HackerRank recently posted a challenge on Hackathon Hackers Facebook group. There's a web page which shows you a sample text and its obfuscated version. You need to figure out a way to get the plain text from the obfuscated text.
Link to the challenge: http://protext.herokuapp.com/
The page uses a generated font called ProText. When a normal font is used to display obfuscated text, it is shown as is, but when ProText font is used, it shows the equivalent plaintext. The obfuscated text is just a bunch of random unicode characters shown using HTML entity codes. Looks simple enough, right? Did I mention the generated font changes every 5 seconds? :P
Apart from the default page, the app also supports taking a parameter to show obfuscated version of any text, like so:
http://protext.herokuapp.com?text=Hello+World
Needless to say, this is what I used in my solution.
This is an extremely simple and lazy solution, and it probably doesn't even qualify. :P But here goes...
- Get the obfuscated version for all possible characters (
abcdefghijklmnopqrstuvwxyz1234567890 ABCDEFGHIJKLMNOPQRSTUVWXYZ) - Generate a mapping between the plaintext alphabet and its obfuscated version.
- Get the obfuscated text and convert it to plaintext using the mapping.
Since the obfuscated code changes every 5 seconds, it's important that both the requests (to get the alphabet and the text) are completed in the span of the same 5 seconds, otherwise the alphabet and text may end up using different obfuscation code and fonts. To that end, I've used requests.session() (see), so that the two requests take advantage of HTTP persistent connection, reducing the time taken significantly.
In my testing, the above workaround led to near perfect results and I was able to get the obfuscated text about 90% of the time.
The script is written in Python 3, using standard packages. To execute, type:
python3 protext.py
If you have a better approach, feel free to send a pull request.