HDF5: How to decode UTF8-encoded string from h5dump output? - php

Question

I'm writing an attribute to an HDF5 file using UTF-8 encoding. As an example, I've written "äöüß" to the attribute "notes" in the file.
I'm now trying to parse the output of h5ls (or h5dump) to extract this data back. Either tool gives me an output like this:
ATTRIBUTE "notes" {
DATATYPE H5T_STRING {
STRSIZE 8;
STRPAD H5T_STR_NULLTERM;
CSET H5T_CSET_UTF8;
CTYPE H5T_C_S1;
}
DATASPACE SIMPLE { ( 1 ) / ( 1 ) }
DATA {
(0): "\37777777703\37777777644\37777777703\37777777666\37777777703\37777777674\37777777703\37777777637"
}
}
I'm aware that, e.g., \37777777703\37777777644 somehow encodes ä as 0xC3 0xA4, however, I have a really hard time coming up with how this encoding works.
What's the magic formula behind this and how can I properly decode it back into äöüß?

Answer 1

The strings are encoded using base 8. I've decoded them in the PHP backend using:
$line = "This is the text including some UTF-8 bytes \37777777703\37777777644\37777777703\37777777666\37777777703\37777777674\37777777703\37777777637";
// extract UTF-8 Bytes
$octbytes;
preg_match_all("/\\\\37777777(\\d{3})/", $line, $octbytes);
// parse extracted Bytes
for ($m = 0; $m < count($octbytes[1]); ) {
$B = octdec($octbytes[1][$m]);
// UTF-8 may span over 2 to 4 Bytes
$numBytes;
if (($B & 0xF8) == 0xF0) { $numBytes = 4; }
else if (($B & 0xF0) == 0xE0) { $numBytes = 3; }
else if (($B & 0xE0) == 0xC0) { $numBytes = 2; }
else { $numBytes = 1; }
$hxstr = "";
$replaceStr = "";
for ($j = 0; $j < $numBytes; $j++) {
$match = $octbytes[1][$m+$j];
$dec = octdec($match) & 255;
$hx = strtoupper(dechex($dec));
$hxstr = $hxstr . $hx;
$replaceStr = $replaceStr . "\\37777777" . $match;
}
// pack extracted bytes into one hex string
$utfChar = pack("H*", $hxstr); // < this will be interpreted correctly
// replace Bytes in the input with the parsed chars
$parsedData = str_replace($replaceStr,$utfChar,$line);
// go to next byte
$m+=$numBytes;
}
echo "The parsed line: $line";