def main():
# Get a list of json.bz2 files to read
paths = glob.glob('data/*.json.bz2')
paths = paths[0:4] # Truncate list for example
# Open file for writing results
f = bz2.BZ2File('term_volume_counts.bz2', "w")
# Start a feature reader with the paths and pass the mapping function
feature_reader = FeatureReader(paths)
results = feature_reader.multiprocessing(get_term_volume_counts)
# Save the results
for vol, result in results:
for t, c in result.iteritems(): # result.items() in python3
s = "{0}\t{1}\t{2}\t{3}\n".format(vol[0], vol[1], t, c)
f.write(s.encode('UTF-8')) # For python3, use str(s)
f.close()
def old():
# Get a list of json.bz2 files to read
paths = glob.glob('data/*.json.bz2')
paths = paths[0:4] # Truncate list for example
# Open file for writing results
f = bz2.BZ2File('term_volume_counts.bz2', "w")
# Start a feature reader with the paths and pass the mapping function
feature_reader = FeatureReader(paths)
results = feature_reader.multiprocessing(get_term_volume_counts)
# Save the results
for vol, result in results:
for t,c in result.iteritems(): # result.items() in python3
s = "{0}\t{1}\t{2}\t{3}\n".format(vol[0], vol[1],t,c)
f.write(s.encode('UTF-8')) # For python3, use str(s)
f.close()
def generic_processor(map_func,
result_func,
paths,
outpath=None,
batch_size=1000):
if outpath:
f = bz2.BZ2File(outpath, "w")
else:
f = sys.stdout
csvf = csv.writer(f)
n = 0
m = math.ceil(float(len(paths)) / batch_size)
logging.info("Script started")
while (True):
start = time.time()
batch, paths = (paths[:batch_size], paths[batch_size:])
n += 1
logging.info("Starting batch {0}/{1}".format(n, m))
feature_reader = FeatureReader(batch)
results = feature_reader.multiprocessing(map_func)
result_func(results, csvf)
logging.info("Batch of {0} volumes finished in in {1}s".format(
len(batch),
time.time() - start))
if outpath:
logging.debug("Output filesize is currently: {0}Gb".format(
os.stat(outpath).st_size / (1024**3)))
if len(paths) == 0:
break
logging.info("Script done")
f.close()
