Kennedy has also created a number of Python training courses. The first, Python Jumpstart by Building 10 Apps, launched its Kickstarter exactly a year ago this month, and was quickly followed later in the year with Python for Entrepreneurs on Kickstarter and Write Pythonic Code Like a Seasoned Developer.

I started and finished Python Jumpstart by Building 10 Apps late last year and loved it. It was a very different learning experience than the University of Michigan’s Python for Everybody class on Coursera. There is an assumption with the Talk Python training courses that you have some basic understanding of computer science or programming. I don’t, so I typically go a little slower and take my time with the courses.

Looking back. there are a few things I liked about the Jumpstart by Building 10 Apps course and I was glad to see continue in this latest course:

  • Michael makes it very easy to follow along in the beginning of the courses. Everyone learns differently, but one of the ways I learn best is to follow along by typing the code as he does in the video, helping me commit it to memory.
  • After teaching you a core concept and coding it into one of the apps, Michael recaps what you’ve just learned in its own “Concept” video. This summarizes the concept you just put into practice and reinforces what you’ve learned.
  • Compared to some of the other online courses I’ve taken, I really like that I know I’m learning from someone well known in the community and I believe I’m not just learning how to code, but coding best practices. I don’t know if I’m explaining this right, but as an example: A few of the online classes I’ve taken haven’t had me put the code into functions and then call them in a main(): function, for example.
  • The source code to the examples Michael teaches you is on Github. You can download it, star it, fork it – but it’s available if you want to follow along, code along as the course goes, or just save it for reference for the future.

I’ve shared my enthusiasm for the Talk Python training courses here and on Twitter and when Michael reached out to me last week asking if I was interested in having a sneak peek at his latest course, Consuming HTTP Services in Python, I jumped at it (after making sure he knew I was still a novice early in my Python learning curve). I took a look at the course overview and this is right in my wheelhouse of what I need to learn. A core part of the app I want to build is exactly what this course is about – using the requests module to download at least a half dozen JSON feeds and then building my app around that. (My app is to build the scoring for a custom NFL Pool league – it’s not a fantasy league, it’s different. All of the data comes from MySportsFeeds, who provides sports data via JSON or XML which I will consume, store in a database, and then write a Python program to calculate the league and player scores to be displayed on the league website.)

What I really liked about this course was that it was focused on one thing: consuming services. I’ve taken a few different Python courses online as I try and learn Python, and most are throwing all the basics that you need to know – everything you’d expect in a beginner course, but it does get overwhelming. This was the first course I’ve taken that was focused on getting you really good at one thing, and in a few different ways that you might need to do it.

Immediately, I learn something new. I only knew of requests from I learned using Google and Stack Overflow. When I started playing around and putting together the building blocks of my app, I wrote the following code. MySportsFeed currently using HTTP Basic Authentication, so I have a separate file called that stores my username and password – I may be new to Python, but I’m smart enough to have created that, import it and add it to my .gitignore file!

This code polls the Playoff Team Standings feed on MySportsFeeds and then I have some (ugly) Python code that runs a for loop to rank each of the two NFL Conferences teams from 1 to 16.

response = requests.get(
    auth=HTTPBasicAuth(secret.msf_username, secret.msf_pw))

rawdata = response.content
data = json.loads(rawdata.decode())

And what did I learn? As I tweeted last week:

Now my code looks like this:

response = requests.get(
    auth=HTTPBasicAuth(secret.msf_username, secret.msf_pw))

data = response.json()

It’s not a lot, it’s just one line of code, but it’s these little things. I had no idea the power of requests – this is just one specific example of something I learned from this course. Another thing I learned? I should be taking the URL in the above eample, create a base_url variable and then append the feed name as another variable. This is covered in a later chapter of the course – Consuming RESTful HTTP services. This chapter has a ton of great examples I’m going to be referencing when writing my app and using.

The Consuming RESTful HTTP services chapter is where the course really starts to take off. I ran into this with the Jumpstart course as well – Michael does a great job in teaching you the building blocks and then the course seems to go from 0-60. This is where having previous programming experience is helpful as that jump from learning what each puzzle piece does to how you put the puzzle together clicks. For someone like me, without any programming experience, it’s a big jump, but possible.

