mirror of
https://github.com/zulip/zulip.git
synced 2025-11-02 04:53:36 +00:00
Fixes #14960. The default of 6 thread may not be appropriate in certain configurations. Taking half of the numer of CPUs available to the process will be more flexible.
56 lines
2.0 KiB
Python
56 lines
2.0 KiB
Python
import argparse
|
|
import os
|
|
import tempfile
|
|
from typing import Any
|
|
|
|
from django.conf import settings
|
|
from django.core.management.base import BaseCommand, CommandError, CommandParser
|
|
|
|
from zerver.data_import.slack import do_convert_data
|
|
|
|
|
|
class Command(BaseCommand):
|
|
help = """Convert the Slack data into Zulip data format."""
|
|
|
|
def add_arguments(self, parser: CommandParser) -> None:
|
|
parser.add_argument('slack_data_zip', nargs='+',
|
|
metavar='<slack data zip>',
|
|
help="Zipped slack data")
|
|
|
|
parser.add_argument('--token', metavar='<slack_token>',
|
|
type=str, help='Slack legacy token of the organsation')
|
|
|
|
parser.add_argument('--output', dest='output_dir',
|
|
action="store", default=None,
|
|
help='Directory to write exported data to.')
|
|
|
|
parser.add_argument('--threads',
|
|
dest='threads',
|
|
action="store",
|
|
default=settings.DEFAULT_DATA_EXPORT_IMPORT_PARALLELISM,
|
|
help='Threads to use in exporting UserMessage objects in parallel')
|
|
|
|
parser.formatter_class = argparse.RawTextHelpFormatter
|
|
|
|
def handle(self, *args: Any, **options: Any) -> None:
|
|
output_dir = options["output_dir"]
|
|
if output_dir is None:
|
|
output_dir = tempfile.mkdtemp(prefix="converted-slack-data-")
|
|
else:
|
|
output_dir = os.path.realpath(output_dir)
|
|
|
|
token = options['token']
|
|
if token is None:
|
|
raise CommandError("Enter slack legacy token!")
|
|
|
|
num_threads = int(options['threads'])
|
|
if num_threads < 1:
|
|
raise CommandError('You must have at least one thread.')
|
|
|
|
for path in options['slack_data_zip']:
|
|
if not os.path.exists(path):
|
|
raise CommandError(f"Slack data directory not found: '{path}'")
|
|
|
|
print("Converting Data ...")
|
|
do_convert_data(path, output_dir, token, threads=num_threads)
|