GIF2 files Python reader

I create this script on the basis of the code to read the Millennium II data (the same used here) provided by Mike Boylan-Kolchin. Being allowed to read the Fortran code to write and read the GIF2 files I could adapt this script to exactly fit this problem.

import numpy as np
import sys

class head:
    def __init__(self, fname):
        import types
        # start by reading in header:    
        if type(fname) is types.StringType:
            f=open(fname, 'rb')
        elif type(fname) is not types.FileType:
            raise TypeError('argument must either be an open file or ' + 
                            'a string containing a file name')

After the usual imports we define a class to read and contain the file header. These first lines check the “filename” argument.

self.pad = np.fromfile(f, count=1, dtype=bo_mark+'i4')

In the unformatted Fortran binary files the “pad” is a 4-byte space to separate the values of different quantities in the file. bo_mark contains the endianess (“bo” means byte order) of the system allowing Python to correctly interpret the numbers from the binary file. count sets the number of item of type dtype to be read. For example count=3 and dtype=bo_mark+'i4' will store three 4-byte integer in the variable, with the byte order expressed by bo_mark, < for big-endian, > for little-endian.

# npart is an array containing the number of particles in the
        # file divided by type (gas, ...)
        self.npart = np.fromfile(f, count=6, dtype=bo_mark+'i4')
        self.massarr = np.fromfile(f, count=6, dtype=bo_mark+'f8') = np.fromfile(f, count=1, dtype=bo_mark+'f8')
        self.redshift = np.fromfile(f, count=1, dtype=bo_mark+'f8')
        self.flag_sfr = np.fromfile(f, count=1, dtype=bo_mark+'i4')
        self.flag_feedback = np.fromfile(f, count=1, dtype=bo_mark+'i4')
        self.nall = np.fromfile(f, count=6, dtype=bo_mark+'i4')
        self.cooling_flag = np.fromfile(f, count=1, dtype=bo_mark+'i4')
        self.numfiles = np.fromfile(f, count=1, dtype=bo_mark+'i4')
        self.boxsize = np.fromfile(f, count=1, dtype=bo_mark+'f8')
        self.Omega = np.fromfile(f, count=1, dtype=bo_mark+'f8')
        self.OmegaL0 = np.fromfile(f, count=1, dtype=bo_mark+'f8')
        self.Hubblepar = np.fromfile(f, count=1, dtype=bo_mark+'f8')
        self.version = np.fromfile(f, count=1, dtype=bo_mark+'a96')
        self.pad2=np.fromfile(f, count=1, dtype=bo_mark+'i4')
        if type(fname) is types.StringType: f.close()

These lines read and store the values of the quantities saved in the header of the file.

def read_gif2_file(fname):
    bo_mark = '>'
    f=open(fname, 'rb')
    # start by reading in header:    
This is the function we use to read the files: it set the endianness to "big-endian", open the file in read-only mode, read the header and extract the total number of particles.    
````python, 1)
    pos=np.fromfile(f, count=npt*3, dtype=bo_mark + 'f4').reshape((npt, 3))
    return pos, from_what) move the pointer through the file to read bits from one position to another. The position is computed from adding offset to a reference point; the reference point is selected by the from_what argument. A from_what value of 0 measures from the beginning of the file, 1 uses the current file position, and 2 uses the end of the file as the reference point. from_what can be omitted and defaults to 0, using the beginning of the file as the reference point (from the Python documentation).

This last piece of code read the GIF2 galaxy catalogue (an ASCII file with “space separated values”):

#! /usr/bin/env python

file = open('lcdm_galaxy_cat.z0.00', 'rb')
i = 0
for riga in file.readlines():
    parole = riga.split()
    if len(parole) == 11:
        print "iterazione ", i 
        print "aggiungo "
comments powered by Disqus