if (out_date_fn!=None):
out_date_file = open(out_date_fn, 'w')
if (out_msg_fn!=None):
out_file = codecs.open(out_msg_fn, 'w', encoding='ascii', errors='ignore')
if (out_cnt_fn!=None):
out_cnt_file = codecs.open(out_cnt_fn, 'w', encoding='ascii', errors='ignore')
# Create UTF8 -> ascii hash
rewrite_hash = create_utf8_rewrite_hash()
# Loop over files in input list
for input_file in list_file:
input_file = input_file.rstrip()
print 'Reading in file: {}'.format(input_file)
transactions = tt.load_tweets(input_file)
print 'Done'
for key in transactions.keys():
value = transactions[key]
if (key_search!=None):
if (value.has_key(ky1)):
if (value[ky1] != ky2):
continue
else:
continue
if (out_date_fn!=None):
dt = value['date']
fmt_in = '%a %b %d %H:%M:%S %Z %Y'
fmt_out = '%Y%m%d'
dt_python = datetime.strptime(dt, fmt_in)
outfile = os.path.join(destdir, out_fn)
print 'outfile name is : {}'.format(outfile)
if (os.path.exists(outfile)):
print "Outfile already exists, skipping ..., delete to regenerate: {}".format(
outfile)
continue
tmpfile = os.path.join(tmpdir, out_fn)
cmd = "scripts/tweet_to_dict.py --in {} --out {} --verbose {} {}".format(
fn, tmpfile, debug, additional_args)
print "Running command: {}".format(cmd)
os.system(cmd)
# Now add simple metadata -- at tags, hashtags, links
print "Adding simple metadata ..."
transactions = tt.load_tweets(tmpfile)
os.unlink(tmpfile)
tsm.add_simple_metadata(transactions, debug)
# Normalize
print "Performing normalization ..."
tnm.normalize_msgs(transactions, debug)
# Language id on text
# print "Performing language recognition ..."
# tlid.add_lid(transactions, debug)
# Save it
print "Saving to serialized file ... ",
print "outfile: {}".format(outfile)
tt.save_tweets(transactions, outfile)
# Parse input command line options
parser = OptionParser()
parser.add_option("--input_file",
help="input pickled file of tweets",
metavar="FILE")
parser.add_option("--output_file",
help="output pickled file of tweets",
metavar="FILE")
parser.add_option("--verbose",
help="verbosity > 0 -> debug mode",
metavar="FILE",
default=0)
(Options, args) = parser.parse_args()
input_file = Options.input_file
output_file = Options.output_file
debug = int(Options.verbose)
if (input_file == None or output_file == None):
print "Need to specify input and output files -- run with --help for syntax"
exit(1)
print 'Reading in file: {}'.format(input_file)
transactions = tt.load_tweets(infile)
print 'Done'
add_simple_metadata(transactions, debug)
outfile = open(output_file, 'w')
pickle.dump(transactions, outfile)
outfile.close()
if (os.path.exists(outfile_pckl)):
print 'Graph {} already exists, exiting ...'.format(outfile_pckl)
exit(0)
listfile = open(listfn, 'r')
G = nx.DiGraph()
for fn in listfile:
fn = fn.rstrip()
print "Loading file: {}".format(fn)
sys.stdout.flush()
# Load in tweets
fn = fn.rstrip()
xact = tt.load_tweets(fn)
# Add to graph
for ky in xact.keys():
val = xact[ky]
# user node
user = '******' + val['userid'].lower()
if (not G.has_node(user)):
G.add_node(user)
G.node[user]['type'] = 'user'
# Mention list
if (val.has_key('mentions')):
mention_list = copy.deepcopy(val['mentions'])
else:
print "predicted language lui: {}".format(lang)
print
# Main driver: command line interface
if __name__ == '__main__':
# Parse input command line options
parser = OptionParser()
parser.add_option("--input_file", help="input pickled file of tweets", metavar="FILE")
parser.add_option("--output_file", help="output pickled file of tweets", metavar="FILE")
parser.add_option("--verbose", help="verbosity > 0 -> debug mode", metavar="FILE", default=0)
(Options, args) = parser.parse_args()
input_file = Options.input_file
output_file = Options.output_file
debug = int(Options.verbose)
if (input_file==None or output_file==None):
print "Need to specify input and output files -- run with --help for syntax"
exit(1)
print 'Reading in file: {}'.format(input_file)
transactions = tt.load_tweets(infile)
print 'Done'
add_lid(transactions, debug)
outfile = open(output_file, 'w')
pickle.dump(transactions, outfile)
outfile.close()
