alex/gotohack
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Some refactorings, also tabs replaced to spaces

Browse Source
This commit is contained in:
Aleksey Lobanov
2016-02-28 02:42:23 +03:00
parent 30dbc8e16c
commit 614bf2c0d4
1 changed files with 27 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

54
push-pazans-to-mongo.py
View File

@@ -23,36 +23,36 @@ ids = set()
idsFile = sys.argv[3]
if not os.path.isfile(idsFile):
idsRegex = re.compile('\\"_id\\": (.+?),')
with open(usersJsonFile, "r") as file_name:
for line in file_name:
groups = idsRegex.search(line)
uid = int(groups.group(1))
ids.add(uid)
idsRegex = re.compile('\\"_id\\": (.+?),')
with open(usersJsonFile, "r") as file_name:
for line in file_name:
groups = idsRegex.search(line)
uid = int(groups.group(1))
ids.add(uid)
with open(idsFile, "w") as file_name:
for uid in ids:
file_name.write(str(uid) + "\n")
with open(idsFile, "w") as file_name:
for uid in ids:
file_name.write(str(uid) + "\n")
else:
with open(idsFile, "r") as f_ids:
for line in f_ids:
f_ids.add(int(line))
with open(idsFile, "r") as f_ids:
for line in f_ids:
f_ids.add(int(line))
for file_name in os.listdir(dirWithIds):
print("parsing {}".format(file_name))
print("parsing {}".format(file_name))
with open(os.path.join(dirWithIds, file_name), "r") as file:
for line in file:
id = int(line)
if id in ids:
pazan = pazansCollection.find_one(id)
if pazan is None:
pazansCollection.insert_one({"_id": id, "groups": [file_name]})
elif file_name not in pazan["groups"]:
pazan["groups"].append(file_name)
pazansCollection.update_one(
{"_id": pazan["_id"]},
{"$set": {"groups": pazan["groups"]}}
)
with open(os.path.join(dirWithIds, file_name), "r") as f_out:
for line in f_out:
uid = int(line)
if uid in ids:
pazan = pazansCollection.find_one(uid)
if pazan is None:
pazansCollection.insert_one({"_id": uid, "groups": [file_name]})
elif file_name not in pazan["groups"]:
pazan["groups"].append(file_name)
pazansCollection.update_one(
{"_id": pazan["_id"]},
{"$set": {"groups": pazan["groups"]}}
)
print("- done")
print("- done")