How do I extract a ieee-be binary file embedded in a zipfile?

Question

I have a set of zip files which contains several ieee-be encoded binary and text files. I have used Pythons ZipFile module and can extract the contents of the text file

def readPropFile(myZipFile):
    zf = zipfile.ZipFile(myZipFile,'r') # Open zip file for reading
    zFileList=zf.namelist() # extract list of files embedded in _myZipFile_

    # text files in _myZipFile_ contain the word 'properties' so 
    # get a list of the property files here
    for f in zFileList:
        if f.find('properties')>0:
            propFileList.append(f)

    # open first file in propFileList
    pp2 = cStringIO.StringIO(zf.read(propFileList[0])) 
    fileLines = []
    for ll in pp2:
        fileLines.append(ll)

    # return the lines in the property text file
    return fileLines

Now I would like to do the same sort of thing except reading the data in the binary files and creating an array of floats. So how would I proceed?

Update 1

The format of the binary files is such that in MATLAB I after extracting them to a temporary location I can read them with the following

>>fid=fopen('dataFile.bin','r','ieee-be');
>>dat=fread(fid,[1 inf],'float');

Update 2

I now have a simple function in attempt to read the binary data something like

def readBinaryFile(myZipFile):
    zFile = zipfile.ZipFile(myZipFile,'r')
    dataFileName = 'dataFile.bin'
    stringData = zFile.read(dataFileName)
    ss=stringData[0:4]
    data=struct.unpack('>f',ss)

but the value I get does the same as the value reported in MATLAB.

Update 3

first float in my binary file

• HEX value: BD 98 99 3D
• float : -.07451103

Answer 1

Most of what you need is in this answer http://stackoverflow.com/questions/1053121/how-do-i-convert-a-python-float-to-a-hexadecimal-string-in-python-2-5-nonworking

See the stuff about struct.pack.

More details on struct are in the Python docs

Answer 2

In the snippet from the question, the "properties files" (whatever that is) are detected, in a rather loose fashion, by the presence of the string 'properties' in their contents, when read as text. I don't know what the equivalent of this would be for binary IEEE-le files.

However, with Python, a easy way to read ieee-le (or other formats) files is with SciPy's io.fopen module.

Edit
Since anyway reading such a binary file requires one to know the structure, you can express this in a struct.pack() format as desribed in Michael Dillon's response! This only requires the std library, and is just as easy!

Answer 3

You could also try the Numpy extension (here), which is a bit lighter than SciPy. Numpy has lots of I/O routines. For example,

import numpy
f = file ('example.dat')
data_type = numpy.dtype ('float32').newbyteorder ('>')
x = numpy.fromfile (f, dtype=data_type)

gives you a numpy array. There's probably a less-clunky way to specify the data type.