We build json with freqs of ids
This commit is contained in:
@@ -2,17 +2,21 @@
|
|||||||
# -*- coding: utf-8 -*-
|
# -*- coding: utf-8 -*-
|
||||||
|
|
||||||
import sys
|
import sys
|
||||||
|
import json
|
||||||
import os
|
import os
|
||||||
from os.path import join
|
from os.path import join
|
||||||
|
|
||||||
all_ids = set()
|
all_ids = {}
|
||||||
|
|
||||||
for dirpath, dirnames, filenames in os.walk(sys.argv[1]):
|
for dirpath, dirnames, filenames in os.walk(sys.argv[1]):
|
||||||
for f in filenames:
|
for f in filenames:
|
||||||
fp = os.path.join(dirpath, f)
|
fp = os.path.join(dirpath, f)
|
||||||
for uid in open(fp):
|
for uid in open(fp):
|
||||||
all_ids.add(int(uid))
|
uid = int(uid)
|
||||||
|
if uid in all_ids:
|
||||||
|
all_ids[uid] += 1
|
||||||
|
else:
|
||||||
|
all_ids[uid] = 1
|
||||||
|
|
||||||
f_out = open(sys.argv[2],'w')
|
f_out = open(sys.argv[2],'w')
|
||||||
for uid in all_ids:
|
f_out.write(json.dumps(all_ids))
|
||||||
f_out.write(str(uid) + '\n')
|
|
||||||
|
|||||||
Reference in New Issue
Block a user