https://mijailovic.net/2022/12/25/hkpropel/
Nemanja Mijailovic's Blog
Reverse engineering yet another ebook format
Dec 25, 2022
In the distant past I was super into removing DRM from books and
magazines (check out my posts on that topic: Removing Edge Magazine
DRM, Removing Zinio DRM). This interest eventually faded away, mostly
because I just stopped using websites that wouldn't allow me to
download and own the products I paid for (kudos to eBooks.com for
doing the right thing).
Few weeks ago I decided to buy this book from Human Kinetics. Holding
and reading a massive hardcover book before bedtime is not really my
cup of tea, which is why I opted to buy it in digital form. The
website wasn't mentioning the format of their ebooks anywhere, so
that instantly sounded the alarm in my head. On top of that, the book
description said "Access Duration: 84 Months", and that didn't sound
promising, either (7 years is generous, but anything less than
forever is not generous enough in my book). Despite all the warning
signs, I went ahead and bought the ebook. And as I had previously
suspected, the download option was nowhere to be found. All I got was
this:
[2022-12-25]
Gah, not the custom web viewer! Old me would have started reversing
the website's API immediately, but the new me didn't want to waste
any time on that, so I chose to simply ask for a refund. That turned
out to be surprisingly difficult--it took me more than 10 minutes to
find the refund instructions. I happily clicked the big "Check if
your Ebook is eligible for a refund" button, entered my order code
and then this happened:
Please enter a valid Ebook code.
The Ebook must have been purchased from a Human Kinetics online
store.
Are you kidding me? Had this been a cheaper book, I would have just
given up. But $82 is a lot of money in the book world! My old
instincts kicked in and I decided it would be more fun to hack and
blog for a week than to waste any time dealing with the customer
support.
Maybe just print the book?
You know you are getting old when your first idea is "can't I just
print the book as PDF and move on with my life?". Printing the
webpage directly wouldn't have worked, though, because all viewer
controls would have been printed, too. But based on the looks of the
page, I was pretty sure that part of it was an iframe that contained
the actual content of the book. So I inspected the main HTML element
in developer tools and found this:
[2022-12-25]
There it was, the iframe with the raw HTML content of the book (more
precisely, the current chapter)! I tried to print it and the results
were pretty decent:
[2022-12-25]
Finishing the job would have been simple: save all chapters of the
book as PDF, then maybe even merge all of them into a single
document. But part of me would have been ashamed of half-assing it
like that (although I did create a mental note that the solution was
good enough for all practical purposes, just in case I fail to find
anything better). I was still curious to see if there was a way to
retrieve the book in the original format, whatever that format was,
so I continued my investigation.
Inspecting the network traffic
I had already figured out how to retrieve the HTML content of a
single chapter. My next idea was to check how the viewer was getting
URLs of all chapters, hoping that I would uncover some API that I
could use to get the book metadata. I knew that the file name of the
first chapter was E6372ch01_d78.html, so I looked at all network
requests that came before fetching the chapter itself:
[2022-12-25]
I examined all of these requests, and two of them stood out. Here is
the entire container.xml file:
And here are some interesting parts of package.opf: