I have a binary file that I would like to pack as a JSON array like so:
{
"content": [0, 23, 45,...]
}
Right now I dump the file using hexdump into a separate file (to print as unsigned u8's with commas)and manually paste those contents in the array:
hexdump -ve '1/1 "%u," foo.bin > foo_arr
Looking for a better way to achieve this preferably over the command line (jq, standard *nix tools), JavaScript could work as well but I'd rather avoid it.
Here's one option:
hexdump -ve '1/1 "%u\n"' foo.bin | jq -s '{content: .}'
Here, I use jq's -s flag ("slurp") to read in all the lines of stdin as a single array, then simply use that array as the value of content.
For example:
$ python -c 'open("foo.bin", "wb").write(b"abc")'
$ hexdump -ve '1/1 "%u\n"' foo.bin | jq -s '{content: .}'
{
"content": [
97,
98,
99
]
}
Using perl:
$ printf "\x00\x17\x2d" > foo.bin
$ perl -0777 -nE '@bytes = map { ord } split //, $_;
$" = ","; # Delimiter when inserting an array into a string
say qq/{"content":[@bytes]}/' foo.bin
{"content":[0,23,45]}
Reads the entire file at once (-O777 -n) and splits it into an array of bytes, and then outputs the JSON with those byte values.
jq can read raw input using -R (or --raw-input) and convert it to its codepoint numbers using the explode builtin:
jq -Rs '{content: explode}' foo.bin
If the binary file is not too large, you can also read it into a variable using --arg and then apply explode on that.
jq -n --arg bin "$(cat foo.bin)" '{content: $bin | explode}'
Note: jq operates on Unicode codepoints, whereas your current hexdump approach translates single byte counts, so the results may differ to this regard.