carray_test.py

# Adapted from
# http://pytables.github.io/usersguide/libref/homogenous_storage.html#the-carray-class

import os
import numpy
import tables
from itertools import groupby
from operator import itemgetter
from time import time

print_every = 1000

# full DBPedia data
fileName = os.path.expanduser('~/data/dbpedia/carray_all.h5')
infileName = os.path.expanduser('~/data/dbpedia/adjacency_all.npy')
N = 3141881

# # test data from 30K edges
# fileName = os.path.expanduser('~/data/dbpedia/carray_test30k.h5')
# infileName = os.path.expanduser('~/data/dbpedia/adjacency_test30k.npy')
# N = 6463

shape = (N, N)
atom = tables.Float64Atom()
filters = tables.Filters(complevel=5, complib='zlib')
h5f = tables.open_file(fileName, 'w')
ca = h5f.create_carray(h5f.root, 'carray', atom, shape, filters=filters)
print 'Reading data from {}.'.format(infileName)
coords = numpy.load(infileName)
coords.sort()
print 'Sorted data.'
Nrecs = len(coords)
print 'Writing data to {}.'.format(fileName)
tic = time()
lap = tic
totWrittenRecs = 0 # in total
writtenRecs = 0 # every `print_every` iterations
for i, (r, c, w) in enumerate(coords):
    ca[r, c] = w
    writtenRecs += 1
    if i % print_every == 0:
        totWrittenRecs += writtenRecs
        prev_lap, lap = lap, time()
        intertime = lap - prev_lap
        speed = float(writtenRecs) / intertime
        p = float(totWrittenRecs)/Nrecs*100
        print '* {} of {} records ({:.2f}%) written to disk in {:.4f} seconds '\
                '({:.2e} recs/s)'.format(totWrittenRecs, Nrecs, p, intertime, speed)
        writtenRecs = 0
h5f.close()
toc = time()
print 'TOTAL: Wrote {} records to disk in {:.2f} seconds ({:.2e} recs/s)'.format(Nrecs,
        toc - tic, float(Nrecs) / (toc - tic))