diff options
| -rwxr-xr-x | please | 21 |
1 files changed, 18 insertions, 3 deletions
@@ -14,16 +14,28 @@ from datetime import datetime HDFS_DIR = "hdfs:///user/bnewbold/sandcrawler" HBASE_HOST = "wbgrp-svc263.us.archive.org" +def rebuild_python(): + print("Rebuilding python venv...") + cmd = """cd mapreduce; + export PIPENV_VENV_IN_PROJECT=1; + pipenv install --deploy + tar -czf venv-current.tar.gz -C .venv .""" + subprocess.call(cmd, shell=True) + +def rebuild_scalding(): + print("Rebuilding scalding jar...") + cmd = """cd scalding; sbt assembly""" + subprocess.call(cmd, shell=True) + def run_backfill(args): + if args.rebuild: + rebuild_python() print("Starting backfill job...") output = "{}/output-{}/{}-backfill".format( HDFS_DIR, args.env, datetime.strftime(datetime.now(), "%Y-%m-%d-%H%M.%S")) cmd = """cd mapreduce; - export PIPENV_VENV_IN_PROJECT=1; - pipenv install --deploy - tar -czf venv-current.tar.gz -C .venv . pipenv run ./backfill_hbase_from_cdx.py \ --hbase-host {hbase_host} \ --hbase-table wbgrp-journal-extract-0-{env} \ @@ -58,6 +70,9 @@ def main(): parser.add_argument('--qa', help="run against qa HBase table", action='store_true') + parser.add_argument('--rebuild', + help="rebuild whatever artifact gets sent", + action='store_true') subparsers = parser.add_subparsers() sub_backfill = subparsers.add_parser('backfill') |