Bash detect ascii art

Question

For bash script, is there a easy way to detect if a text contains ascii art?

An example of ascii art:

Answer 1

First, there's a problem: you didn't mention encoding in your question, which makes me inclined to believe that you aren't aware of how it factors into this. Character encodings determine the meaning of individual bytes. This means we need to start with a small history lesson, and the answer to whether or not this is easy is already "no".

Your example technically isn't ASCII. The extended characters that you're discussing first made their appearance in IBM Code Page 437 (known by various names such as as CP437, OEM 437, and IBM437), but they were not part of the more commonly used internet character sets (i.e. ISO 8859-1 AKA Latin-1) until UTF-8 became the de facto standard. UTF-8 brought these characters back in the form of Block Elements and Box-drawing Characters, among others. It should be noted that the Unicode blocks that I mentioned contain characters that were not originally present in CP437, which makes font support spotty.

With that much understanding, we can break this down into two steps:

1. Determine what encoding a byte stream should be interpreted with.
2. Determine which combinations of bytes are "interesting" to you.

As others have already stated, it's impossible to detect if commonly used text characters (i.e. basically anything in the Latin-1 codepage, including the 7-bit ASCII set) should be interpreted as artistic. There are ranges of characters that you can take more interest in than usual due to a much higher probability that they are being used in art.

For the sake of everyone's sanity, I'm going to assume that we're discussing UTF-8 here. Dissecting the example you provided, we see characters from the following Unicode blocks:

• Geometric Shapes
• Block Elements (which we talked about earlier)

From here on out, everything is implementation specific. You need to examine each character and determine whether or not they fall into a Unicode block that you're interested in. Doing this by hand without experience can be tricky, as UTF-8 is a multi-byte encoding: this means that individual characters outside of the basic 7-bit ASCII set are represented by more than one byte. Using software that is UTF-8 aware will simplify this greatly. I do not recommend trying this using bash alone as the title of your question implies.