With that said, this chapter is fantastic. While I had a cursory knowledge of HTTP commands like GET and PUT, the API Michael built for the course is awesome. You have the opportunity to create your own examples and interact with the API and blog explorer app – this isn’t something you see with most online courses out there.

I also learned that I only want to use requests, and not built-ins. Though I do now have an understanding of the urlib built-in for Python 3.x if I’m ever cornered and have to use it.

I will admit to skipping the chapter on SOAP. I’m a hobbyist, not an enterprise developer who may encounter SOAP. But it’s great this available for those who may need it as part of this course. This, combined with learning how to use JSON, XML, and screen scraping makes it a complete course.

The last chapter is on screen scraping. There are a ton of of tutorials and classes available on the web about screen scraping. I’ve taken a few of them – one of the challenges I have with my app is figuring out the playoff seeding and I thought about scraping, but that’s a different story. This chapter kicks off with an example of using a site’s sitemap.xml – an example I’ve never seen before that makes so much sense once you learn about it. And if a website you want to scrape doesn’t have a sitemap.xml, shame on them for not being search engine friendly. But if they don’t, Michael goes through other ways to scrape a website using Beautiful Soup and does it in the most Pythonic way I’ve seen yet in a course.

I enjoyed Consuming HTTP Services in Python. With the requests module and JSON being a cornerstone of the app I hope to write, it was great to learn about everything I need to know to make that happen. Michael’s delivery is conversational and he makes it easy to follow along and do the code examples with him, if you choose to. If you have programming experience or are coming from a different language, the videos themselves will probably teach you what you need to do in Python. If you’re like me, a complete novice to Python, you’ll be able to follow along, but be prepared for the jump the course will make in the Consuming RESTful HTTP Services chapter – this moves pretty quickly, but if you’ve forked the Github repo you’ll have access to the program Michael has written and you can (and should) write your own examples to interact with the API on the blog explorer. For $39, you’re getting a well developed course from someone well known in the Python community teaching you the Pythonic way interact with services. While other online training sites might have “sales” that are cheaper, as someone new to Python who has taken some of those courses, trust me – the Talk Python courses are well worth the money.

I’m still early in my Python journey and the two courses I’ve finished from Talk Python have been the best learning resources I’ve used out of all the books and training I’ve purchased (and it’s a lot). I’m still working my way through Python for Entrepreneurs and am really looking forward to two of the upcoming courses using SQLAlchemy as this database stuff is way over my head right now. Thanks again to Michael for allowing me to have a preview of the Consuming HTTP Services course – now it’s time for me to take his advice from the last chapter of the course and write some code – the best way to actually learn.

February 14, 2017

The macOS apps I’ll miss the most

I have been considering switching back to GNOME full-time and finally pulled the trigger last week and did, installing Fedora 25 on both my iMac and MacBook Pro. I installed GNOME on my iMac a couple months ago, but didn’t do the installation correctly and screwed up my MBR, resulting in only GNOME being an option. I’ve fixed that this time and have kept dual boot (for just in case and for iTunes on my iPhone and iPad).

The more I’ve thought about this over the last couple months, the more I have wanted to go back to GNOME. The privacy concerns I have about the big tech companies continues to nag at me and there is something about the open source ethos that appeals to me. I may even switch back to Android from iOS if this works well.

I will still be tied to the Apple ecosystem with my work laptop. That’s both good and bad as I think about the few apps that have held me back from making the switch full time. The only alternative would be to switch to Windows, which is never going to happen. I haven’t used Windows since 2004 and considering what Microsoft has done with tracking in Windows 10…

There are a handful of apps on macOS that just don’t have a Linux equivalent, or if they do, aren’t close from a usability experience. The last three are the big ones for me. I also see the irony in that those three apps are some of most expensive applications I’ve purchased through the Mac App Store. You do get what you pay for and I really shouldn’t be comparing these, especially the last two which Apple has featured as apps of the year previously, to free and open source apps. I should be grateful that there are programmers out in the open source world making applications and offering them without charge rather than trying to compare them to Mac equivalents.

