I would like to use regex to increase the speed of my searches for specific records within a large binary image. It seems like regex searches always outperform my own search methods, so that's why I'm looking into this. I have already implemented the following, which works, but is not very fast.
My binary image is loaded into a Numpy memmap as words.
I_FILE = np.memmap(opts.image_file, dtype='uint32', mode='r')
And here is start of my search loop currently (which works):
for i in range(0, FILESIZE - 19):
if (((I_FILE[i] + 1 == I_FILE[i + 19]) or (I_FILE[i - 19] + 1 == I_FILE[i])) and I_FILE[i] < 60):
...do stuff...
This is seeking out records that are 19 bytes long that start with a decimal sequence number between 0 and 59. It looks for an incrementing sequence on either a record before or after the current search location to validate the record.
I've seen a few examples where folks have crafted variables into string using re.escape (like this: How to use a variable inside a regular expression?) But I can't seem to figure out how to search for a changing value sequence.