In no particular order, the apps I’ll most the most:


I love text messaging from my desktop (and the immediacy of the notifications). I’m old, shouting Get Off My Lawn and just don’t like tapping on virtual keyboards compared to a real keyboard hooked up to a computer. But I can live without this.
Status: Can live without this.


The web client is pretty good and I’ll probably continue to use the iPad as the primary reading device for Pocket. I can live without this. Firefox has a save to Pocket add-on that works just fine.
Status: Can live without this.


Reeder is my RSS reader of choice, and there are a number of RSS readers available on Linux. Feedbin, the replacement service for Google Reader that I pay for annually, also has a decent web interface. New links open in a tab in the browser instead of Reeder’s readability feature. I’ll miss Reeder.
Status: Can live without it.
Update: I’ve found FeedReader in the Fedora 25 repositories. Version 1.6 is in the repo, but the developer has also made a Flatpak available for version 2.0 that was released two days ago and I’m now running. A few thoughts:

  • This has fantastic usability. Almost to the level of Reeder. This is a slam dunk as far as RSS readers go.
  • I installed the Flatpak because version 2.0 adds support for both Feedbin and Pocket as a read it later service. Feedbin suport is working great and after upgrading from the 2.0 beta to 2.0 final, Pocket support is working flawlessly. FeedReader automatically added Pocket as a service since I had it configured in GNOME Online Accounts.
  • A big thank you and shout out to the developers for taking the time to release a Flatpak making it easy for users to upgrade to the latest version.

Updated Status: Found a replacement that is just as good as one of the best Mac apps.


Considering all the work I did over the Christmas holiday to change weak passwords to strong passwords and removing duplicates, and also the integration with iOS, this is a big loss as there is no Linux client for 1Password. There are a few password management alternatives on Linux, but I don’t know how good they are. Ryan C. Gordon aka icculus did write a 1Password script for Linux that may be worth checking out:
Status: More research needed and may just need to switch to Encryptr or Enpass.


Ouch. This one hurts. I love Twitter, it’s the only social network I’m active on. I love syncing my Twitter reading experience between all my devices, which Tweetbot does better than any other application out there, regardless of platform or operating system. I’ve installed Corebird on Fedora and it’s ok, but it’s not Tweetbot.
Status: This one hurts. I can probably confine myself to Twitter on iOS and use Pocket to save and read links.


I love, love, love writing in Ulysses. It’s hands down the best writing app I’ve ever used after trying Scrivener, Hemingway and others. The iCloud integration is great, making it easy to jump to and from other devices, including iOS. I am using Ulysses to not only write for my blog and journal (then importing into Day One) but also as an Evernote replacement after Evernote screwed everyone over with their privacy settings (though they would later backtrack, I’ve lost all trust in them). Like most of the great Mac apps, they’re Apple only. If I’m writing anything, I’m always starting in Ulysses.

I’m using Dropbox Paper right now to try it out as a replacement for Ulysses, and while Paper is close, it’s lack of true Markdown support while writing bugs me. It’s not too bad if I open it in its own browser window and then use it in its own workspace – this makes it feel like more of a writing app and not a browser. I’ve spent significant time learning Markdown for both Ulysses and Day One, so Dropbox Paper missing real keyboard shortcuts for Markdown kind of sucks (some work, like strong and italics, but others, like headings, don’t). I’ve installed the Markdown plugin in WordPress, making it easy to copy and paste drafts from Ulysses to my blog or to Day One. It is possible to export Dropbox Paper as Markdown and after a cursory glance there are some decent looking Markdown editors available on Linux, so there may be hope.
Status: Can probably live without it. But I’m not happy about it.

Day One

This is probably the biggest one for me. If I love Ulysses, I love Day One more. And like Ulysses, Day One is exclusively in the Apple ecosystem. Ironically, I don’t write in my journal nearly as much as I should. But I love the integration with IFTTT and use it to track all of my exercise entries from Endomondo. I spent an hour looking at journaling options on Linux last week, and there are a couple, but I don’t see a way to sync the entries between computers, which is a must have feature. One option is to continue to use Day One on my work laptop or use a Markdown editor on Linux, save in Dropbox, and then import. I’ve also come across jrnl, a command line journaling app that says it works with Day One, but I really love the user experience of Day One’s app. This one hurts the most – Day One was one of the first apps I ever bought in the Mac App Store and I have years of journal entries in there.
Status: Ouch. I really don’t want to miss this. I’m not ready to start journaling in another app, so I’ll probably just write drafts in Dropbox Paper and then use my work laptop for journal entries.

February 10, 2017

Why I’m going back to Linux after five years of using macOS

I’ve been a supporter of the Electronic Frontier Foundation since 2004. Their work on privacy, free expression and technology are all things I am passionate about. For the last year or so, I have become more concerned with privacy issues in technology. The rise in big data and how everything is tracking everything we do has given me significant concerns. I’ve been giving a lot of thought to which ecosystems I want to stay in. I’m not going to say I trust any of these technology companies, but I can control (or minimize) my footprint with some of these companies.

Last year I took a number of steps in this direction:

  • I deleted my Facebook and Instagram accounts. I don’t think I need to go into detail here, but Facebook isn’t something you would ever equate with the word “privacy”.
  • After Evernote said they would access your notes and data (only to backtrack later) I quickly stopped using Evernote.
  • I’m paying cash for most of my personal purchases and now shopping local and not online – even if I have to pay a bit more for things such as records, books or cycling gear.
  • I went through and deleted over a hundred online accounts over the Christmas break and used a password manager to make sure I wasn’t using duplicate passwords online and also that I was using secure passwords.
  • I’m no longer using Flickr (and Yahoo services in general) for my photos and I have a tough decision to make about whether I delete that account and remove access to the photos there. (Wikipedia using a number of my Green Bay Packer photos under a Creative Commons license).
  • I switched to DuckDuckGo instead of Google as my default search engine.
  • As much as I’m intrigued by Amazon’s Alexa and Google Home, I won’t buy a voice activated device. Just think about what data it knows about you – what smart devices in your house, what your saying around it – and the recent story in the news how a police department wants the data scares the shit out of me.
  • I’m not using TouchID on my iOS devices. Courts have ruled multiple times that your fingerprint is not protected under the Fifth Amendment – but a passcode is.

Yes, I sound paranoid. But at the end of the day, this is my decision and my choice. I may not have anything to hide, but I don’t believe just because we have the technology means that it always needs to be used to collect everything about you. While I will never be able to erase everything about me online or with these technology companies – nor would I necessarily want to – I can control with whom I do business and make conscious choices about it. This way I can be eyes wide open that yes, I’ve been using Gmail since it first launched and that Google knows almost everything about me. But that’s my choice to stay within Google’s ecosystem (for now). even if I start to use less of their services, such as switching to DuckDuckGo for internet searches.

I stopped using Microsoft Windows in 2003 when I switched to using Linux full time until about 2012 when I started using macOS after buying my first MacBook. I love Apple’s hardware and I like macOS – the same Unix internals underneath, lots of polish, and excellent apps. Everything just works – you don’t have to fiddle with video card drivers or wireless. But you will have to do things the way Apple wants you to (see: iTunes). Integreation with iOS is great – answer phone calls on your Mac, reply to text messages. But who knows what Apple is tracking as well as the apps you’re using (I’m looking at you Evernote). And don’t get me started on the Touch Bar on the new MacBooks. (No Escape key? Really?)

So I’m going back to using Linux on the desktop after five+ years away. There is no question that the macOS user experience is significantly better. But using the GNOME desktop on Fedora is pretty close and gets better every release. I’ll know my computing experience is secure and private. I’ll probably share some thoughts on what key applications I’ll miss most in a separate blog post. I’ll still need to use macOS at my day job, but I can control what I use at home and have the peace of mind that nothing is tracking me (outside of what’s in my web browser) when using my own computers.